$QWHFHGHQWVDQG&RQVHTXHQFHVRI0DUNHWLQJ6WUDWHJ\0DNLQJ$0RGHODQGD7HVW $XWKRUVf$QLO0HQRQ6XQGDU*%KDUDGZDM3KDQL7HM$GLGDPDQG6WHYHQ:(GLVRQ 6RXUFH-RXUQDORI0DUNHWLQJ9RO1R$SUfSS 3XEOLVKHGE\$PHULFDQ0DUNHWLQJ$VVRFLDWLRQ 6WDEOH85/http://www.jstor.org/stable/1251943 . $FFHVVHG Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact firstname.lastname@example.org. . American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access toJournal of Marketing. http://www.jstor.org This content downloaded from 22.214.171.124 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
Anil Menon, Sundar G. Bharadwaj, Phani Tej Adidam, & Steven W. Edison Antecedents and Consequences of Marketing Strategy Making: A Model and a Test There is a strong rekindling of academic and practitioner interest in the marketing strategy making (MSM) process and its effect on firm performance. However, there is a dearth of research on process issues in marketing strategy. This study attempts to fill this important gap in the marketing strategy literature by using a discovery-oriented approach to develop a (1) multifaceted conceptualization of MSM, (2) model of the antecedents and consequences of MSM, and (3) test of this model using data on more than 200 marketing mix-related decisions. After ruling out common method bias, the authors find that innovative culture is the fundamental antecedent of effective MSM. They also find that the components of MSM (situation analysis, comprehensiveness, emphasis on marketing assets and capabilities, cross-functional integration, communication quality, consensus commitment, and resource commitment) have differential effects on the outcomes measured, strategy creativity, organizational learning, and market performance. The authors find that strategy creativity affects market performance and organizational learning directly and as a mediator variable. Plans are nothing. Planning is everything.... -Dwight D. Eisenhower n his treatise on the rise and fall of strategic planning, Mintzberg (1994, p. 66) concludes that the "missing detail" in the vast strategy literature is an understanding of how strategies are made. He contends that, for all the tools, techniques, and schools of thought, managers and academics lack a framework outlining a set of activities and processes needed to develop and execute plans. Similarly, several marketing scholars have called for more research on process issues in marketing strategy (Bonoma 1985; Hutt, Reingen, and Ronchetto 1988; Kohli and Jaworski 1990; Menon, Bharadwaj, and Howell 1996; Ruekert and Walker 1987). Recent articles in the popular business press also have expressed the sentiment that there is a lack of understanding of how effective strategies are made (BusinessWeek 1996; Fortune 1996). The study reported in this article is an initial effort toward improving our understanding of how effective marketing strategies are constructed. A review of the literature suggests two potential reasons for this apparent limited understanding of how such strategies are made. First, because scholars generally have distinAnil Menon is Assistant Professor of Marketing, and Sundar G. Bharadwaj is Assistant Professor of Marketing, Goizueta Business School, Emory University. Phani Tej Adidam is Assistant Professor of Marketing, College of Business Administration, University of Nebraska at Omaha. Steven W. Edison is Assistant Professor of Marketing, University of Arkansas. The first two authors contributed equally to the article. The authors thank the Strategic Planning Institute for its support of this study. The authors thank Terry Clark, Roy Howell, Shelby Hunt, Rajan Varadarajan, Jim Wilcox, and the three JM reviewers for their insightful comments on previous versions of this article. guished marketing strategy formulation issues from marketing strategy implementation issues (see Moorman and Miner 1998), research has progressed along these two separate domains. In addition, academic research also has progressed along a dichotomy of rationa) versus incremental planning. The rational school is founded on the core belief that a specialized group or top management deliberately formulates strategy plans, whereas the incremental school argues that strategy plans emerge within the organization. Therefore, researchers in the rational tradition do not include organizational and individual dynamics in their conceptualizations. Conversely, researchers in the incremental tradition tend to deemphasize analysis (Barney 1997; Hart and Banbury 1994; Mintzberg and Quinn 1996). Moreover, researchers have further fragmented the field by narrowing their focus to single dimensions within these domains. For example, in marketing, researchers have conducted depth analysis of dimensions such as integration (e.g., Ruekert and Walker 1987), interaction (e.g., Andrews and Smith 1996), and plan consistency and completeness (e.g., John and Martin 1984).1 Because of these false dichotomies, we lack cumulative knowledge and conceptual frameworks for investigating the process of marketing strategy (Bonoma 1985; Hart 1992; Hutt, Reingen, and Ronchetto 1988). Second, much of the research on strategy planning is based on either researcher conceptualizations of what it should be or what it is in practice (Mintzberg 1994). Although models of strategy planning that are developed only by researchers (without managerial input) are exhaustive IA detailed table summarizing the marketing and strategy research on this issue is available from the authors. Journal of Marketing Vol. 63 (April 1999), 18-40 18 / Journal of Marketing, April 1999 This content downloaded from 126.96.36.199 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
and inclusive, they tend to be complicated and unwieldy for practical application. In contrast, though models of strategy planning developed only by managers (without theoretical underpinnings) are practical, they tend to be too idiosyncratic to their organizations. Thus, it appears that there is a need not only for a conceptualization of marketing strategy making (MSM) that integrates theory and practice, but also for its systematic empirical investigation and validation. Against this backdrop, the key contribution of this article is that it develops and proposes a framework for conceptualizing MSM in the tradition of "integrative science" (Kerin 1996, p. 7). The critical aspects of the proposed MSM framework are threefold. First, it is drawn from extant strategy and marketing theory and integrated with managerial mental models of strategic planning, using a discoveryoriented approach. Second, this multicomponential representation of MSM is holistic and incorporates relevant aspects of marketing strategy formulation and implementation, as well as incremental and rational planning modes. Third, it is based on rigorous empirical investigation using data from 202 marketing mix decisions. Another contribution of this article is that it tests the MSM framework within a nomological net of antecedents and consequences. This departs from prior research, which has focused primarily on either antecedents (cf. Bourgeois and Eisenhardt 1988; Bryson and Bromiley 1993; John and Martin 1984; Sinha 1990) or consequences of strategy planning (cf. Eisenhardt 1990; Ramanujam, Venkatraman, and Camillus 1986), but not both. The few extant empirical studies to investigate both antecedents to and consequences of strategy making have focused on a limited set of components. The choice of antecedents was motivated by a lacuna identified in the literature by Capon, Farley, and Hoenig (1990). These scholars, on the basis of a meta-analysis of 320 studies on the key factors affecting a firm's performance, conclude that the marketing and management strategy field is "badly in need of more work" (Capon, Farley, and Hoenig 1990, p. 1158) on the role of the organization in strategic planning. Consequently, our study focuses on the effect of key organizational antecedents to MSM. A final, unique contribution of this study is the examination of the effect of MSM components on multiple performance measures. Given the general recognition in the strategy literature of the need to measure performance multidimensionally (cf. Chakravarthy 1986), we include multiple outcomes (market performance, organizational learning, and strategy creativity) when assessing the effect of MSM on a firm. By adopting this approach, we offer a framework to managers for how to enhance the quality of the strategy process. In terms of a contribution, it provides managers an a priori basis for focusing efforts on specific components of the whole MSM process and enables post hoc analysis of prior strategies. The MSM Study Overview This study was conducted in three stages using a discoveryoriented approach for (1) developing, qualitative testing, and revising a preliminary model of MSM,2 (2) a rigorous quantitative test of the revised conceptual model, and (3) supplementary managerial input on the empirical results (See Figure 1). A discovery-oriented approach involves supplementing the literature findings with focus group discussions (field-based perspective) involving senior managers in diverse industries (see also Kohli and Jaworski 1990; Lincoln and Guba 1985; Parasuraman, Zeithaml, and Berry 1985). In addition, we follow recommendations in marketing for triangulation by using qualitative methods (i.e., focus group interviews) for theory construction, quantitative methods for theory verification, and qualitative methods for explicating any counterintuitive results. MSM: The Construct Comparing Literature and Field Perspectives The literature on planning in business and marketing strategy is broadly made up of two schools of thought: rational and incremental. The rational planning school's primary emphasis is on the formulation issues in strategy, whereas the incremental school emphasizes implementation issues (Barney 1997; Grant 1995; Nutt 1993). Given these predispositions, it is not surprising to find that the strategy research on formulation and implementation developed as separate streams. Research on formulation traditionally has emphasized assessment of firms' strengths, weaknesses, and external opportunities and threats. The formulation stream of research believes that strategies are made through a great deal of deliberation and a rational evaluation of alternative strategies that is guided by a vision for the organization (cf. Kerin, Mahajan, and Varadarajan 1990; Thompson, Strickland, and Fulmer 1987). Thus, vision has been characterized as a critical component of the planning process because of its ability to provide constancy of purpose for the strategy planning process (Dess and Miller 1993). Researchers belonging to the rational school also have proposed that synergy is a critical component of the strategy process (cf. Ansoff 1965; Day 1986; Kerin, Mahajan, and Varadarajan 1990). In contrast, implementation research has investigated issues related to achieving organizational buy-in for, commitment for, and involvement in the planning process. Because the implementation stream views the planning as a craft, it regards creativity as a key component of the planning process (Mintzberg 1994). However, the viewpoint that the rational/incremental and formulation/implementation dichotomies are problematic is gaining acceptance. For example, some contend, with regard to the rational/incremental modes of planning, that a strategy planning process that emphasizes a single mode is likely to be less successful than one that emphasizes both modes (cf. Hart and Banbury 1994). There is also a growing 2The development of the preliminary model involved a focus group interview with a panel of 26 senior managers who belonged to the Strategic Planning Institute's (SPI) Council on Market Strategy (discussion protocol available from the authors). This preliminary model subsequently was revised, in this stage, through a qualitative face validity test with an additional set of 25 senior SPI managers. Marketing Strategy Making 119 This content downloaded from 188.8.131.52 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
FIGURE 1 Study Overview Stage 1: Discovery-Oriented Model Development Stage 2: Model Test Stage 3: Research Results and Implications Z =Z = These boxes involved drawing on extant literature. w = .j ,iii ii These boxes involved collecting data from senior managers These boxes involved researchers' assimilation of literature and through focus groups and mail surveys. field-based perspectives and model testing. 0 C0 CD 0D -This content downloaded from 184.108.40.206 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
recognition that formulation and implementation converge in time (Bonoma 1985; Hutt, Reingen, and Ronchetto 1988; Moorman and Miner 1998). In addition, empirical research finds that the separation of formulation and implementation hurts organizational performance (Khatri 1996). Moreover, Grant (1995) argues that, independent of whether strategies evolve rationally or incrementally, an analysis of strengths, weaknesses, opportunities, and threats (SWOTs), as well as an evaluation of alternative plans, should be components of any planning process. In summary, the current state of the received view in the literature is that the strategic planning process should include components from both strategy formulation and implementation. The key components are as follows: vision; assessment of SWOTs; evaluation of alternative strategies; buy-in of people; involvement of multiple functions; and creativity in the planning process. In addition to these components, the resource-based school of strategy suggests an additional component for strategy making: an explicit exploitation of the firm's core capabilities. The resource-based view argues that traditional planning approaches and viewpoints do not make an explicit evaluation of the firm's core capabilities relative to the strengths of competitors and its ability to create a competitive advantage (Barney 1991). A mere matching of internal strengths to external opportunities fails to consider if these strengths are superior to competitors (i.e., rare) and are historical and sustainable sources of competitive advantage (i.e., valuable and inimitable). The comparison of the preceding views in the literature with the viewpoints of managers yielded several insights. First, there was a high degree of convergence between the emerging view in the literature and the field perspective. For example, though managers used terms differently and often had fuzzy definitions, the importance of integrating incremental and rational approaches to planning was evident in much of the discussion. Similarly, the intertwined nature of formulation and implementation was well recognized. Thus, the field perspective offered several critical components of MSM: scanning, possession of pertinent skills, motivation, planning for contingencies or alternatives, communication during the strategy, and staying the course. Without exception, managers stated that resource commitment or staying the course was a central element of the planning process for strategy success. In explaining the reasons for failures of strategy, one manager observed that "when the going gets tough, the sponsors often pull out;" another concurred, noting that a "lack of understanding of real resource requirements was the bane of strategy implementation." In discussing scanning, managers identified several aspects, including surfacing critical opportunities and threats, multiple scenario generation, and considering strategic options. Managers contend that strategic thinking too often involves only "our friend Rosy Scenario," without the use of "Darth Vader" scenarios to help balance the analysis of opportunities with threats. Second, the literature perspective on MSM was revised in several key respects on the basis of managerial input. For example, though the literature on planning stresses the value of synergy in developing plans, the managers almost unanimously agreed that synergy was not an aspect of the strategy making process and, at best, was a content issue. Similarly, though the strategy literature identifies vision as an important component of strategy making, the managers were divided about the relevance of vision at a marketing strategy level, with some arguing that it was more appropriate for corporate-level strategy. Third, following the traditional strategy literature, we had initially conceptualized creativity as a component of MSM. However, the discussions with the managers revealed a distinction between the content and the process of creativity. Managers were not able to able to come to a consensus or agreement on defining the process of creativity in MSM. However, much of the discussion seemed to revolve around the creativity of the content of marketing strategy. Therefore, we developed measures of creativity that clearly observed only the outcome of the MSM process in terms of strategy content.3 On the basis of this qualitative test, we develop and propose the following parsimonious conceptualization of MSM. Explicating the Hybridized Conceptualization of MSM On the basis of the two-stage discovery-oriented approach described in the preceding section, we define MSM as follows: a complex set of activities, processes, and routines involved in the design and execution of marketing plans. Therefore, MSM can be conceptualized in terms of seven components: situational analysis, comprehensiveness, emphasis on marketing assets and capabilities, cross-functional integration, communication quality, consensus commitment, and resource commitment. Because these components incorporate traditional views of both formulation and implementation, as well as rational and incremental approaches to strategy, together they form the basis for our hybrid conceptualization of MSM. Situational analysis. Situation analysis refers to rational and systematic consideration of the organizational SWOTs in a marketing strategy domain (cf. Bourgeois and Eisenhardt 1988; Kohli and Jaworski 1990). Because a strategy should articulate how an organization will achieve its goals, a systematic analysis of the market context is necessary to align the strategy and the environment. Some of the managers noted that their firms "were too internally focused to recognize external developments," whereas others countered that their firms were too focused on the external analyses of the opportunities and threats, with little concomitant attention to internal analysis (strengths and weaknesses). Comprehensiveness. In our field research, managers raised the criticality of alternative generations during strategy making and suggested several approaches to alternative 3Our conceptualization of strategy creativity is consistent with the emerging literature that is more focused on creativity in organizations. For example, Ford (1995, p. 19) identifies a "product" definition of creativity, which refers to a judgment made about "some public outcome rather than to a process or a specific person." Also, our measure is similar to the outcome-based measure used by Andrews and Smith (1996). Marketing Strategy Making / 21 This content downloaded from 220.127.116.11 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
strategy generation. In addition to the "Darth Vader" scenario approach mentioned previously, others included approaches such as "locate the exit door," "considering obligatory strategy alternatives," "fishbowl review," and so forth. Thus, though managers and their firms seem to employ different techniques, they all were suggestive of what the academic literature labels "comprehensiveness." Comprehensiveness refers to the systematic identification and indepth evaluation of multiple alternatives to chosen strategy (cf. Eisenhardt 1989; Fredrickson 1983). The literature on comprehensiveness in the strategy domain has focused on the generation and internal consistency of multiple alternatives (cf. Day 1986; Fredrickson and Mitchell 1984). The importance of comprehensiveness in MSM stems from its potential to generate a wide range of strategy options, refine and improve selected strategy, and enhance the confidence in the chosen strategy. Moreover, rejected alternatives can serve as contingency plans. However, one of the factors that influence the value of comprehensiveness is the number of strategies that are considered. Generating too few strategy alternatives increases the likelihood of missed opportunities, whereas the generation of too many options increases time and cost of planning. Thus, comprehensiveness should focus not only on the scope (number or breadth of alternatives), but also on the depth of analysis of the alternatives (Eisenhardt 1989). Emphasis on marketing assets and capabilities. Emphasis on marketing assets and capabilities refers to the historical and ongoing core marketing-related processes, resources, and skills on which the marketing strategy is based (Bharadwaj, Varadarajan, and Fahy 1993; Day 1994). One of the consistent themes that emerged during the discussions with managers was the importance of core capabilities, either resident or developed specifically for the strategy. Managers proposed that strategies should not be developed without the explicit identification and objective evaluation of a firm's ability to execute them. Several managers raised the need for documenting the firm's capabilities as a necessary part of any formal strategy document. This focus on capabilities is consistent with the emerging literature on the resource-based theory of the firm (Barney 1991; Bharadwaj, Varadarajan, and Fahy 1993), which suggests that a firm's marketing advantage stems from its assets and capabilities (Day 1994). Assets refer to investments in scale and scope (e.g., advertising, promotions), brand image, and locational and channel advantages. Capabilities refer to marketing processes and applications of assets such as pricing, customer service capabilities, innovation, and product development (Day 1994). Cross-functional integration. Cross-functional integration refers to the extent to which the MSM team reflects the main organization, includes adequate representation from relevant functional areas, and is well organized and coordinated (Ayers, Dahlstrom, and Skinner 1997; Miller 1987). This definition of cross-functional integration is consistent with Olson, Walker, and Ruekert's (1995), who point out that recent approaches to integration in marketing practice have revolved around the concept of having teams with multiple functional skills and responsibilities. Several managers in the focus group discussions emphasized the importance of having the "right set of functional skills." One manager noted that strategies fail when they do not assess the implementing team's functional mix. Another manager pointed out that, in addition to possessing the appropriate skills set, the team should be well connected to the main organization in functional and political terms. Communication quality. Communication quality refers to the nature and extent of formal and informal communications during the strategy making process (Bonoma 1985; Miller 1987). In our focus groups, the critical role of communication quality in MSM was highlighted in discussions of successful and unsuccessful strategies. Managers pointed out that strategy making is generally not an individual exercise because it requires multiple groups to interact and coordinate activities and functions. Managers ascribed strategy breakdowns, misalignments of activities, and strategy "drifts" to a lack of good interaction among the team members. Conversely, several of the successful strategies discussed in the focus group enjoyed close formal and informal communications among the strategy team members. Good interactions were critical to correcting deviations from intended strategy and for coordination of overall strategy. Resource commitment. Resource commitment refers to the adequate level of people, time, and money allocated to the pursuit of the marketing strategy (Day 1986; Ramanujam, Venkatraman, and Camillus 1986). For example, one of the participants contended that his firm's strategy failed because "funds seem to fall through immediately prior or during implementation." Another manager stated, "the problem is that higher management verbally indicate their commitment to the strategy and plan but fail to support it with resources." Although money is the most obvious, managers also raised the issue of time as a resource constraint; thus, time commitment-for example, meeting deadlines-can be viewed as another aspect of resource commitment. Consensus commitment. Consensus commitment refers to the extent to which members of the strategy team agreed with and supported the chosen strategy (Wooldridge and Floyd 1989). This implies that affected groups within the organization not only "buy into" the strategy, but also are enthusiastic about it. Because consensus reflects a shared understanding of the decision-making process, it reduces uncertainty and allows managers to focus on the substance of their decision (laquinto and Fredrickson 1997). When team members are in consensus about a strategy decision, less time is spent in internal politicking and conflict resolution; the freed-up time can be used more constructively for strategy execution. Because consensus commitment denotes an acceptance of strategy goals by team members, as well as of their roles in the strategy and operational activities, it increases the predictability of behaviors and motives. Conceptual Model of the Antecedents to and Consequences of MSM The conceptual model of MSM developed and tested in this study is composed of three sets of factors: (1) the MSM components, (2) the organizational context factors that are antecedent to MSM, and (3) the consequences of MSM. We 22 / Journal of Marketing, April 1999 This content downloaded from 18.104.22.168 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
next discuss the relationships delineated in the model and depicted in Figure 2. Organizational Antecedents to MSM In this study, organizational context is viewed in terms of two core elements: organizational structure and organizational culture. Organizational structure is defined in terms of centralization and formalization, because they are the most frequently studied in the strategy literature (Menon and Varadarajan 1992) and are among the most significant organizational predictors of firm performance (see the review by Rajagopalan, Rasheed, and Datta 1993). However, the impact of organizational culture on a firm's marketing strategies has been given little scholarly attention (Deshpande and Webster 1989; Mahajan, Varadarajan, and Kerin 1987). In this study, we focus on one type of culture, namely, innovative organizational culture. Centralization. Centralization refers to the extent of decision-making authority concentrated at the higher levels of an organization (Dewar and Werbel 1979). There are two schools of thought on the impact of centralization on situational analysis and comprehensiveness in MSM. One traditional school argues that centralization leads to increased situational analysis and comprehensiveness (cf. Hofer and Schendel 1978). Research in this perspective suggests that, in centralized organizations, planning processes employ specialized tools, techniques, and personnel, all of which tend to increase the level of situational analysis and comprehensiveness in MSM (Shrivastava and Grant 1985). This research also suggests that centralization lowers political activity, thereby enhancing rational strategy making. In contrast, Eisenhardt (1989) and Bourgeois and Eisenhardt (1988) find that centralization is associated with higher levels of political activity and, thereby, with less emphasis on situational analysis and comprehensiveness. This alternative perspective also contends that centralization discourages situational analysis and comprehensiveness because it imposes time and cognitive restraints on senior decision makers (Miller 1987). Several managers in our qualitative research echoed this theme. They complained that their senior management had a narrow and short-term focus, tended not to do a good job in analyzing the strategy context and evaluating alternative strategies, and therefore missed business opportunities. Given these alternative hypotheses, we do not posit any directional relationships between centralization and situational analysis and comprehensiveness. The rationale for the relationship between centralization and emphasis on internal marketing assets and capabilities stems from two streams of research. In the strategy literature, the upper-echelons theory posits that senior managers pursue strategic persistence and show a higher commitment to the status quo (cf. Hambrick and Mason 1984). Also, marketing strategy research suggests that managers emphasize marketing strategies and tactics that historically have provided competitive advantages for the firm in a given market (Gatignon 1984). These research viewpoints were echoed in the focus groups by managers who noted that it is "business as usual" when decisions are completely driven by senior management. Taken together, these arguments point to the hypothesis that centralization leads to a greater FIGURE 2 A Model of the Antecedents to and Outcomes of MSM Outcomes Antecedents Process Marketing Strategy Making / 23 This content downloaded from 22.214.171.124 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
emphasis on historical and ongoing marketing assets and capabilities in MSM. The marketing and strategy literature on the impact of centralization on strategy making suggest that higher levels of centralization are associated with lower levels of crossfunctional integration, communication quality, and consensus commitment. Research in marketing suggests that greater centralization increases levels of alienation, lowers the degree of participation in decision making, and inhibits the healthy exchange of ideas and constructive criticism within an organization (Barclay 1991; John and Martin 1984; Ruekert and Walker 1987). Also, as centralization increases, involvement, communication, and interaction among the different strategic business units (SBUs) and levels of the organization decrease because decision making is concentrated at the top management level (Guth and MacMillan 1986; Wooldridge and Floyd 1989). Increased centralization also can lead to a decline in strategic awareness and understanding in the middle and lower levels of the organization, where strategies are implemented. Consequently, greater centralization can have a deleterious effect on consensus commitment for MSM. In contrast to these negative relationships, we hypothesize that higher levels of centralization have a positive effect, in that necessary resources for MSM are more likely to be allocated by senior management. Rationale for this hypothesis is based on research that suggests that centralization allows top managers to be assertive in allocating the resources necessary to make a plan work, whereas decentralization leads to fragmentation and jockeying for resources, thereby leading to suboptimal resource allocation (cf. Miller 1987). Therefore: Hia: Centralization is associated with situational analysis in MSM. H lb: Centralization is associated with the degree of comprehensiveness in MSM. H lc: Centralization is associated positively with the emphasis on marketing assets and capabilities in MSM. Hid: Centralization is associated negatively with crossfunctional integration in MSM. Hie: Centralization is associated negatively with communication quality in MSM. Hlf: Centralization is associated negatively with consensus commitment in MSM. H g: Centralization is associated positively with resource commitment in MSM. Formalization. Formalization refers to the extent to which rules, procedures, instructions, and communications are written and standardized and the degree to which roles are clearly defined (Pugh et al. 1968). Research has found that increased formalization in an organization leads to higher levels of rationality in planning, recruitment of planning specialists, and more formal analysis, evaluation, and reporting systems (Fredrickson 1986; Miller 1987). Therefore, we hypothesize that formalization is associated positively with situational analysis. However, we propose the inverse relationship between formalization level and degree of comprehensiveness in MSM. The rationale for this is that standardized processes impose time and resource constraints on the degree of comprehensiveness that can be brought to the strategy. Because formalization can lead to ritualistic planning (Lyles and Lenz 1982), efforts are expended in the pursuit of a "document" rather than a comprehensive plan of action (Miller, Droge, and Toulouse 1988). Because formalization also leads to dependence on previously used information and places emphasis on solutions that were successful historically (Fredrickson 1986), marketing strategies tend to reflect historical and existing marketing assets and capabilities. Therefore, we posit that higher levels of formalization will lead to a greater emphasis on marketing assets and capabilities. Formalization has been shown to clarify interdepartmental linkages, functional responsibilities, roles, and skills (Inkson, Hickson, and Pugh 1968), all of which can facilitate the development of cross-functional integration in MSM. Miller (1987) finds a positive relationship between formalization and the involvement and participation of personnel from multiple groups during strategy making. A manager raised the issue that implementing teams fail because of "poor definition of responsibilities ... overlap of responsibilities can allow everyone to drop the ball ... [and] gaps can result from lack of clarity." Such lack of clarity can lower the extent and quality of communications in MSM. Similarly, Jaworski and Kohli (1993) suggest that formalization leads to improved sharing of information within an organization. Therefore, we propose that increased formalization leads to greater communication quality. Other research evidence suggests that formalization is related to lower role conflict, less role stress, and less role ambiguity among marketing personnel (Michaels et al. 1988), which we posit can lead to enhanced consensus for the strategy within the team. Finally, formalization imposes objective criteria and procedures in resource allocations and commitments. By reducing the subjective element, formalized procedures create a greater likelihood that strategies are supported appropriately with needed resources. Therefore, we hypothesize that formalization will be associated positively with resource commitment. In summary: H2a: Formalization is associated positively with situational analysis in MSM. H2b: Formalization is associated negatively with comprehensiveness in MSM. H2c: Formalization is associated positively with the emphasis on marketing assets and capabilities in MSM. H2d: Formalization is associated positively with crossfunctional integration in MSM. H2e: Formalization is associated positively with communication quality in MSM. H2f: Formalization is associated positively with consensus commitment in MSM. H2g: Formalization is associated positively with resource commitment in MSM. Innovative organizational culture. An innovative culture refers to the extent to which there exists within an organization an emphasis on inventiveness, openness to new ideas, and quick response decision making (Menon and Varadarajan 1992; Zaltman 1986). On the basis of a review of the different definitions and conceptualizations of culture, Kerin, Mahajan, and Varadarajan (1990) identify three core dimensions of culture: meanings, communication, and sharedness. 24 Journal of Marketing, April 1999 This content downloaded from 126.96.36.199 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
Meanings refer to frames of reference used by decision makers to describe corporate practices; communications refer to informal and formal codes of behavior that reinforce meanings; and sharedness refers to "shared doings" (e.g., participation is the norm), "shared sayings" (e.g., we respond quickly to change), and a "shared sense of trust" among groups. Consistent with this emphasis on process, our conceptualization of innovative culture refers to a complex set of beliefs and ways of doing things that influence an organization's perspective on how innovation and change should be managed. An innovative culture will tend to foster situation analysis and comprehensiveness in MSM because it motivates systematic attempts to develop, scrutinize, and reconcile divergent perspectives on a strategic option (Miller 1987). Because innovative cultures focus on information and rational processes (Denison and Mishra 1995), decisions usually are based on "numbers" developed by extensive analysis of markets and SWOTs. Support for the proposition that innovative cultures encourage comprehensiveness in MSM is based on the rationale that innovative cultures create a climate that encourages the search for multiple options and new solutions in MSM. Such a climate increases the propensity to analyze information, fosters in-depth examination of strategic alternatives, and generates a desire to find newer and better ways of doing things. One of the managers, reflecting a general sentiment in the focus group discussions, stated that without an innovative culture, "things are done the same way for so long that people neither feel they can nor have a strong motivation to change." Therefore, innovative cultures encourage comprehensiveness in MSM. Innovative cultures encourage exploration and experimentation to develop new businesses within existing business and the renewal or revival of ongoing businesses (Slater and Narver 1995). This is consistent with the central arguments of the core capabilities and innovation, which assert that innovative firms are successful because they exploit and leverage their core capabilities in unique and superior ways (Barney 1991). Extending this line of argument, we hypothesize that, when developing marketing strategies, firms with innovative cultures are likely to choose strategies that emphasize historical and ongoing marketing assets and capabilities. One of the most important roles culture can play in organizations is through its impact on the process of MSM (Deshpande, Farley, and Webster 1993; Deshpande and Webster 1989; Ruekert and Walker 1987). An innovative culture facilitates implementation in several ways. First, by discouraging empire building and resource jockeying, innovative cultures tend to lower functional fixedness and divisional turf battles. In such an internal environment, innovative cultures encourage cross-functional integration in MSM because of the involvement of different groups and personnel in the decision process (Ruekert and Walker 1987; Wooldridge and Floyd 1989). The composition of the strategy-making team is predicated more on the functional skills requirement of the strategy than on departmental parochialism. Second, innovative cultures thrive because diverse functional groups find it easy to share and exchange ideas and also to challenge and communicate openly, all of which we believe contribute to good interaction and communications among the members of the MSM team. Third, innovative cultures foster higher levels of consensus commitment on the part of personnel because plans are developed in an open climate, in which all members are given the opportunity to participate. People "buy into" the strategy and therefore are more likely to devote necessary time and effort to making the plan succeed. Because innovative culture reduces fragmentation and resource jockeying, all efforts can be focused toward the chosen strategy. Therefore, we posit that, in an innovative culture, strategies are backed with appropriate levels of resources. H3a: Innovative culture is associated positively with situational analysis in MSM. H3b: Innovative culture is associated positively with comprehensiveness in MSM. H3c: Innovative culture is associated positively with emphasis on marketing assets and capabilities in MSM. H3d: Innovative culture is associated positively with crossfunctional integration in MSM. H3e: Innovative culture is associated positively with communication quality in MSM. H3f: Innovative culture is associated positively with consensus commitment in MSM. H3g: Innovative culture is associated positively with resource commitment in MSM. Outcomes of Effective Strategy Making Effects of situational analysis. Situation analysis should not only facilitate the understanding and identification of environmental enablers and obstacles to the strategy (Grant 1995), but also put the firm in the position to identify the most appropriate strategy-operating environment fit. For example, a senior executive in the nuclear pharmaceutical business explained that "the general strategy was well thought [out] in many respects ... [but] we disregarded strong dynamics in the healthcare market during the planning stage, which had a negative impact on profits." Although a meta-analytic review of the planning and performance literature (Capon, Farley, and Hulbert 1994) finds that firms using formal planning processes moderately outperform firms that do not use formal planning processes, other empirical research finds lower levels of situational analysis associated with poorer performance (Bourgeois and Eisenhardt 1988; Eisenhardt 1989). Therefore, we hypothesize that situational analysis in MSM leads to superior market performance. Situational analysis entails the acquisition and organization of information to evaluate explicitly a firm's internal capabilities against its external opportunities and threats. This exercise offers the potential for new learning because it can suggest goal-directed behavior in response to changing environmental conditions (Alba and Hutchinson 1987). It also can lead to learning when new information is integrated with existing knowledge regarding competition and the drivers of superior performance (Amabile 1988). Therefore, we propose that situational analysis in MSM will be associated positively with learning. In addition to learning, situational analysis can lead to strategy creativity because it highlights opportunities that can be exploited and threats that must be counteracted (Barney 1997; Grant 1995). HighMarketing Strategy Making / 25 This content downloaded from 188.8.131.52 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
er levels of situational analysis help generate a larger database of knowledge, which is necessary to facilitate creativity (Amabile 1995). Therefore, the preceding discussion suggests that situation analysis leads to more strategy creativity. H4a: Situational analysis is associated positively with strategy creativity. H4b: Situational analysis is associated positively with organizational learning. H4c: Situational analysis is associated positively with market performance. Effects of comprehensiveness. As we stated previously, several of the executives in our study raised the importance of developing multiple strategies and contingency plans to achieve superior organizational performance. Accordingly, we hypothesize that comprehensiveness in MSM is likely to be associated positively with performance, learning, and creativity. We draw on research in decision processes and learning to support this field-based proposition. Decision process literature suggests that generating and evaluating multiple alternatives is more effective because alternatives are difficult to assess in isolation (Eisenhardt 1989). The effectiveness rationale emerges because the process of comparing strategy alternatives enables managers to judge better the viability of different strategy options and the probability that the best option is chosen (Schweiger, Sandberg, and Ragan 1986). Empirical research also has found support for a positive relationship between comprehensiveness and firm performance (McKee, Varadarajan, and Vassar 1990). From a learning perspective, comprehensiveness forces managers into a "hypotheses testing mode" (Eisenhardt 1989, p. 558), which creates new insights, gives a better understanding of the strategy, and increases confidence in decisions and in the managerial decision-making ability (Day 1994). Comprehensiveness in MSM also enhances learning because it facilitates the development of richer and more complex mental maps (Slater and Narver 1995). Similarly, the process of comparing and contrasting strategy alternatives can facilitate creativity because it requires thoughtful reflection and encourages deviations from the status quo (Amabile 1988; Andrews and Smith 1996). Therefore, we propose that H5a: Comprehensiveness is associated positively with strategy creativity. H5b: Comprehensiveness is associated positively with organizational learning. H5c: Comprehensiveness is associated positively with market performance. Effects of emphasizing marketing assets and capabilities. Research suggests that emphasizing a firm's strengths can lead to superior market performance, especially if multiple capabilities are emphasized (Bharadwaj, Varadarajan, and Fahy 1993; Day 1994). For example, a firm that can deliver the best product quality at the lowest cost with superior distribution and excellent promotion should have considerable advantages over its competition. Furthermore, firm strategies that are based on multiple sources of advantage are complex and, therefore, difficult for rivals to interpret. This is consistent with the resource-based perspective, which posits that superior performance is attained in conditions of uncertain imitability; that is, competitors are unable to discern, understand, or imitate the strategies of their competitors (Bharadwaj, Varadarajan, and Fahy 1993). However, organizational learning scholars point out that emphasizing core capabilities carries a danger, in that core capabilities can turn into "core rigidities" or "competency traps" (Leonard-Barton 1992; Sinkula 1994). In other words, emphasizing historical core capabilities reduces the incentive to develop new capabilities and restricts the breadth and depth of new learning. Not surprising, core rigidities are the source of lessened creativity in organizations (Leonard-Barton 1992; Moorman and Miner 1997). Thus, even when new strategies or capabilities might be appropriate, firms are likely to continue emphasizing historical and traditional marketing assets, capabilities, and strategies (Sinkula 1994; Slater and Narver 1995). Therefore: H6a: Emphasis on marketing assets and capabilities is associated negatively with strategy creativity. H6b: Emphasis on marketing assets and capabilities is associated negatively with organizational learning. H6c: Emphasis on marketing assets and capabilities is associated positively with market performance. Effects of cross-functional integration. Although the concept of cross-functional integration has been discussed at some length in prior research (Miller 1987), there has been less research on its effect on performance outcomes (Rajagopalan, Rasheed, and Datta 1993). Recent research in marketing has found that new product success is high when necessary and relevant skills, perspectives, and knowledge bases are cross-functionally integrated within the product strategy team (Ayers, Dahlstrom, and Skinner 1997). This research finding is consistent with several viewpoints expressed in our qualitative research. For example, one manager concluded that strategies were more likely to be successful when "interdisciplinary approaches and workgroups" are used in firms. When multiple functions and skills are well-represented on a strategy team, critical mistakes are detected at the source and corrected more quickly than would be possible under bureaucratic control or hierarchical directive (Olson, Walker, and Ruekert 1995). Therefore, we posit that when multiple functions are represented on the MSM team, the result will be enhanced market performance. In terms of the relationship between learning and crossfunctional integration, Slater and Narver (1995) conclude that people with different perspectives are necessary to avoid learning traps and enhance learning. Because crossfunctional integration brings personnel from multiple functions together, it facilitates the likelihood of learning through intellectual activities such as collaborative inquiry, collaborative diagnosis, and interparticipant knowledge relationships (Cicourel 1990; Sinkula 1994). Therefore, we hypothesize that the degree of cross-functional integration is related positively to learning. Organization theorists long have argued that people from different functional backgrounds frame problems differently, and therefore, integrating multiple functional members on a strategy team should 26 Journal of Marketing, April 1999 This content downloaded from 184.108.40.206 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
lead to more creative solutions (Tyre and von Hippel 1997; Woodman, Sawyer, and Griffin 1993). Cross-functional integration also facilitates creativity in MSM because it reduces a unidimensional perspective and opens the organization to novel and radical approaches to decision making (Andrews and Smith 1996; Stasch and Lanktree 1980). Griffin (1997) finds that multifunctional representation on a strategy team results in the generation of newer and more complex ideas. Therefore, we posit a positive relationship between cross-functional integration and creativity in MSM. H7a: Cross-functional integration is associated positively with strategy creativity. H7b: Cross-functional integration is associated positively with organizational learning. H7c: Cross-functional integration is associated positively with market performance. Effects of communication quality. Empirical research consistently has found a positive relationship between communication quality and market performance (Kohli and Jaworski 1990; Narver and Slater 1990). When team members communicate extensively and openly during implementation, they can make necessary adaptive changes quickly, and the overall coordination of the strategy is enhanced (Moorman and Miner 1998; Slater and Narver 1995). Research suggests that increased communication and interaction among strategy team members can lead to higher levels of mutual understanding, better rapport (Menon, Bharadwaj, and Howell 1996; Wooldridge and Floyd 1989), and more consistent market orientation within a firm (Jaworski and Kohli 1993), thereby leading to better market performance. Learning within an organization is a social process affected by the nature of interactions among members of the strategy team (March and Olson 1975; Tyre and von Hippel 1997). Increased contact and communication among team members can counter myopic thinking, reduce cognitive blinders, and promote open-minded inquiry (Day 1994). Nonaka (1994) contends that interaction among individuals and groups in an organization are fundamental to learning. Similarly, greater frequency and informality of communications within an organization have been associated with innovative behavior and novel solutions during decision making. For example, Moorman and Miner (1997) argue that new ideas and organizational creativity are functions of the extent of conversations among individuals in the firm. Taken together, these perspectives and findings suggest that an increased level of communication quality should be related positively to learning and strategy creativity. H8a: Communication quality is associated positively with strategy creativity. Hsb: Communication quality is associated positively with organizational learning. Hgc: Communication quality is associated positively with market performance. Effects of consensus commitment. Wooldridge and Floyd (1989) conclude that both shared understanding or consensus and commitment to the strategy should exist for the strategy to be successful. Managers provided many examples of strategy failures because of lack of agreement or consensus about the strategy from groups that included uncooperative middle management, unwilling sales force, and nonsupportive internal functions. A senior marketing executive of an electronic components firm bemoaned that "Although the overall strategy was sound, it was not universally agreed to ... [and] implementation was bad." Because consensus commitment is associated with willingness by individuals to exert effort for the strategy and with a sense of identification to the strategy objectives, we expect consensus commitment to the strategy to be related positively to market performance. Organizational learning is fostered when there is consensus about the content of the strategy and its implications for implementation (Slater and Narver 1995). Similarly, others suggest that learning within an organization is a function of its ability to share common understanding and interpretations of information and knowledge about a strategy (Fiol 1994). We hypothesize that the combined effect of consensus with the strategy, buy-in, and enthusiasm toward the strategy among the team members will be positive on learning. Although consensus commitment can encourage learning, it does not need to be novel or different. Research on innovation and creativity suggests that, in the process of reaching consensus commitment, novel and radical ideas are rejected in favor of orthodox and conventional ideas (Schweiger, Sandberg, and Ragan 1986). Therefore, we expect a negative relationship between consensus commitment and creativity in MSM. H9a: Consensus commitment is associated negatively with strategy creativity. H9b: Consensus commitment is associated positively with organizational learning. H9c: Consensus commitment is associated positively with market performance. Effects of resource commitment. A consistent finding of our qualitative research is that the importance of committing appropriate and adequate tangible and intangible (e.g., managerial time) resources to strategy is indisputable. Although there are similar sentiments expressed in the marketing literature (e.g., Day 1983), with the exception of research in the general planning literature (Ramanujam, Venkatraman, and Camillus 1986), there is little empirical research into the effect of resource commitment on firm performance. Therefore, we propose that allocating appropriate levels of resources will be related positively to market performance. Research on organizational slack suggests that the extent of resources committed will directly affect the extent of experimentation with new strategies and the pursuit of nontraditional and radically different strategies (Bourgeois 1981). Nohria and Gulati (1996) find an inverted relationship between innovation and organizational slack, which suggests that too little resource commitment is as bad as too much commitment. However, if and when appropriate types and adequate levels of resources are allocated, an organization not only provides a context in which strategy team members can do what is necessary for success, but also creates the potential climate for learning. Thus, resource commitment will be associated positively with learning and creativity. Marketing Strategy Making / 27 This content downloaded from 220.127.116.11 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
H iOa: Resource commitment is associated positively with strategy creativity. H lob: Resource commitment is associated positively with organizational learning. Hloc: Resource commitment is associated positively with market performance. Relationships Among Outcomes Although an emerging stream of research in marketing has studied some of the antecedents of creativity (Andrews and Smith 1996; Moorman and Miner 1997), the consequences of marketing creativity have not been examined. However, research on new product development has found that innovativeness of products is related positively to profit performance (Nagle and Holden 1995). Creativity involves developing newer and more radical alternatives and, thus, serves as a vehicle for new learning and potential for changes in organizational behavior (Amabile 1995). Similarly, creativity will affect market performance because it provides a mechanism for differentiation (Andrews and Smith 1996), which is an important route to market performance (Bharadwaj, Varadarajan, and Fahy 1993). Therefore, we hypothesize that marketing strategy creativity will be associated positively with organizational learning and market performance. Hlla: Strategy creativity is associated positively with organizational learning Hlb: Strategy creativity is associated positively with market performance. Controlling the effects of environmental turbulence. The external environment has been shown to play a major role in strategy making and performance (cf. Eisenhardt 1989), influence the nature and extent of learning in an organization (Menon and Varadarajan 1992; Sinkula 1994; Slater and Narver 1995), and affect product creativity (Moorman and Miner 1998). Because Rajagopalan, Rasheed, and Datta (1993) conclude that research in strategy should control for the environment, we incorporate environmental instability as the control variable. Testing the Model of MSM Stage 2, the empirical testing stage, included the following: (1) creating new scales or adapting extant scales to measure MSM and its antecedents and consequences, (2) pretesting and purifying the scales, (3) testing the model with a senior executive sample, and (4) conducting a post hoc test of common method bias. Data Collection A random sample of 1569 senior executives of SBUs of Fortune 1000 companies was drawn from a commercial mailing list. On the basis of the focus group and pretest results, the appropriate key informant was determined to be at a senior management level or have general management responsibility for products or markets. To achieve randomization and reduce bias, the respondents were requested to refer to the formulation and implementation of a recent marketing strategy for which performance data were available (cf. Menon, Bharadwaj, and Howell 1996; Sinha 1990). Using a critical incident approach, we explicitly directed the respondents in the cover letter and the instruction sheet with each survey to choose a marketing mix strategy. It is important to note that the choice of a marketing mix strategy as the critical incident defines and operationalizes the context of this study, "marketing strategy making."4 In addition, they were asked to choose a marketing strategy that (1) was not so recent that performance evaluation was premature and (2) not late enough that memory about the strategy and performance was fuzzy. We followed a four-stage modified "Total Design Method" to enhance respondent involvement and response rate. A prenotification letter was sent to each of the respondents, explaining the purpose of the research and the arrival of the survey and promising a summary of the results if the respondent attached a business card with the completed questionnaire. We mailed a copy of the questionnaire along with a cover letter and a return envelope a week later. A reminder postcard was sent the following week, and after two weeks, a replacement copy of the questionnaire was mailed. We contacted 100 randomly chosen respondents to determine nondeliverable and noncompliance rates. We determined that 38% of the mailings were nondeliverable (e.g., incorrect address, respondent no longer with the firm); an additional 16% did not reach the respondents (company gatekeepers apparently prevented the potential respondents from seeing the questionnaires); and 7% of the respondents reported a corporate policy of not responding to academic surveys. The total of 229 (212 usable) returned questionnaires represents a 37.8% response rate, which is quite satisfactory, given that average top management survey response rates are in the range of 15%-20% (Menon, Bharadwaj, and Howell 1996). Evaluating Quality of Data Collected Sixty-five percent of the respondents in the final sample had senior management titles, and the remainder was from middle management. The respondents belonged to manufacturing (74.8%) and service (25.2%) sectors. The SBUs represented in the sample had revenues of less than $10 million (7.2%), $10 million to $150 million (50.5%), and greater than $150 million (42.3%). Finally, the following marketing mix decisions were represented in the final sample: product strategy (75%), promotion strategy (15%), channel strategy (9.8%), and pricing strategy (.2%). Test of nonresponse bias. Tests of nonresponse bias were conducted and indicated that there were no significant differences between the early and late respondents on the means of three variables, namely, sales of the SBU, length of time in the industry, and positional level of the respondent. In addition, the covariance matrix of construct items was not different for the two groups, also suggesting that nonresponse bias was not a problem. Test of key informant competence. Three questions were used to verify the choice of the appropriate respondent. The first question assessed knowledge by using experience as a 4As elaborated in the next section, the respondents' choice of strategy for the critical incident is within the marketing mix domain. 28 Journal of Marketing, April 1999 This content downloaded from 18.104.22.168 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
proxy for knowledge (Phillips 1981). The second question assessed involvement in planning and implementation of reported strategy. The third question assessed importance of the project to the respondents and their organizations. On the basis of this assessment, ten respondents were dropped.5 Test of data poolability. When the appropriateness of key informants had been identified, responses needed to be tested for cross-sectional pooling of three categories of groups: organizational level, sales revenue, and industry sector. Using a test for poolability, we concluded that pooling the manager groups, sales levels, and industry sectors in the sample was appropriate and justified.6 An omnibus Ftest, to test if the marketing mix decisions could be pooled, indicated no differences across the three groups on the vari5After eliminating three respondents who had I year or less of experience in the firm, on average, respondents had 10 years of experience with the firm and approximately 17.2 years of industry experience, levels comparable to other samples of top management informants (Menon, Bharadwaj, and Howell 1996). Five respondents who were not directly involved in the strategy making but received reports were dropped from the final sample. Finally, we eliminated four respondents who scored two or below (1 = Not important; 5 = Very important) on importance of the strategy. Of these four respondents, two were also ineligible because they were not involved directly in the strategy. 6To test for the suitability of pooling respondents from the groups previously described, we examined if they responded differently on the constructs of interest to the study. The omnibus Ftest indicated no statistical difference between the two manager groups, providing justification for pooling respondents. To test if we could pool SBUs from the three different sales levels, SBUs with revenues of $10 million to $150 million a year were compared with SBUs with sales level greater than $150 million and were found to be not statistically different from each other on the crucial constructs. The SBUs with sales less than $10 million were too small to test statistically, but examination of the mean scores suggests that the mean values of the variables of interest are roughly similar to the means for the other two groups. Finally, the test for industry sector bias suggests that the manufacturing and service sectors were similar on mean values of these variables. ables of interest for this study.7 Thus, pooling across strategy types appears justified. Questionnaire When available, existing measures were adapted for our study context using input provided by managers in Stages 1 and 2. For constructs that did not have existing scales, measure development was guided by conceptual definitions advanced in Stage I and refined in Stage 2 (see Figure 1) of the study. Following a pretest of the draft questionnaire with 7 senior managers and 20 academic domain experts, the questionnaire was modified and administered to the full sample. Table 1 presents the descriptive characteristics of the scales, and the Appendix lists the items used. Antecedents constructs. The centralization four-item scale was developed to reflect Dewar and Werbel's (1979) definition of centralization as the concentration of decisions. The formalization three-item scale was adapted from Barclay's (1991) and Menon, Bharadwaj, and Howell's (1996) research to measure the degree to which the organization explicitly specifies the procedures and systems to accomplish tasks. The innovative culture seven-item scale captures the degree to which the firm emphasizes innovation, dynamism, openness, and change (Menon and Varadarajan 1992). The MSM components. The situation analysis four-item scale captures the extent to which the SBU systematically conducts a SWOT analysis while formulating strategy (Ramanujam, Venkatraman, and Camillus 1986). Emphasis on (13) marketing assets and capabilities was measured as a product of the extent of emphasis (five-point Likert-type scale) and whether they were historical and ongoing sources of advantage (ordinal scale 0, 1). Because both these constructs were measured using formative scales, calculating coefficient alphas for them is not appropriate. 7We conducted a test of differences across the first three strategy types because pricing strategy was too small to conduct a meaningful analysis. TABLE 1 Descriptive Statistics Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. Centralization 2.97 .60 .62 2. Formalization 2.72 .78 .32a .72 3. Innovative culture 3.46 .76 -.56a -.15a .89 4. Environmental turbulence 3.44 .64 -.02 -.03 .08 NA 5. Situational analysis 3.63 .79 -.11 .01 .20b .15b NA 6. Emphasis on marketing assets and capabilities 2.36 3.92 .05 .08 .21b -.08 .13 NA 7. Comprehensiveness 3.34 .81 -.11 -.02 .21b .07 .45a -.08 .62 8. Communication quality 3.82 .69 -.19b .04 .42a .08 .23a .02 .30a .83 9. Cross-functional integration 3.79 .76 -.15b .09 .34a .04 .40a .18b .49a .52a .77 10. Consensus commitment 3.66 .80 -.17b .11 .49a .01 .18b .24b .28a .54a .52a .77 11. Resource commitment 3.42 .81 -.15b .01 .44a .08 .10 .19b .14b .40a .44a .59a .80 12. Strategy creativity 3.68 .75 -.04 .04 .06 .07 .15 -.15b .29a .25a .13b .10 .10 .89 13. Organizational learning 3.71 .96 -.15a .06 .26b .11 .23b -.01 .22b .12 .17b .21b .11 .36a .77 14. Market performance 3.48 1.01 -.04 .01 .21b .06 .12 .12 .25b .21b .26b .30a .24a .21a .63a .84 aSignificant at p < .0001. bSignificant at p < .05. Note: Numbers on the diagonal are reliabilities. NA indicates a formative scale. Off-diagonal elements are Pearson correlation coefficients. S.D. refers to standard deviation Marketing Strategy Making / 29 This content downloaded from 22.214.171.124 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
Comprehensiveness was measured with a three-item scale that assessed the consistency and flexibility of alternative strategies before selection of a strategy (cf. Fredrickson 1986). The five-item cross-functional integration construct measured the extent to which the MSM team was well integrated, coordinated, organized, and functionally balanced (cf. Bourgeois and Eisenhardt 1988; Miller 1987; Ramanujam, Venkatraman, and Camillus 1986; Ruekert and Walker 1987). The four-item communication quality scale measured the extent to which there was continuous interaction and communication among implementation team members (cf. Menon, Bharadwaj, and Howell 1996). The three-item consensus commitment scale measures the extent to which there was consensus and buy-in about the strategy among MSM team members (Porter et al. 1974). The three-item resource commitment scale measured the extent to which both tangible (e.g., people) and intangible (e.g., managerial time) resources were made available to achieve the goals of the strategy (Porter et al. 1974; Ramanujam, Venkatraman, and Camillus 1986). Outcome measures. Creativity was measured with a five-item scale developed for this study by drawing on the literature and managerial input in Stages I and 2. A marketing strategy is creative if it deviates from previous strategies, is radical, and "breaks the rules of the game" (see Andrews and Smith 1996). The three-item scale to measure organizational learning, developed for this study, assessed the development of managerial skills on the strategy team, the understanding of the product market in which the firm competed, and the likelihood of procedural changes in the future as results of this marketing strategy (see Naman and Slevin 1993).8 Market performance was measured with a three-item scale that captured the extent to which the performance of the strategy met expectations for overall performance, sales, and profits (Menon, Bharadwaj, and Howell 1996). Although objective measures may have been more ideal, recent research points out that managerial assessments of financial and market performance are consistent with objective performance measures (cf. Hart and Banbury 1994; Naman and Slevin 1993). Finally, environmental turbulence was measured with a formative scale that captured the degree of change in seven components of a SBU's market environment. General Theory Testing Approach The hypotheses were tested through a series of ten seemingly unrelated regression (SUR) models.9 Precedent for 8Although we recognize that learning is multidimensional in nature (Sinkula 1994), the correlation between our scale and a global item of learning ("overall improvement in our organizational learning because of lessons learned from this strategy") is .75. 9To minimize the potential for Type I error that could arise from running ten separate regression, an omnibus test must be conducted. Following the recommendations of Dillon and Goldstein (1984) for the omnibus test, we conducted a canonical correlation analysis incorporating all dependent and independent variables. Because the result of the omnibus canonical correlation is significant (Wilks' lambda = .52, F = 4.48, p < .0001), we conclude that it is appropriate to conduct individual multiple regression analyses for each dependent variable. using a SUR approach in circumstances in which the model (see Figure 2) has multiple outcomes that are directionally (positively or negatively) but not causally related has been set in the literature. Following the recommendations of Lastovicka and Thamodaran (1991), factor scores rather than an unweighted linear composite of the item scores were used in the regression analysis. The resulting regressions are as follows: (1) MSMi = a + PIXl + P2X2 + 33X3 + e; i = 1 to 7; (2) Strategy creativity = a + P1MSMi + 32MSM2 + P3MSM3 + 04MSM4 + P5MSM5 + P6MSM6 + 37MSM7 + [8Environmental turbulence + ?; (3) Organizational learning = a + P1MSMI + 12MSM2 + P3MSM3 + P4MSM4 + P5MSM5 + 36MSM6 + 7MSM7 + P8Creativity + p9Environmental turbulence + e; and (4) Market performance = a + PiMSM1 + 032MSM2 + P3MSM3 + P4MSM4 + 35MSM5 + 16MSM6 + 37MSM7 + 38Creativity + 39Environmental turbulence + c, where MSMi = situational analysis, MSM2 = comprehensiveness, MSM3 = emphasis on marketing assets and capabilities, MSM4 = cross-functional integration, MSM5 = communication quality, MSM6 = consensus commitment, MSM7 = resource commitment, XI = centralization, X2 = formalization, X3 = innovative culture, and ? = an error term with 0 mean. Results Measurement analysis. After data collection, the measures first were subjected to maximum likelihood exploratory factor analysis (using oblique rotation) and confirmatory factor analysis. Because of the large number of constructs and measures employed in the study, we followed the recommendations in the literature on confirmatory factor analysis that scales should be assessed by examining smaller confirmatory factor models (cf. Bentler and Chou 1987), and therefore, measure analyses were based on groups of related sets of measures: (1) antecedents, (2) MSM, and (3) outcomes. This approach is well established in empirical research in marketing (cf. Moorman and Miner 1997). The results of the exploratory factor analysis indicate that the original assignment of 43 of the 50 items to the 11 reflective scales was appropriate. Similar to the exploratory factor analysis results, the overall fit of the initial confirmatory factor analysis with all 50 items was somewhat disappointing. Because the normalized residuals had a pattern similar to that of the cross-loadings in the exploratory factor analysis, and taking into account the need for statistical, practical, and conceptual fit, the 7 "problem" items were deleted. Subsequently, the fit was adequate [Antecedents: X2 = 165.38 with 82 degrees of freedom (d.f.) (p = .0001), goodness-of-fit index (GFI) = .91, Tucker-Lewis index (TLI) = .91, comparative fit index (CFI) = .93; MSM: x2 = 235.39 with 151 d.f. (p = .0002), GFI = .90, TLI = .93, CFI = .95; Outcomes: X2 = 80.95 with 51 d.f. (p = .0002), GFI = 30 / Journal of Marketing, April 1999 This content downloaded from 126.96.36.199 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
.93, TLI = .95, CFI = .96 ]. Furthermore, all the item loadings on the respective constructs were statistically significant (smallest t-value = 5.65, average t-value = 10.86). The reliability estimates for all the reflective constructs are greater than .60, with nine of eleven constructs greater than .70 and an average reliability of .77. The exploratory factor analysis indicated that the constructs were unidimensional (i.e., a single eigenvalue greater than 1). Discriminant validity was assessed by examining whether the confidence intervals (? two standard errors) around the correlation estimates between any two factors included 1.0. In none of the cases did the confidence intervals contain 1.0, thereby indicating that discriminant validity was upheld. In addition, two-factor confirmatory factor analysis of pairs of constructs was conducted twice: once constraining the correlation between the latent variables to unity and once freeing the parameter. A chi-square difference test then was used to test whether the chi-square value of the unconstrained model was significantly lower, in which case discriminant validity would be upheld. The critical value (AX2(i) > 3.85) indicates that discriminant validity was upheld in all 55 pairwise tests. Post hoc testing for common method bias. When dependent and independent variable data are collected from a single informant, common method bias can be a potential problem. Following Podsakoff and Organ (1986), we use the Harman's one-factor test to examine the extent of the bias. The result of the principal components factor analysis reveals that there are 12 factors with eigenvalues greater than 1.0, which accounts for 70% of the total variance. Because several factors were identified, because the first factor did not account for the majority of the variance (only 20%), and because there is no general factor in the unrotated factor structure, common method variance does not appear to be a problem (Podsakoff and Organ 1986). Regression analysis results. Because the model presented in Figure 2 has multiple outcomes, which are expected to be positively or negatively but not causally interrelated, SUR was used to test the hypotheses. The standardized results of the SUR procedure are reported in Tables 2 and 3. The individual results are summarized as follows: HI: As we predicted, centralization was associated positively with emphasis on marketing assets and capabilities (Hlc), and resource commitment (Hlg). The other hypothesized relationships between centralization and MSM were not supported. H2: As we predicted, formalization was associated positively with level of cross-functional integration (H2d), communication quality (H2e), and consensus commitment (H2f), providing partial support for H2. The other hypothesized relationships between formalization and MSM were not supported. H3: As we predicted, innovative culture was associated positively with all seven components of MSM (H3a-H3g), providing complete support for H3. H4: As we predicted, situational analysis was associated positively with organizational learning (H4b). Contrary to our a priori expectation, siati tional analysis was related negatively to market performance (H4c). No support was found for the relationship between situational analysis and strategy creativity (H4a). H5: As we hypothesized, comprehensiveness was associated positively with creativity (Hsa) and market performance (Hs5), providing general support for H5. H6: As we predicted, emphasis on marketing assets and capabilities was related negatively to creativity (H6a) and related positively to market performance (H6C). Thus, H6 is generally supported. H7: As we predicted, cross-functional integration was associated positively with strategy creativity (H7a). However, support was not present for the hypothesized relationships between cross-functional integration and organizational learning (H7b) and market performance (H7c). H8: As we predicted, communication quality was associated positively with strategy creativity (H8a). However, support was not present for the hypothesized relationships between communication quality and organizational learning (H8b) and market performance (H8c). H9: As we predicted, consensus commitment was associated positively with organizational learning (H9b). No support was found for H9a and H9c. HIo: As we predicted, resource commitment was associated positively with organizational learning (Hlob) and market performance (Hioc), providing general support for Hio. H11: As we predicted, strategy creativity was related positively to organizational learning (H Ila) and market performance (Hllb), providing complete support for HI1. Limitations There are several limitations that should be taken into consideration. Because of its cross-sectional design, causality cannot be established from this study. A longitudinal study is needed to establish causality. Data were collected by the key informant approach. Although some have pointed out that a study must capture the perceptions of multiple informants, especially in a self-report study (cf. Phillips 1981), others suggest that choosing the appropriate key informant helps alleviate some of the potential problems (Huber and Power 1985). Moreover, research has found that senior managers provide data as reliable and valid as multiple informants and objective data do (Tan and Litschert 1994; Zahra and Covin 1993). The evaluation of the quality of strategy was based on historical retrospectives by managers, which could be influenced or biased by performance and other factors (Curren, Folkes, and Steckel 1992; Golden 1992). However, recent research suggests that the danger of retrospective reports is overstated because these studies do not control for the errors in measurement (Miller, Cardinal, and Glick 1997). Although we provide a reasonable test of common method bias, it is at best post hoc and its results should not be interpreted unequivocally (Podsakoff and Organ 1986). Therefore, further research should attempt to obtain data on the dependent and independent variables from multiple sources, using multiple methods. Finally, it is important to note that, though we conceptualized comprehensiveness in MSM in terms of depth and breadth of evaluation of multiple strategies, our final operationalization did not measure depth. Because of the importance of comprehensiveness in MSM, future researchers are encouraged to expand the scale for comprehensiveness with additional items that tap into the depth dimension of comprehensiveness in MSM. Marketing Strategy Making / 31 This content downloaded from 188.8.131.52 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
C13 C cD co TABLE 2 Regression Results of Relationship Between Antecedents and Components of MSM Dependent Variables CrossIndependent Situational CompreMarketing Assets Functional Communication Consensus Variables Audit hensiveness and Capabilities Integration Quality Commitment Centralization .03 .02 .23** -.08 .02 .03 (.07) (.08) (.39) (.09) (.09) (.08) Formalization .02 .02 .06 .23** .12** .20** (.05) (.06) (.32) (.07) (.07) (.06) Innovative culture .29** .29** .36** .36** .47** .53** (.06) (.07) (.34) (.08) (.07) (.07) F value (p level) 4.69 5.09 6.04 13.83 16.07 23.37 (.0035) (.0021) (.0018) (.0001) (.0001) (.0001) R2 (R2 adj) .07 (.06) .08 (.06) .09 (.08) .19 (.18) .21 (.20) .28 (.27) *Standardized betas significant at less than .10 (directional, one-tailed test). **Standardized betas significant at less than .05 (directional, one-tailed test). Note: Standard errors are in parentheses. Resource Commitment .11* (.08) .05 (.07) .51 ** (.06) 15.64 (.0001) .21 (.20) This content downloaded from 184.108.40.206 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
TABLE 3 Regression Results of Relationship Between MSM and Outcomes Dependent Variables Strategy Organizational Market Independent Variables Creativity Learning Performance Situational analysis .10 .10* -.16** (.09) (.06) (.09) Comprehensiveness .17** .05 .19** (.08) (.07) (.08) Emphasis on marketing -.16** -.01 .10* assets and capabilities (.02) (.01) (.02) Cross-functional integration .14* -.05 .07 (.09) (.08) (.08) Communication quality .12** .04 .07 (.06) (.08) (.09) Consensus commitment -.11 .12* .07 (.09) (.07) (.08) Resource commitment -.01 .18** .27** (.08) (.07) (.08) Strategy creativity .38** .16** (.06) (.07) Environmental turbulence .05 .20** .17** (.10) (.09) (.09) F value (p level) 3.56 (.0006) 7.51 (.0001) 8.98 (.0001) R2 (R2 adj) .14 (.10) .28 (.25) .32 (.29) System weighted R2 .24 *Standardized betas significant at less than .10 (directional, one-tailed test). *Standardized betas significant at less than .05 (directional, one-tailed test). Note: Numbers in parentheses are standard errors. Guidelines for MSM and a Research Agenda In Stage 3, we triangulated the study results with additional insights from the literature and the field. Field-based views were drawn from 17 senior executives of the SPI Council for Market Strategy at a focus group-type workshop. These managers were used to provide practical insights on the study findings, find possible explanations for any counterintuitive results, and develop implications for best practices and additional research in MSM. Discussion By assessing the comprehensive effect of the organizational context on the multiple components of MSM in the same model, we can assess the differential effect of key organizational context variables more accurately. As is evident from the results, we find that, though formalization and centralization significantly influence some of the MSM components, innovative culture affects all of the MSM components identified in this study. Thus, our findings provide empirical support for Deshpande and Webster's (1989) assertion that organizational culture is central to managing the marketing function. Although the executives in the final stage of our qualitative research were receptive to the finding that innovative cultures have a broad-ranging impact, they also were sensitive to the difficulties in developing such cultures. One of them raised an interesting point, noting that managers were more likely to change organization structures, because it is easier and within their control, even though the impact of such changes is limited. Another called it a "short-term fix," but one that is more actionable. We find that centralization affects MSM positively by making it more likely that strategies are based on core capabilities and competencies and that appropriate levels of resources are allocated to execute them. Contrary to our expectations, centralization did not appear to be related positively or negatively to other MSM components. When the positive findings in this study are juxtaposed with negative findings identified in the extant literature, it suggests that the relationship between centralization and MSM is complex and remains worthy of further research. Consistent with our expectations, formalization facilitates MSM by improving cross-functional integration, communication quality, and consensus commitment, all of which are central to coordination and execution of a strategy. Thus, it appears that formalization enhances strategy implementation by enabling the smooth coordination and alignment of people, activities, and team members, ensuring appropriate functional and divisional representation on the team. Furthermore, MSM quality is enhanced not only when team members communicate quickly and frequently about the strategy during its execution, but also when they are committed to making it work. The managers raised the theme that "clear definition of responsibilities is more than a programming task. The clarity with which the instructions are communicated is as important as the actual allocation of the work load." Consistent with our findings, managers arMarketing Strategy Making /33 This content downloaded from 220.127.116.11 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
gued that poor definition "is a common critical failing. Overlap of responsibilities can allow everyone to drop the ball. Gaps can result from lack of clarity." The effects of the seven components of MSM vary in terms of their effect on the outcomes. Creativity appears to be enhanced by comprehensiveness, cross-functional integration, and communication quality and inhibited by emphasis on marketing assets and capabilities. Of these, managers cited communication quality as the most difficult to carry out because "managers lack critical communication skills ... they engage in discussion but not true dialogue ... communication is hard work ... confidentiality is the common excuse for not communicating." Others claimed that "extensive communication is avoided ... [because] managers lack confidence in the strategy, [or] fear measurement and failure ... [or because] ... communication draws out the critics; managers prefer not to rock the boat, to avoid conflict." With respect to cross-functional teams, several managers pointed out that "management by cross-functional teams is the in thing ... but, the truth is that many teams are really groups. It takes work and practice to make a group into a team." And, one other manager concluded, "the gap between team theory and group reality really hurts." Organizational learning appears to be facilitated by conducting situational analysis, consensus commitment, resource commitment, and marketing creativity. Market performance is enhanced by comprehensiveness, emphasis on marketing assets and capabilities, and resource commitment. With respect to emphasizing marketing assets and capabilities, several managers bemoaned the infrequency with which capability assessment is made part of a strategy process. One manager pointed out that "in evaluating acquisitions, my company always asks if it has the capabilities to run it. However, we seldom ask the same question about implementing a new strategy." The significant negative relationship between situational analysis and market performance needs further explication, as it was contrary to our expectations and struck some of the managers to whom we presented the findings in Stage 3 as surprising and counterintuitive. Others suggested that though situational analysis may be important in turbulent and changing markets, extensive situational analysis may be counterproductive in stable markets because it diverts valuable time and money resources from other potentially higher payoff activities. To test the managers' moderator hypothesis, we split the sample into high and low market turbulence groups and found empirical support for it.10 However, an alternative hypothesis for this negative finding could be that low-performing firms are more likely to conduct a situational analysis more extensively than firms that perform better.11 10We conducted a post hoc test by splitting the sample into high and low environmental turbulence groups to examine if the correlation between situational analysis and market performance was different in the two groups. We found the Pearson correlation between situational analysis and market performance in low turbulence environments to be -.20 (p < .05) and in high turbulence environments to be .33 (p < .05). I Unfortunately, we do not have data on the instrumental variables necessary to test the direction of causality implied in this alternative hypothesis, which was suggested by one of the reviewers. The mediating role of creativity in the effects of MSM on market performance and organizational learning is an important finding of this study.12 Although prior research has suggested that cross-functional integration and communication quality effect performance and learning directly, we show their effects on these outcomes to be completely mediated through creativity. In other words, without creativity, the positive effects of cross-functional integration and communication quality on learning and market performance disappear. Creativity also completely mediates the relationship between two MSM components-comprehensiveness and emphasis on marketing assets and capabilities-and organizational learning, but it only partially mediates the effects of comprehensiveness and emphasis on marketing assets and capabilities on market performance. Implications for Research and Theory in Marketing Strategy The Marketing Science Institute designated research on marketing organization and planning processes a priority topic for 1998-2000 (Marketing Science Institute 1998). This study has several implications for research and theory in this domain. It provides a comprehensive conceptualization for studying the process of MSM. By following a systematic, multistage, iterative research method to identify the seven critical components of MSM and testing it rigorously in a broad nomological net that includes its antecedents and consequences, we provide new theoretical insights for marketing strategy research and theory. The seven components explicated in this study also provide researchers in marketing strategy a useful conceptualization for studying process issues. By bringing together different schools of thought to conceptualize MSM, this research contributes to the integrative tradition in marketing strategy research (Kerin 1996) by providing a single hybrid framework for examining rational and incremental modes of strategy making with the formulation and implementation components of strategy making. Furthermore, by going beyond traditional market performance measures and including strategic, organizationwide impact variables such as learning and creativity, this study provides a richer and broader perspective on the effectiveness of MSM. Although the issue of organizational learning has garnered interest among marketing researchers (cf. Day 1994; Sinkula 1994; Slater and Narver 1995), empirical work focusing on the drivers of learning is lacking. This study contributes to the learning literature in marketing in several ways. First, we show that strategy creativity is one of the critical drivers of organizational learning. Second, we extend theory on the organizational structure and culture of learning by examining a new set of antecedents of learning, that is, the strategy making process. Third, we find a negative mediated relationship between an emphasis on core marketing assets and capabilities and learning, which provides suggestive support for the "competency trap" (Cohen and Levinthal 1990). 12We tested for mediation by creativity, following Venkatraman (1989). 34 / Journal of Marketing, April 1999 This content downloaded from 18.104.22.168 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
With the emergence of hypercompetitive markets and the rapid commoditization of products, differentiation through marketing creativity is a strategic imperative. This study advances the emerging literature base on marketing creativity in three ways. First, this is the first study that demonstrates empirically the direct and positive effect of strategy creativity on market performance and learning. Second, it contributes to marketing strategy theory by explicating the complex role that creativity plays in the effects of the MSM process. Specifically, the study shows that the effect of MSM on market performance and organizational learning is mediated by strategy creativity. Third, this study builds on the extant literature on marketing creativity (cf. Andrews and Smith 1996) by proposing and testing a new set of antecedents that drive strategy creativity. This study also adds to the emerging body of research on the resource-based theory of the firm. The resource-based literature contends that socially complex capabilities cause casual ambiguity and thereby become a potential source of sustainable competitive advantage. Social complexity refers to capabilities being dispersed throughout the organization and different elements of the capability residing in different groups or individuals. Individual elements that compose the capability may not have much value in and of themselves, but the combination of the disparate elements constitutes a firm-specific capability. Because MSM draws on multiple capabilities and different functions, it can be viewed as a socially complex capability. Furthermore, a capability is viewed as a higher order capability when it is composed of several lower order capabilities (Grant 1995). Because seven components together make up MSM, it is illustrative of higher order capabilities. This study also contributes to the marketing research methodology by demonstrating the value of the discoveryoriented approach for building MSMs; substantiating the benefits of triangulating literature and managerial views on the study findings; and illustrating, with a series of tests, an approach for imputing greater validity to a sample and data from biases such as informant fallibility, pooling, nonresponse, and common method variance. Implications for Best Practices in MSM This study has clear implications for MSM practice, the most obvious of which is that the effects of MSM on firm performance must be measured multidimensionally. It also points to the differential effect of the MSM components on market performance, learning, and creativity, demonstrating that factors that might not affect performance directly can have a long-term effect through creativity or learning. Instead of attributing more importance to strategy implementation than to strategy formulation, this article suggests studying MSM in a holistic and comprehensive manner and determining how and where MSM components can make a significant difference. Thus, a key contribution of this study is that it offers a framework for the manager on how to enhance the quality of the strategy process. It gives managers an a priori basis for focusing efforts on specific components of the whole process and enables post hoc analysis of prior strategies. The study offers empirical support for the conventional wisdom that resource commitment is a critical component of strategy success. We find that resource commitment affects not only a firm's market performance, but also the strategic viability of the firm by influencing the amount of organizational learning. Thus, allocating an appropriate amount of resources can enable the execution of the strategy as intended and, by signaling the importance of the strategy within the organization, can increase the likelihood of creating behavioral and cognitive learning. Because we find that comprehensiveness provides several strategic benefits to market performance, creativity, and learning, managers should exploit multiple tools and techniques to instill comprehensiveness in a firm's strategymaking exercise. Macro-level techniques and tools might include strategic optioning methods, such as scenarioand value-based planning, and, at the micro-level, dialectical inquiry and devil's advocacy techniques. Furthermore, it is critical for the organization to create systems that generate multiple, equally viable options, a framework that surfaces the contingent implications of these options, and optimization techniques to maximize performance by matching strategy to the market conditions. The importance of developing innovative cultures in which such techniques are adopted easily and used for strategy making is a clear implication of this research. Our evidence of the fundamental role of innovative cultures on the MSM process reiterates recent calls for focusing managerial attention on cultural issues. We provide strong support for creating an organizational climate that encourages people to take informed risks and share information openly within the organization, is open to change, and rewards entrepreneurial activities by employees. Directions for Further Research Innovative culture-we know it when we see it, we know now of its importance to MSM, but we know little about how to create it. Because current understanding of culture creation comes from anecdotes and case histories, there is ample opportunity for a comprehensive study on the antecedents and consequences of innovative marketing cultures. In our final focus group discussion, managers emphasized the importance of rewards to the development of innovative cultures, suggesting another avenue for additional research into the role of rewards on innovative cultures and MSM process. In particular, such research should examine the role that behavioral and outcome-based compensation mechanisms play on the choice of strategy and the process through which strategy is executed. Strategy creativity has a significant effect on firm performance and organizational learning. Although there is an emerging base of research on new product creativity (Moorman and Miner 1997) and product program creativity (Andrews and Smith 1996), further research should begin to synthesize and extend these research findings by developing an integrative framework of the organizational and individual drivers of marketing creativity. In addition, several of the findings on creativity suggest avenues for further research. First, our understanding of what drives marketing strategy creativity still is limited. Our model explains, at best, only 14% of the variance in marketing strategy creativity, which suggests that other variables, such as rewards, individual and team effects, and product/market characterisMarketing Strategy Making / 35 This content downloaded from 22.214.171.124 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
tics, need investigation. Second, consider the negative relationship between emphasis on marketing assets and capabilities and creativity. Although this finding was consistent with our competency trap hypothesis, it raises an interesting research issue: How do creative firms break this potential deleterious impact of extant competencies? Can use of creativity techniques overcome an organizational predisposition toward a competency trap or a core rigidity? Third, further research is needed to conceptualize the process of creativity in MSM and its effect on learning and market performance. The determination of the appropriate time to measure or evaluate outcomes of strategy is a perennial issue in strategy research. Researchers in advertising, research and development, and first mover advantages, for example, have raised the issue of when the actual impact of these activities should be measured and how long the impact persists. To our knowledge, this is an issue that remains a philosophical and empirical conundrum. Although we were sensitive to this issue, rather than specifying a period (say a year), we chose to let the respondent determine the appropriate time lag. It is possible that creativity has an immediate impact on learning and potentially a lagged and/or cumulative, longterm impact on learning. What this suggests is that, in all likelihood, we are capturing only a partial effect of creativity on learning and performance in this study. The notion of creativity's "wear-out" is an interesting concept and could be studied in an experimental context. Because our study investigates only the main effects of the MSM components on outcomes, additional research should examine more complex models of the antecedents and consequences of MSM. For example, research could explore the consequences of interactions among the components of MSM on strategy creativity, learning, and market performance. Because environmental turbulence demonstrated a significant association with learning and market performance and moderated the relationship between situational analysis and performance, further research should examine other moderators and controls, such as market characteristics, as well as other dimensions of the environment, such as environmental munificence and complexity. The nature of the strategy decision also may moderate the relationships implied in the MSM framework. For example, research could examine if the relationships vary depending on whether the strategy decision is a new product introduction versus a product line extension. Summary The recent BusinessWeek (1996) cover story, proclaiming that "Strategy Planning is Back," underscores a resurgence in practitioner interest in the process of strategy making. There is a corresponding rekindling in the marketing academic community of interest in how strategy decisions are made and implemented. This study was developed to provide insights into the process of MSM and contribute to the knowledge base in marketing theory and practice. To that end, this article (1) adopts a discovery-oriented approach to meld practitioner and academic perspectives on MSM; (2) builds, in an iterative and systematic manner, a hybrid conceptualization of the components of MSM; (3) develops and proposes a conceptual model that includes the antecedents and consequences of MSM; (4) empirically investigates the model using a sample of 202 marketing mix-related decisions from Fortune 1000 firms; (5) discusses the implications of the study results for marketing theory and practice; and (6) concludes with a set of directions for further research on MSM. APPENDIX Measures A. Centralization (Strongly Disagree/Strongly Agree) (Developed for this study) *In this division, decisions tend to be made at high levels. *Managers generally make decisions as they relate to their work without checking with anyone else (R). eThe individual decision maker has wider latitude in the choice of means to accomplish goals (R). *People are allowed flexibility in getting work done (R). B. Formalization (Strongly Agree/Strongly Disagree) (Adapted from Barclay 1991 and Menon, Bharadwaj, and Howell 1996) *Plans must be rigidly followed during implementation. *There is a "standard operating procedure" for almost all major decisions. *There are rules and procedures for most things. C. Innovative Organizational Culture (Strongly Agree/Strongly Disagree) (Based on Menon and Varadarajan 1992) *People in this division stress quick response to changing market conditions. *Our division's management style encourages a high level of participation. *Our division is dynamic and entrepreneurial. *Information is credibly and openly shared. *Our division emphasizes innovation and change. *There is a general feeling of trust and confidence between different groups. *People feel that their ideas and information are listened to by others. D. Situational Audit (Strongly Agree/Strongly Disagree) (Adapted from Ramanujam, Venkatraman, and Camillus 1986) During the development of this strategy, *The decision makers systematically considered organizational strengths. *The decision makers systematically considered organizational weaknesses. *The decision makers systematically considered environmental opportunities. *The decision makers systematically considered environmental threats. E. Comprehensiveness (Strongly Agree/Strongly Disagree) (Adapted from Fredrickson 1986) *Many alternative courses of action were explicitly considered before we chose this strategy. 361 Journal of Marketing, April 1999 This content downloaded from 126.96.36.199 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
*The chosen strategy was flexible and allowed for various contingencies. *Alternative strategies were adequately analyzed before they were dropped. F. Emphasis on Marketing Assets and Capabilities (Not Emphasized/Strong Emphasis; Historically, a major source of advantage for your company) (Developed for this study) In developing this strategy, to what degree did this strategy (a) emphasize the following and (b) is a major source of advantage for your company: *Pricing below competitors. *New products. *Broad range of products. *Extensive customer service capabilities. oBuilding brand image. *Developing and refining existing products. *Premium quality products and services. *Strong influence over channels of distribution. *Focus on specific geographic markets. *Promotion, advertising expenditures above industry average. *Products in higher priced market segments. *Products in lower priced market segments. *Innovation in marketing techniques. G. Cross-Functional Integration (Strongly Agree/Strongly Disagree) (Developed for this study) *The marketing unit responsible for implementation of this strategy was well integrated with the main business. *The members of the team implementing this strategy had the necessary skills and motivation to carry it out. *The implementation team was well organized. oThere was smooth coordination of the activities of group members during implementation. *The team responsible for the implementation of this strategy had adequate representation from other departments. H. Communication Quality (Strongly Agree/Strongly Disagre6e, (Adapted from Menon, Bharadwaj, and Howell 1996) oThe key players involved had continuous interaction during implementation of the strategy. *The strategy's objectives and goals were communicated clearly to involved and concerned parties. *Team members openly communicated while implementing this strategy. *There were extensive formal and informal communications during implementation. I. Consensus Commitment (Strongly Agree/Strongly Disagree) (Adapted from Porter et al. 1974) *All involved parties worked hard to make sure that the strategy was implemented successfully. .Consensus was evident during the implementation of this strategy. *All parties "bought in" or were "on board" with the strategy. J. Resource Commitment (Strongly Agree/Strongly Disagree) (Adapted from Ramanujam, Venkatraman, and Camillus 1986) *The right kinds of resources were allocated to the implementation efforts. *Every person was committed to make sure that they met their deadlines. *Adequate resources were allocated to the implementation efforts. K. Strategy Creativity (Strongly Agree/Strongly Disagree) (Developed for this study) *The chosen strategy was very different from others developed in the past in this division. *The strategy included some new aspects compared to previous strategies within this division. *The strategy broke some of the "rules of the game" within the product/market. *This strategy was innovative. *Compared to our previous, similar strategies, at least some parts were daring, risky, or bold. L. Organizational Learning (Very LowNery High) (Developed for this study) *Improvement in understanding about the market as a result of this strategy. eLikelihood of changes in the way we do things as a result of this strategy. .Development of managerial skills of the key players. M. Market Performance (Very LowNery High) (Adapted from Menon, Bharadwaj, and Howell 1996) *Overall strategy performance compared to expectations. *Net profits relative to expectations. *Growth in sales relative to expectations. N. Environmental Turbulence (No Change/ Very Frequent Change) (Developed for this study) Please indicate the level of change in your division's market environment over the past three years for the following: *Market opportunities. eProduction/process technology in your industry. *Products/services innovation in your industry. *Research and development (R&D) activity in your industry. *Legal and political constraints (e.g., government regulations). *Competitive intensity. *Customer preferences and expectations. Note: All scales are Likert-type. The anchors to the scales are within parentheses. REFERENCES Alba, Joel and Wesley Hutchinson (1987), "Dimensions of Consumer Expertise," Journal of Consumer Research, 13 (March), 411-54. Amabile, Teresa M. (1988), "A Model of Creativity and Innovation in Organizations," in Research in Organizational Behavior, Vol. 10, Barry M. Staw and L.L. Cummings, eds. Greenwich, CT: JAI Press, 123-67. (1995), "Discovering the Unknowable, Managing the Unmanageable," in Creative Actions in Organizations, Cameron Ford and Dennis Gioia, eds. Thousand Oaks, CA: Sage Publications, 77-82. Andrews, Jonlee and Daniel C. Smith (1996), "In Search of the Marketing Imagination: Factors Affecting the Creativity of Marketing Programs for Mature Products," Journal of Marketing Research, 33 (May), 174-87. Ansoff, Igor (1965), Corporate Strategy. New York: McGraw-Hill. Marketing Strategy Making / 37 This content downloaded from 188.8.131.52 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
Ayers, Doug, Robert Dahlstrom, and Steven J. Skinner (1997), "An Exploratory Investigation of Organizational Antecedents to New Product Success," Journal of Marketing Research, 34 (February), 107-16. Barclay, Donald (1991), "Interdepartmental Conflict in Organizational Buying: The Impact of the Organizational Context," Journal of Marketing Research, 28 (May), 145-59. Barney, Jay (1991), "Firm Resources and Sustained Competitive Advantage," Journal of Management, 17 (March), 99-120. (1997), Gaining and Sustaining Competitive Advantage. Reading, MA: Addison-Wesley. Bentler, Peter M. and Chih-Ping Chou (1987), "Practical Issues in Structural Modeling," Sociological Methods and Research, 16 (August), 78-117. Bharadwaj, Sundar G., P. Rajan Varadarajan, and John Fahy (1993), "Sustainable Competitive Advantage in Service Industries: A Conceptual Model and Research Propositions," Journal of Marketing, 57 (October), 83-99. Bonoma, Thomas V. (1985), The Marketing Edge: Making Strategies Work. New York: The Free Press. Bourgeois, L.J., III (1981), "On the Measurement of Organizational Slack," Academy of Management Review, 6 (January), 29-39. and Kathleen M. Eisenhardt (1988), "Strategic Decision Processes in High Velocity Environments: Four Cases in Microcomputer Industry," Management Science, 34 (July), 816-35. Bryson, John M. and Philip Bromiley (1993), "Critical Factors Affecting the Planning and Implementation of Major Products," Strategic Management Journal, 14 (July), 319-38. BusinessWeek (1996), "Strategic Planning," (August 26), 46+. Capon, Noel, John U. Farley, and Scott Hoenig (1990), "Determinants of Financial Performance: A Meta-Analysis," Management Science, 36 (October), 1143-59. , and James M. Hulbert (1994),"Strategic Planning and Financial Performance: More Evidence," Journal of Management Studies, 31 (January), 105-10. Chakravarthy, Balaji S. (1986), "Measuring Strategic Performance," Strategic Management Journal, 7 (September/October), 437-58. Cicourel, Aaron (1990), "The Integration of Distributed Knowledge in Collaborative Medical Diagnosis," in Intellectual Team Work: Social and Technological Foundations of Cooperative Work, J. Galagher, R. Kraut, and C. Edigo, eds. Hillsdale, NJ: Lawrence Erlbaum Associates, 221-42. Cohen, Wesley M. and Daniel A. Levinthal (1990), "Absorptive Capacity: A New Perspective on Learning and Innovation," Administrative Science Quarterly, 35 (September), 128-52. Curren, Mary, Valerie Folkes, and Joel Steckel (1992), "Explanations for Successful and Unsuccessful Marketing Decisions: The Decision Maker's Perspective," Journal of Marketing, 56 (April), 18-31. Day, George S. (1983), "Gaining Insights Through Strategy Analysis," Journal of Business Strategy, 4 (Summer), 51-58. (1986), Analysis for Strategic Market Decisions. St. Paul, MN: West Publishing Co. (1994), "The Capabilities of Market-Driven Organizations," Journal of Marketing, 58 (October), 37-52. Denison, Daniel R. and Aneil K. Mishra (1995), "Toward a Theory of Organizational Culture and Effectiveness," Organization Science, 6 (March/April), 204-23. Deshpande, Rohit, John U. Farley, and Frederick E. Webster Jr. (1993), "Corporate Culture, Customer Orientation, and Innovativeness in Japanese Firms: A Quadrad Analysis," Journal of Marketing, 57 (January), 23-37. and Frederick E. Webster Jr. (1989), "Organizational Culture and Marketing: Defining the Research Agenda," Journal of Marketing, 53 (January), 3-15. Dess, Gregory G. and Alex Miller (1993), Strategic Management. New York: McGraw-Hill. Dewar, Robert and James Werbel (1979), "Universalistic and Contingency Predictions of Employee Satisfaction and Conflict," Administrative Science Quarterly, 24 (September), 426-48. Dillon, William R. and Michael G. Goldstein (1984), Multivariate Analysis. New York: John Wiley & Sons. Eisenhardt, Kathleen M. (1989), "Making Fast Strategic Decisions in High-Velocity Environments," Academy of Management Journal, 32 (September), 543-76. (1990), "Speed and Strategic Choice: How Managers Accelerate Decision Making," California Management Review, 32 (Spring), 39-54. Fiol, Marlene C. (1994), "Consensus, Diversity, and Learning in Organizations," Organization Science, 5 (August), 403-20. Ford, Cameron (1995), "Creativity is a Mystery," in Creative Actions in Organizations, Cameron Ford and Dennis Gioia, eds. Thousand Oaks, CA: Sage Publications, 12-49. Fortune (1996), "Business As Wargame," (September 30), 190+. Fredrickson, J.W. (1983), "Strategic Process Research: Questions and Recommendations," Academy of Management Review, 8 (October), 565-75. (1986), "An Exploratory Approach to Measuring Perceptions of Strategic Decision Process Constructs," Strategic Management Journal, 7 (September/October), 473-83. and T.R. Mitchell (1984), "Strategic Decision Processes: Comprehensiveness and Performance in an Industry Within an Unstable Environment," Academy of Management Journal, 27 (June), 399-423. Gatignon, Hubert (1984), "Competition as a Moderator of the Effect of Advertising on Sales," Journal of Marketing Research, 21 (November), 387-98. Golden, Brian (1992), "The Past Is the Past-Or Is It? The Use of Retrospective Accounts as Indicators of Past," Academy of Management Journal, 35 (October), 848-61. Grant, Robert (1995), Contemporary Strategy Analysis. Cambridge, MA: Blackwell Publishers. Griffin, Abbie (1997), 'The Effect of Project and Process Characteristics on Product Development Cycle Time," Journal of Marketing Research, 34 (February), 24-35. Guth, William and Ian C. MacMillan (1986), "Strategy Implementation Versus Middle Management Self-Interest," Strategic Management Journal, 7 (July-August), 313-27. Hambrick, Donald C. and Phyllis A. Mason (1984), "Upper Echelons: The Organization as a Reflection of Its Top Managers," Academy of Management Review, 9 (April), 193-206. Hart, Stuart L. (1992), "An Integrative Framework for StrategyMaking Processes," Academy of Management Review, 17 (April), 327-51. and C. Banbury (1994), "How Strategy-Making Process Can Make a Difference," Strategy Management Journal, 15 (May), 251-70. Hofer, C. and Dan Schendel (1978), Strategy Formulation: Analytical Concepts. St. Paul, MN: West Publishing Co. Huber, George P. and R. Power (1985), "Retrospective Reports of Strategic-Level Managers: Guidelines for Increasing Their Accuracy," Strategic Management Journal, 6 (April-June), 171-80. Hutt, M.D., P.H. Reingen, and J.R. Ronchetto Jr. (1988), "Tracing Emergent Processes in Marketing Strategy Formulation," Journal of Marketing, 52 (January), 4-19. laquinto, A.L. and J.W. Fredrickson (1997), "Top Management Team Agreement About the Strategic Decision Process: A Test of Some of Its Determinants and Consequences," Strategic Management Journal, 18 (January/February), 63-76. Inkson, Kerr, David Hickson, and Derek Pugh (1968), "Administrative Reduction of Variance in Organization and Behavior," paper presented to British Psychological Society, London. 38 / Journal of Marketing, April 1999 This content downloaded from 184.108.40.206 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
Jaworski, Bernard J. and Ajay K. Kohli (1993), "Market Orientation: Antecedents and Consequences," Journal of Marketing, 57 (July), 53-70. John, George and John Martin (1984),"Effects of Organizational Structure of Marketing Planning on Credibility and Utilization of Plan Output," Journal of Marketing Research, 21 (May), 170-83. Kerin, Roger A. (1996), "In Pursuit of an Ideal: The Editorial and Literary History of the Journal of Marketing," Journal of Marketing, 60 (January), 1-13. , Vijay Mahajan, and P. Rajan Varadarajan (1990), Contemporary Perspectives on Strategic Market Planning. Needham Heights, MA: Allyn and Bacon. Khatri, Naresh (1996), "Further Evidence for Logical Incrementalism: A Study of Computer, Banking and Utility Industries in the United States," in Academy of Management Best Paper Proceedings. Cincinnati, OH: Academy of Management, 21-25. Kohli, Ajay K. and Bernard J. Jaworski (1990), "Market Orientation: The Construct, Research Propositions, and Managerial Implications," Journal of Marketing, 54 (April), 1-18. Leonard-Barton, Dorothy (1992), "Core Capabilities and Core Rigidities: A Paradox in Managing New Product Development," Strategic Management Journal, 13 (Summer), 111-25. Lincoln, Yvonne and Egon Guba (1985), Naturalistic Inquiry. Beverly Hills, CA: Sage Publications. Lyles, Marjorie A. and R.T. Lenz (1982), "Managing the Planning Process: A Field Study of the Human Side of Planning," Strategic Management Journal, 3 (January-February), 105-18. Mahajan, Vijay, P. Rajan Varadarajan, and Roger A. Kerin (1987), "Metamorphosis in Strategic Market Planning," in Contemporary Views on Market Practice, Gary L. Frazier and Jagdish N. Sheth, eds. Lexington, MA: Lexington Books, 67-110. March, James G. and J.P. Olson (1975), "The Uncertainty of the Past: Organizational Learning Under Ambiguity," European Journal of Political Research, 3, 147-71. Marketing Science Institute (1998), Research Priorities: A Guide to MSI Research Programs and Procedures, 1998-2000. Cambridge, MA: Marketing Science Institute. McKee, Daryl O., P. Rajan Varadarajan, and John Vassar (1990), "A Taxonomy of Marketing Planning Styles," Journal of the Academy of Marketing Science, 18 (Spring), 131-42. Menon, Anil, Sundar G. Bharadwaj, and Roy D. Howell (1996), "The Quality and Effectiveness of Marketing Strategy: Effect of Functional and Dysfunctional Conflict in Intraorganizational Relationships," Journal of Academy of Marketing Sciences, 24 (Fall), 299-313. and P. Rajan Varadarajan (1992), "A Model of Marketing Knowledge Use Within Firms," Journal of Marketing, 56 (October), 53-71. Michaels, Ronald, William Cron, Alan Dubinsky, and Erich Joachimsthaler (1988), "The Influence of Formalization on the Organizational Commitment and Work Alienation of Salespeople and Industrial Buyers," Journal of Marketing Research, 25 (November), 376-83. Miller, Chet C., Laura B. Cardinal, and William H. Glick (1997), "Retrospective Reports in Organizational Research: A Reexamination of Recent Evidence," Academy of Management Journal, 40 (February), 189-204. Miller, Danny (1987), "The Structural and Environmental Correlates of Business Strategy," Strategic Management Journal, 8 (January-February), 55-76. , Cornelia Droge, and Jean-Marie Toulouse (1988), "Strategic Process and Content as Mediators Between Organizational Context and Structure," Academy of Management Journal, 31 (September), 544-69. Mintzberg, Henry (1994), Rise and Fall of Strategic Planning. New York: The Free Press. and James B. Quinn (1996), The Strategy Process. Upper Saddle River, NJ: Prentice Hall. Moorman, Christine and Anne S. Miner (1997), "The Impact of Organizational Memory on New Product Performance and Creativity," Journal of Marketing Research, 34 (February), 91-106. and (1998), 'The Convergence of Planning and Execution: Improvisation in New Product Development," Journal of Marketing, 62 (July), 1-20. Nagle, Thomas T. and Reed K. Holden (1995), Strategy and Tactics of Pricing. Englewoods Cliffs, NJ: Prentice Hall. Naman, J.L. and D.P. Slevin (1993), "Entrepreneurship and the Concept of Fit: A Model and Empirical Tests," Strategic Management Journal, 14 (February), 137-54. Narver, John and Stanley Slater (1990), 'The Effect of Market Orientation on Business Profitability," Journal of Marketing, 54 (October), 20-35. Nohria, Nitin and Ranjay Gulati (1996), "Is Slack Good or Bad for Innovation?" Academy of Management Journal, 39 (October), 1245-64. Nonaka, Ikujiro (1994), "A Dynamic Theory of Organizational Knowledge Creation," Organization Science, 5 (February), 14-37. Nutt, Paul C. (1993), 'The Formulation Processes and Tactics Used in Organizational Decision Making," Organization Science, 4 (May), 226-51. Olson, Eric M., Orville C. Walker Jr., and Robert W. Ruekert (1995), "Organizing for Effective New Product Development: The Moderating Role of Product Innovativeness," Journal of Marketing, 59 (January), 48-62 Parasuraman, A., Valarie A. Zeithaml, and Leonard Berry (1985), "A Conceptual Model of Service Quality and Its Implications for Future Research," Journal of Marketing, 49 (Fall), 41-50. Phillips, Lynn W. (1981), "Assessing Measurement Error in Key Informant Reports: A Methodological Note on Organizational Analysis in Marketing," Journal of Marketing Research, 17 (November), 395-415. Podsakoff, Philip M. and Dennis Organ (1986), "Self-Reports in Organizational Research: Problems and Prospects," Journal of Management, 12 (Winter), 531-43. Porter, Lyman W., Richard M. Steers, Richard T. Mowday, and Paul V. Boulian (1974), "Organizational Commitment, Job Satisfaction, and Turnover Among Psychiatric Technicians," Journal of Applied Psychology, 59 (October), 603-609. Pugh, Derek, David Hickson, C. Hinnings, and C. Turner (1968), "Dimensions of Organizational Structure," Administrative Science Quarterly, 13 (June), 65-105. Rajagopalan, Nandini, Abdul M.A. Rasheed, and Deepak K. Datta (1993), "Strategic Decision Processes: Critical Review and Future Directions," Journal of Management, 19 (Summer), 349-84. Ramanujam, V., N. Venkatraman, and J.C. Camillus (1986), "Multi-Objective Assessment of Effectiveness of Strategic Planning: A Discriminant Analysis Approach," Academy of Management Journal, 29 (June), 347-72. Ruekert, Robert W. and Orville C. Walker Jr. (1987), "Marketing's Interaction with Other Functional Units: A Conceptual Framework and Empirical Evidence," Journal of Marketing, 51 (January), 1-19. Schweiger, David M., William R. Sandberg, and James W. Ragan (1986), "Group Approaches for Improving Strategic Decision Making: A Comparative Analysis of Dialectical Inquiry, Devil's Advocacy, and Consensus Approaches to Strategic Decision Making," Academy of Management Journal, 29 (March), 51-71. Shrivastava, Paul and John H. Grant (1985), "Empirically Derived Models of Strategic Decision Making Processes," Strategic Management Journal, 6 (April/June), 97-113. Marketing Strategy Making / 39 This content downloaded from 220.127.116.11 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
Sinha, Deepak K. (1990), "The Contributions of Formal Planning to Decisions," Strategic Management Journal, 11 (October), 479-92. Sinkula, James M. (1994), "Market Information Processing and Organizational Learning," Journal of Marketing, 58 (January), 35-45. Slater, Stanley F. and John C. Narver (1995), "Market Orientation and the Learning Organization," Journal of Marketing, 59 (July), 63-74. Stasch, Stanley F. and Patricia Lanktree (1980), "Can Your Marketing Planning Procedures Be Improved?" Journal of Market.ing, 44 (Summer), 79-90. Tan, Justin J. and R.J. Litschert (1994), "Environment-Strategy Re'ationship and Its Performance Implications: An Empirical Study of the Chinese Electronics Industry," Strategic Management Journal, 15 (January), 1-20. Thompson, Arthur A., A.J. Strickland, and William Fulmer (1987), Readings in Strategic Management. Plano, TX: Business Publications. Tyre, Marcie J. and Eric von Hippel (1997), "The Situated Nature of Adaptive Learning in Organizations," Organization Science, 8 (January/February), 71-83. Venkatraman, N. (1989), "The Concept of Fit in Strategy Research: Toward Verbal and Statistical Correspondence," Academy of Management Review, 14 (July), 423-44. Woodman, Richard, John Sawyer, and Ricky Griffin (1993), "Toward a Theory of Organizational Creativity," Academy of Management Review, 18 (April), 293-321. Wooldridge, Bill and Steven W. Floyd (1989), "Strategic Process Effects on Consensus," Strategic Management Journal, 10 (May/June), 295-302. Zahra, Shaker A. and Jeffrey G. Covin (1993), "Business Strategy, Technology Policy, and Firm Performance," Strategic Management Journal, 14 (September), 451-78. Zaltman, Gerald (1986), "Knowledge Utilization as Planned Social Change," in Knowledge Generation, Exchange, and Utilization, George Beal, Wimal Dissanayake, and Sumiye Konoshima, eds. Boulder, CO: The Westview Press, 433-62. Books from the American MarketinQ Association will keep you current. Dictionary of Marketing Terms Peter D. Bennett, editor Pennsylvania State University his completely revised and E expanded edition is an essential reference for business professionals and students alike. Fully ' . . cross-referenced for ease of use, this comprehensive resource lists more than 2500 up-to-date definitions of today's most important marketing terms. Covering both the day-to-day terminology and the specialized vocabulary in corporate and academic use, the Dictionary helps everyone from newcomers to senior-level marketing executives gain a more thorough understanding of critical marketing concepts. $32.95 AMA Members/$39.95 Nonmembers 1995. w AMA Marketing Encyclopedia Issues and Trends Shaping the Future Jeffrey Heilbrunn, editor ith an emphasis on recent and emerging issues, a panel of leading marketing : L l /i experts surveys the key areas of marketing, identifies current trends, and provides realistic prescriptions for the future. Contributors include industry leaders and topname academics, as well as pioneers and inventors of key marketing strategies and techniques: Tony Alessandra, John Hauser, Philip Crosby, Philip Kotler, Jack Trout, Al Ries, Don E. Schultz, Jagdish Sheth, and many others. The past, present, and future of marketing is shaped in this renowned book. $42 AMA Member/$4795 Nonmembers 1995. CALL TOLL-FREE 1-800-AMA1 150 TO PLACE YOUR ORDER. HAVE YOUR CREDIT CARD AND AMA MEMBERSHIP NUMBER HANDY. REFER TO JM499 WHEN YOU CALL. A E R I C A N M A R K E T I N G A SSOC I A T I O N 3 1 1 S . W ac ke r D ri ve, S u i te 5 8 00, C h i c a g o, IL 60606 .XcS5 e 0759< S~0W2S " ~ ,9 /'' " "~" " "~a0""a'~" 0 5"~'~ 2 40 / Journal of Marketing, April 1999 This content downloaded from 18.104.22.168 on Mon, 16 Jun 2014 17:24:41 PMAll use subject to JSTOR Terms and Conditions
THR EE ESSAYS ON MARKETI NG DECISION MAKING UNDER UNCERTAINTIES By CHENXI ZHOU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
Â© 2014 Chenxi Zhou
To my Mom
4 ACKNOWLEDGMENTS I would like to thank my committee members and other faculty members in the Department of Marketing at the University of Florida. Special thanks are to Professor Jinhong Xie and to Professor Barton Weitz for their invaluable guidance on the preparation of this dissertation and supporting the datasets. I thank all my friends who have supported me on this special journey and, most importantly, my parents, who gave me the strength and courage to overcome the obstacles I encountered along the way.
5 TABLE OF C ONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ..................... 9 C HAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 1.1 Buyer Uncertainty ................................ ................................ ............................. 11 1.2 Co untry level Uncertainty ................................ ................................ .................. 12 1.3 Product level Uncertainty ................................ ................................ .................. 13 2 ADVANCE SELLING INVESTMENT GOODS ................................ ........................ 15 2.1 Introductory Remarks ................................ ................................ ........................ 15 2.2 Basic Model: Pure Advance Selling Strategy ................................ .................... 25 2.2.1 Assumptions ................................ ................................ ............................ 26 2.2.2 Model ................................ ................................ ................................ ...... 30 2.2.3 Optimal Strategies ................................ ................................ ................... 32 2.2. 3.1 Impact of behavioral issues on the pricing strategy ....................... 32 22.214.171.124 Impact of refunds policy ................................ ................................ . 34 126.96.36.199 Impact of overconfidence ................................ ............................... 39 2.3 Model Extension: Mixed Advance Selling Strategy ................................ ........... 44 2.4 Summations ................................ ................................ ................................ ...... 52 2.4.1 Research Contribution ................................ ................................ ............. 53 2.4.2 Managerial Insights an d Future Research ................................ ............... 55 3 AN EMPIRICAL S TUDY OF THE COMPLETION OF C ROSS B ORDER MERGERS & ACQUISITIONS ................................ ................................ ................ 64 3.1 Introductory Remarks ................................ ................................ ........................ 64 3. 2 Theory and Hypothesis ................................ ................................ ..................... 69 3.2.1 M&A Procedure and Deal Completion ................................ ..................... 69 3.2.2. Country Level Factors ................................ ................................ ............ 72 3.2.3. Firm Level Factors ................................ ................................ .................. 78 3.3 Empirical Analysis ................................ ................................ ............................. 83 3.3.1 Model ................................ ................................ ................................ ....... 83 3.3.2 Data ................................ ................................ ................................ ......... 84 3.3.3 Variables ................................ ................................ ................................ . 85 3.3.4 Results ................................ ................................ ................................ .... 88
6 3.4 Summations ................................ ................................ ................................ ...... 95 3.4.1 Research Contributions ................................ ................................ ........... 96 3.4.2 Managerial Implications ................................ ................................ ........... 98 3.4.3 Limitations and Further Research ................................ ......................... 100 4 STOP VOLUNTARY DISCLOSURES? ................................ ................................ . 109 4.1 Introductory Remarks ................................ ................................ ...................... 109 4.2 Theory and Hypotheses ................................ ................................ .................. 113 4.2.1 Comparable Store Sales in Retailing Industry ................................ ....... 113 4.2.2 The Poor Performance Explanation ................................ ...................... 116 4.2.3 The Loss of Relevance Explanation ................................ ...................... 117 4.2.4 The Short Termism Explanation ................................ ............................ 120 4.2.5 The Information Spillover Explanation ................................ ................... 123 4.3 Empirical Analysis ................................ ................................ ........................... 125 4.3.1 Model ................................ ................................ ................................ ..... 125 4.3.2 Data ................................ ................................ ................................ ....... 126 4.3.3 Variables ................................ ................................ ............................... 128 4.3.4 Empirical Results ................................ ................................ ................... 131 4.4 Further Empirical Evidence ................................ ................................ ............. 132 4.4.1 Market Reaction to the Announcement of Stoppage ............................. 132 4.4.2 Drivers of Abnormal Returns ................................ ................................ . 135 4.4.3 The Moderating Role of Product Format ................................ ............... 137 4.5 Summations ................................ ................................ ................................ .... 138 4.5.1 Research Contributions ................................ ................................ ......... 139 4.5.2 Managerial Implications ................................ ................................ ......... 141 4.5.3 Limitations and Future Research ................................ ........................... 142 5 CONCLUSION ................................ ................................ ................................ ...... 150 LIST OF REFERENCES ................................ ................................ ............................. 152 BIOGRAPH ICAL SKETCH ................................ ................................ .......................... 161
7 LIST OF TABLES Table page 2 1 Industry Examples of Advance Selling the Investment Goods ........................... 58 2 2 Summary of Key Notations ................................ ................................ ................. 58 3 1 Inbound and Outbound M&As Distribution ................................ ....................... 102 3 2 Descriptive Statistics ................................ ................................ ........................ 102 3 3 Impact of Country and Firm Level Factors on Cross Border M&A Completion 103 3 4 Impact of Country and Firm Level Factors on Cross Border M&A Completion 104 3 5 Robustness of Results ................................ ................................ ...................... 106 3 6 Distribution of Cross border M&As by Developed Countries ............................ 107 3 7 Distribution of Cross border M&A by Industry ................................ ................... 107 3 8 Estimation Results Using Several Sub Samples ................................ .............. 108 4 1 Cited R easons for the M onthly CSS S toppage ................................ ................. 145 4 2 Distribution of Retailing Format According to SIC Code ................................ ... 145 4 3 Sample Selection Procedures ................................ ................................ .......... 146 4 4 Distribution of Stoppage Events across Time ................................ ................... 146 4 5 Determinants of Giving Up Monthly CSS Announcement ................................ . 147 4 6 Consequences o f Giving Up Monthly CSS: Stock Reactions towards Stoppage ................................ ................................ ................................ .......... 147 4 7 Drivers of Abnormal Returns ................................ ................................ ............ 148 4 8 Additional Tests: Determinants of Giving Up Monthly CSS in Sub Category ... 149
8 LIST OF FIGURES Figure page 2 1 Optimal Pricing under Pure Advance Selling with Refunds Strategy and Lemma 1 ................................ ................................ ................................ ............ 59 2 2 Optimal Pricing under Pure Advance Selling without Refunds Strategy and Lemma 2 ................................ ................................ ................................ ............ 60 2 3 Optimal Pricing under Pure Advance Selling with Refunds When and Lemma 3 ................................ ................................ ................................ ............ 60 2 4 Impact of Overconfidence for Pure Advance Selling with Refund Model and Proposition 3. ................................ ................................ ................................ ...... 61 2 5 Optimal Pricing under Mixed Advance Selling with Refunds Strategy and Lemma 4 ................................ ................................ ................................ ............ 62 2 6 Impact of Buyer Overconfidence under Mixed AS with Refunds and Proposition 4 ................................ ................................ ................................ ....... 63 3 1 M&A Procedure ................................ ................................ ................................ 105 3 2 Concep tual Framework ................................ ................................ .................... 105 4 1 An E xample of Monthly Same Store Sales Disclosure ................................ ..... 144
9 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THREE ESSAYS ON MARKETING DECISION MAKING UNDER UNCERTAINTIES By Chenxi Zhou A ugust 2014 Chair: Jinhong Xie Major: B usiness Administration This dissertation focuses on the opportunities and challenges faced by marketers in tailoring marketing activities to risk factors from several levels (e.g., consumer , firm , in dustry ,and country level). T hree issues of special interest are (1) How do the buyers behavior issues (e.g., present biased preference and over confidence) interact with the buyer uncertainty under a dvance selling strategy (C hapter 1) , (2) the impact of country level uncertainty and firm (C hapter 2) , and (3) how does the product level uncertainty affect the incentive for marketing communication strategy (e.g., the monthly same store sales announcement) in C hapter 3 . The thesis consists of three studies, all of which collectively investigate issues related to the marketing/other disciplines (e.g., psychology, finance) interface under the uncert ain environment present biased preference and overconfidence) affect the classic advance selling model and investigates how seller profit and consumer surplus change when incorporating behavioral iss ues in to the model . Chapter 3 s t udies the cross border M&A completion related to the emerging countries , the fastest growing market but with faster changing
10 environment s . While cross borde r mergers and acquisitions (M&A ) involving emerging markets have been increasing in recent years, a high percentage of them collapse before completion. This study investigates how the predictors of cross border M&A completion involving emerging markets depend upon the direction of glob al expansion, i.e., investment inbound to a developing market or investment outbound to a developed market. Chapter 4 studies what determines the retailers withdraw of monthly same store sales announcement, a disclosure ritual which has been adopted by re tailers for a long period. Moreover, how do the product characteristic s (e.g., fashion vs. staple) affect the incentive s towards the stopping decision .
11 CHAPTER 1 INTRODUCTION Uncertainties pose opportunities and challengers to the marketers. Risks stem from several sources , including capital market volatility (via beta) at macroeconomic levels (e.g., exchange rate and interest risk); or at the sector level (some industries by t heir nature are more or less stable), product market competition (competition can be stronger or weaker than anticipated), or product level outcomes (projects such as new product launches may fare better or worse than expected) . This dissertation covers th ree separate projects, all of which collectively investigate issues related to the marketing and other disciplines interface . Moreover, for each project, the decision make r has to deal with certain type s of uncertainties . 1.1 Buyer U ncertainty In Chapter 2 , I examine the research question on how seller shall advance sell for investment goods when buyers have time inconsistent preference and are overconfiden t about their abilities to be time consistent. Numerous social and laboratory studies have found a sys tematic pattern of the above two behavioral issues. After incorporating behavioral issues into the classic advance selling model, I found that the pricing structure has been greatly changed. Specifically, the seller can offer a lower level of refunds. Mor eover, a higher (lower) level of price can even go with a lower (higher) level of refunds. Second, I found a new function of refunds. From the buyers perspective, refunds can serve as a self control tool. From the sellers perspective, refunds can help th e seller to explore the decision bias from the buyers during the multiple decision process. Third, reducing buyers advance purchase risk by offering
12 refunds and reminding buyers tendency of being overconfident can both lead to a WIN WIN situation for fir m and consumers. This study contributes to the literature on buyer uncertainty ( e.g., Shugan and Xie ( 2000, 2005 ), Xie and Shguan ( 2001 ), Guo ( 2006) , Fay and Xie (2008), Png and Wang (2010) , Fay and Xie (2010 )) by showing how the firm should adjust its opt imal pricing structure when behavioral issues prevail and interplay with buyer uncertainty (e.g., uncertain about future consumption states) in the advance selling setting. 1.2 Country level Uncertainty In C hapter 3 , I exam the global marketing strategy related to emerging markets and focus on one type of expansion strategy, namely merger and acquisition. While cross border mergers and acquisitions (M&As) involving emerging markets have been increasing in recent years, a high percentage of them collapse be fore completion. This study empirically investigates how the predictors of cross border M&A completion involving emerging markets depend upon the direction of global expansion, i.e., investment inbound to a developing market or investment outbound to a de veloped market. Analysis based on 30 years of data from the two largest emerging markets, China and India, reveals fundamental differences in the predictors of inbound vs. outbound M&As completion. Inbound transactions are found to be most influenced by co untry level risk factors reflecting differences in political, trade, and legal environments. By contrast, outbound transactions are most influenced by firm level risk factors such as level factors, (e. g., percentage of stake sought by the acquirer, whether the deal size is disclosed, and whether the deal is a cash transaction), have significantly different effects on inbound and outbound M&As.
13 Th is study is closely related to two types of uncertaint ies : country level risk and firm level risk. For the macroeconomic risk, in the international marketing field, given the inherent risks of entering foreign markets such as information asymmetry (Kogut and Singh 1988) or liability of foreignness (Zaheer 1995), the global marketers will select different entry paths (e.g., greenfield, alliance, and merger and acquisitions) . In addition to country level risk, firm s idiosyncratic risk such as knowledge about foreign markets, size and financial strength will also affect the expansion outcome. My study contributes to the literature by highlighting the importance of focusing and managing specific type risks according to home country and host country characteristic s during the global expansion process. 1.3 Product level Uncertainty In Chapter 4 , I investigate the voluntary disclosure policy adopted by retailers and how this policy influences retailer performance. Retailers routinely disclose strategic information regarding their businesses, stores, and produ cts to analysts, investors, and the general public. Interestingly, a significant portion of such information is not mandated by regulations. Literatures in marketing, finance, and accounting propose several reasons why firms voluntarily disclose such infor mation. Unfortunately, empirical support is scant and cannot disentangle these hypotheses. My study contribute s to this literature by focusing on why retailers stop making one such disclosure on the monthly basis that helps distinguish organic growth and e xpansion growth comparative store sales (CSS). The CSS disclosure stoppage by some retailers and continuance by others provides a quasi experiment setting that helps empirically tease apart four alternate theories of information and strategy, not possibl e otherwise. Using a unique dataset collected over 17 years for 164 retailers, I find that the disclosure stoppage
14 to information postponement, loss of metric relevan ce, or long term focus. Moreover, in contrast with extant research, I provide evidence that investors respond favorably towards disclosure stoppage when they can rationally attribute it to reduced competitive spillover. My paper contributes to the stream o f marketing literature related to the new product preannouncement (e.g., Eliashberg and Robertson 1988, Robertson, Eliashberg and Rymon 1995, Lilly and Walters 1997, Sorescu et al. 2007). Previous results, largely based on the survey and laboratory data, a re subject to the response bias and potenti ally omitted variable biases. The monthly CSS announcement can be seen as firms preannounce overall operating ability in the established order prior to the required quarterly earning s announcement by newswires. Hence, by investigating the firms shifting from pre announcing to silenc e , this paper adds to the existing literature by offering cleaner datasets for the theory testing. In addition , we show that product level characteristic wi ll moderate the firm s incentive on the voluntary disclosure. Specifically, retailers with fashion products (e.g., apparel store, shoe store) are more likely to choose silence due to the competitive information spillover effect; however, retailers selling staple product (e.g., mass merchandiser, food store) choose to remain silence mainly because of poor performance.
15 CHAPTER 2 ADVANCE SELLING INVESTMENT GOODS This paper provides insights on how the seller should advance sell investment services suc decision process is influenced by the following behavioral issues: present biased preference and overconfidence. We found that behavioral factors greatly change the optimal pricing structure. S pecifically, the seller tends to offer a lower level of refunds; and a higher (lower) advance price might be related with a lower (higher) refund. Second, for the seller, offering a partial refund for service cancellation should not be always viewed as a c ostly option, but as a potential profit source. Offering such refunds allows the seller retain extra profits when neither consumer present biased preference is severe nor the unit service cost is low. Finally, helping consumers mitigate two potential risks in the advance selling setting: a) allowing them to regret by offering refunds, and b) educating them about the tendency of being overconfident during the multiple decision process could both lead to a win win sit u ation . 2.1 Introduct ory Remarks Technolog ical advancements have facilitated innovative selling strategies. The advance selling strategy, as one popular selling strategy within the industry, refers to a marketing practice in which the seller offers buyers opportunities to make purchase before the time of consumption. In the service industry, many providers often sell such services before their delivery date. For example, consumers often book the hotel, the air tickets, the access for theme parks weeks or months before their travel dates. Previous r esearches have shown that this strategy has created enormous profits for sellers (e.g., Xie and Shugan 2001, Prasad, Eteeke and Zhao 2009, Fay and Xie (2010), Nasiry and Popescu 2012). For instance, Xie and Shugan
16 (2001) and Shugan and Xie (2004) show that advance selling takes advantage of Zhao (2009) show that advance selling is a recommended method to reduce the commi tted. Fay and Xie (2010) show that, different from probabilistic selling strategy, advance selling via homogenizing heterogeneous consumers in fact addresses unobserved buyer heterogeneity by inducing sales involving buyer uncertainty. Naturally, the advan ce sell model itself is a multi stage decision process, in instance, during the advance purchase period, consumers have to decide whether to purchase according to the price off ered; during the consumption period, those buyers will decide whether to consume the product according to their current circumstances. It is such separation of consumption from purchase gives rise to the profit advantage of advance selling. However, this c ould also breed another important issue: buyers often have present biased preferences which might lead to time inconsistency preferences across purchase and consumption period, and buyers, during the advance period, are also likely to incur decision biases due to wrong forecasts about future preferences. For the former, the self control literature (e.g., Thalter and Shefrin 1981, Laibson 1997, O Donoghue and Rabin 1999, Fuddenberg and Levine 2006, Gilpatric 2009, Jain 2009, 2012, Machado and Sinha 2007) con sistently shows that consumers always place a higher discount rate between the present and the next period than between any of the subsequent periods. For the latter, consumers are shown to possess overconfidence about some positive personal attributes (e. g., Larwood and Whittaker 1977, Svenson 1981) and, hence, decision biases due to less precise predictions always occur. A growing
17 number of laboratory studies have documented such systematic and persistent consumer deviations from standard preferences and biases in the decision making process. This paper will mainly address the above issue in the context of consumers buying investment goods or services (e.g., conferences, health club training classes, the smoking clinics services) in advance. Compared to neutral goods, investment goods often cau se an immediate consumption cost and a delayed consumption benefit (DellaVigna and Malmendier 2004). For instance, going to the Yoga class will cause the immediate opportunity cost for the people who want to stay fit but have busy time schedule; Moreover, relative to the opportunity cost of consumption, the benefit from the class (e.g., the improvement in the health condition) will come in the future. This property of investment goods, together with the separation of purchase from consumption in the AS stra control issues due to conflicts of preferences across different decision periods. Here, we will revisit the classic advance selling model by adding an important behavioral layer decision procedure related to this specific type of products, and will provide answers to the following questions: first, in markets of investment goods, how do two well nonstandard preference (e.g., present biased preference) and irrational expectation selling setting? Second, what is the impact of offering refunds for cancellation in markets with above consumers? Third, how the fi control issues into the optimal design of advance selling strategy (e.g., advance price, spot price, and refunds)?
18 This research topic is particularly vital and urgent to industry practitioners because, on the demand side, consumers are increasingly becoming more health extra stress or just simply wanting to feel better, breed tremendous markets for the firms selling investment goods. Accor ding to the Organic Food Association, organic food sales increased 9.4 percent in 2011 from the year before, reaching nearly $30 billion in sales nationwide. Reports show that about two thirds of the U.S. population is striving to lose weight (Emmert 2011) . More importantly, with rapid development of information technology, the transaction costs related to advance sell and refunds processing have been greatly reduced, creating great opportunities for the company to implement this selling strategy. Through a stylized model, we found, in the market of advance selling the investment goods, the behavioral issues significantly change the optimal pricing structure: a) sellers tend to offer lower levels of refunds; b) a higher level of refunds might also be related with a lower advance price. We reveal that refunds offered by sellers for the service cancellation can be used by consumers as a pre commitment device to solve their own preference conflicts across purchase and consumption periods. When consumers purchase the investment services, the consumption cost is perceived higher during the consumption period than the advance period. Hence, they tend to have higher expectations to consume the product during the advance purchase period than they will actually do when the actual consumption date approaches. Realizing this tendency, consumers have preferences to accept a lower level of refunds in order to discourage the future walking away. Furthermore, since consumers are irrational (e.g., overconfidence), they might b e willing to accept a
19 higher price than that during the sophisticated situation, which causes the non positive relationship between refunds and the advance price. Then, we show that offering refunds for service cancellation should be viewed not as a costly option but as a potential profit driver for the seller in the advance selling setting. At the first glance, when consumers possess time inconsistent preference and are with a lower preference during the consumption period, refunds should help the buyers b y allowing them a chance to regret, thus hurting benefits of seller. However, we reveal that such refunds policy in fact helps the seller and can be the most profitable to seller if 1) present biased preference is weak, 2) the service cost is sufficiently high, and 3) consumers possess some deg ree of overconfidence. T his result first confirms previous findings in the literature (e.g., Xie and Shugan 2001) that refunds can help the seller when the marginal cost service is sufficiently high. In addition, we reveal that, due to buyer overconfidence, a medium level of refund induces buyers to incur decision biases and enables the seller to achieve two folded benefits together (e.g., also named as the refund manipulation effect): a higher advance selling price a nd the cost savings when providing service is expensive. The seller achieves the first benefit during the advance selling stage; due to overconfidence, advance buyers wrongly expect that they will consume the product even given an unfavorable consumption s tate, hence accepting a higher advance price charged. The second benefit arrives during the second period; when service cost is quite high, the same level of refunds could also benefit the seller by encouraging those unfavorable state buyers, to automatica lly give up services during only beneficial to the seller when buyers possess weak present biased preference. For markets with severe present biased preference, consumers f alling into
20 unfavorable states tend to self select to walk away during the consumption period even without the incentive of refunds; and the seller could still benefit from these adjusting the advance selling price. Third, compared with a non refund policy that locks customers into a purchase from the service provider, we reveal that refunds for cancellation might even lead to a win win situation if consumer behavioral issues exist. The above case occurs when the cha nce of the very bad state in the future is relatively high, present biased preference is very weak and the marginal service cost is sufficiently high. Recall that consumers with changing preferences are also compensated by the refunds, which allow they to regret once discover actual consumption states. More importantly, such refund compensation effect could dominate the above refund manipulating effect when the very bad future state has a higher chance to appear and present biased preference is very weak, h s increase as well when the service cost is sufficiently high. Moreover, we show that seller might not necessarily benefit from consumer overconfidence and buyers might not be hurt as well. Instead, cons limitation (e.g., overconfidence) during the multiple stage decision process might lead to totally four different scenarios (e.g., win loss, loss win, win win and loss loss), which depend on consumer characteristics (e.g., degree of overconfiden ce and present biased preference), seller characteristics (e.g., the unit service cost) and advance selling strategy (e.g., whether offering spot sales). Without offering spot sales, we find that overconfidence always benefits the seller and could either benefit or hurt buyers. Specifically, buyers could benefit from their own overconfidence when present biased preference is very weak and the marginal service cost is among the medium range. For this marginal cost range, for those
21 naÃ¯ve buyers, the seller h as incentives to offer a medium level of refunds; however, for the sophisticated buyers, the seller only offers a very low level of refunds (e.g., close to zero). Due to a higher refunds compensation effect, the buyers, although manipulated by the seller, them to regret and provide compensations, compared to the sophisticated case where a close to zero level of refunds is offered. After offering the spot sales option, we identify two additional cases (e.g., loss win and loss loss). In particular, the loss loss situation happens when present biased preference is relatively weak, overconfidence is very strong and unit service cost is among the medium range. Once seller adds in a spot sales option, strong overconfidence, together with weak present biased preference, could non proportionally increase the expected value of waiting option relative to the expected ir advance purchase decision. Wh en present biased preference is weak, overconfidence could influences ( e.g., determined by the actual present biased parameter) and the expected value of future consumpt ion ( e.g., determined by the predicted present biased parameter); however, when present biased preference is weak, overconfidence is no longer effective in boosting up the expected value of advance purchase since the buyers already make time consistent dec ision (and overconfidence has no influence ). Hence, after adding a spot sales option, overconfidence could directly force the seller to drop the price to encourage advance buying behavior of early arrivals , which decreases the seller profits. When the un it service cost is among the medium range, due to overconfidence and a not low service cost, the seller always has incentives to use
22 refunds to manipulate advance buyer s . However, in order to retain this part of benefits, the seller also has to reduce the AS price to encourage advance purchase. Compared to the case when buyers are sophisticated (e.g., zero overconfidence level), we could identify a medium service cost range (when the cost is not too high) that, the extra profits from shifting the commitment value from the buyer side might not be high enough to fully cover losses due to a drop in A S price. Hence, overconfidence might negatively influence both seller profits and buyers surpluses. This paper adds to the existing literature in several ways. Firs t, our paper significantly extends advance selling literature (e.g., Xie and Shugan 2001, Fay and Xie 2010, Nasiry and Popescu 2012, Prasad, Steeke and Zhao 2009) by incorporating two important but under explored behavioral factors: present biased preferen ce and overconfidence. Adding this additional behavioral dimension into the advance purchase process best captures the actual decision making process in daily life. The importance of understanding those behavioral factors in pricing strategy should never b e ignored when people make decisions under risks (Bell 1983, 1985). Specifically, we found that the above psychological factors could make the seller become more profitable. However, adding the spot sales option, although extending the sales market, could inconsistent behavior. As a result, our studies further complement the past advance selling literature on how seller reaps extra profits through exploring buyer uncertainties about future consumption st ates (Xie and Shugan 2001), obtaining probability rule of the occurrence of each consumption state (Huang and Chen, 2012).
23 Second, our paper also complements the research by (Xie and Shugan 2001, Xie and Gerstner 2007) on why the seller offers refunds in the advance selling setting. Compared with a no refunds policy, we show that, given some degree of o verconfidence, not severe present biased preference as well as a sufficiently high service cost, a well designed level of refunds could potentially benefit the advance sell in the following ways: Tempting buyers to advance purchase with higher expectations to consume the product; however, eventually, encouraging unfavorable state consumers to self select to cancel services during the consumption period. This profit mechanism offered by refunds is unique within the whole advance selling literature. For insta nce, Xie and Gerstner (2007) shows that, through refunds, the seller gets the opportunity to profit from the multiple selling in services with capacity constraints; Xie and Shugan (2001) suggests that, when buyer valuation for the service is state dependen t, offering refunds could be profitable if the cost of providing the service is sufficiently high. However, our work suggests that, by be time consistent, hence boosting the seller profit; this mechanism only increases seller profit when neither buyer present biased preference is severe nor the service cost is not sufficiently low. Furthermore, our work complements the recent research related to money back guarantee lite rature, where refunds have been widely studied as an effective marketing strategy (e.g., Chu et al. 1995, Fruchter and Gerstner 1999, Moorthy and Srinivasan 1995, Xie and Gersnter 2007). Our results point out that the refunds function might go beyond those identified in the literature. For instance, researchers have shown that offering refunds signals the higher quality of products for the seller (Moorthy and Srinivasan 1995), provides insurance against future dissatisfaction
24 (Fruchter and Gerstner 1999), r educes the opportunistic returns by consumers (Chu et al. 1995) and offers an opportunity to resell the service (Xie and Gernster 2007). New features of refunds are suggested by this paper: in the advance selling setting, commitment device to solve his own self control problem across two periods when perspective, refunds could help the seller explor e re brought by the multi stage decision process. Finally, we shed new insights on the impact brought upon by the information asymmetry between two parties. Most studies show that, for the interaction between two entities (e.g., Shugan and Xie 2000), one par favorable ground to the other party. For instance, Shugan and Xie (2000) suggest that the advance selling strategy benefits the seller because it overcomes the informational disadvantage of sellers in the spot period, during which sellers are Simi l arly their future preference directly benefit the seller? Different from conventional thoughts, our results reveal that, for advance selling of the investment goods, as the widens, the seller might not necessarily benefit from such information advantage. Instead, when offering spot sales option and when present biased preference in the ma rket is weak, very strong overconfidence might significantly increase the attractiveness of the waiting option for the buyer, which forces the seller to reduce the price to encourage advance purchase and causes a reduction in profits. Such finding directly implies that intentionally enlarging or over disadvantages (caused by psychological issues) might eventually hurt firm herself.
25 This contributes to a stream of literature discussing the merits of course related marketing to the firm development. This paper is organized as follows: Section 2 .2, 2.3 exams several advance profits an 2.4 discusses conclusions, managerial impl ications, and the future research. 2.2 Basic Model: Pure Advance Selling Strategy biased preference and overconfidence affect the pricing and profits under advance selling (AS) strategies. For AS, we first consider a situation where the seller has one period to sell: offering sales in the advance period and promising refunds for advance buyers to claim if they find out the consumption is costly in the future. We refer the period consumers are aware of their true consumption states as the consumption period. The period when early arrivals purchase the service but are uncertain about future consumption states is defined as the advance period. Many services offers refunds and charges the advance price well in advance of the service delivery date (see some examples in the listed websites 1 ). In section 2. 3, we then extend our model to a more general setting in which the seller has an opportunity to sell in both advance and consumption p eriods. We organize our current section as follows: Section 2. 2. 1 illustrates model assumptions; 2.2 .2 builds up the basic model for the pure AS with refunds strategy; 2. 2. 3 offers the model results. 1 For examples of advance selling strategy , please see the following web sites : a) http://www.gvhba.org/EventCalendar#id=526&wid=901&cid=120 ; b) http://acecma.org/index.cfm?pk=custom&ac=viewEventDetails&eventId=AC_CELTICS_030514; c)www.cnssa.org ; d) www.nfhs.org/WorkArea/DownloadAsset.aspx?id=9653 ; e) http://www.firpoint.org/spring retreat 2014.html ; f) ht tp://www.cvent.com/events/ckahu april 2014 meeting nahu members only/event summary c28c6ca2a552430a98f5871476d787b9.aspx ;
26 2. 2. 1 Assumptions In this section, we set up a simple th ree period model of consumer behavior and firm pricing strategy. This three period covers: the advance purchase period, the consumption period and the benefits harvesting period. Traditional AS models (e.g., Xie and Shugan 2001) usually assume a two period situation, where the consumers advance purchase during the first period and the consumption occurs in the spot period. However, for the investment goods, the benefit and cost from the consumption are usually separated (e.g., Dellavigna and Malmendier 2004 ), and benefits based on the consumption will come after the consumption cost. To capture this unique property, we add in a new period after the consumption period and refer it as the harvesting period, when the benefits from consuming the product arrive. Consider, for example, the advance purchase of the boxing classes offered by a local gym. During the advance period, early arrivals decide whether to advance purchase fitness training classes given the price announced by the seller and their expectations about future consumption utilities. For this period, early arrivals might be uncertain about their consumption states in the coming period. Although they all know that the training could bring about direct benefits (i.e., health benefits) in the future, so me people might have other activities (suddenly becoming extremely busy due to newly assigned works, sudden visits by old friends, etc.), which are taken as the opportunity cost for the training class. And they are not certain about those potential consump tion costs. Sudden assigned works might cause emotional anxieties, stresses or even time constraints, all of which will result into certain degree of cost from taking the class and, at the worst case, force advance buyers to cancel the service. This cost s tatus dependent assumption is natural and has appeared many times in the literature (e.g., Jeuland and Narasimhan 1985, Lu and
27 Moorthy 2007). During the second period (i.e., consumption period), after realizing their actual consumption cost (e.g., solving the uncertainty about consumption state), the early buyers will decide whether stick to the consumption or claim the refunds (walk away) based on their actual opportunity costs. After the consumption (if consumed), they will receive a delayed benefit durin g the third period. Consumer Behavior. We model the case in which the buyers purchase the investment products/service in the advance period. One characteristic of this type of service is that it will cause immediate consumption costs and delayed benefits. Examples include those products that are related to health care such as sport facilities, conferences, smoking clinics. Accordingly, we assume that a direct consumption cost will occur during the consumption period; and the benefits, from the consumption will arrive during the harvesting period. Second, we assume that early arrivals appear at the first period, and those buyers have uncertainties about future consumption states. Hence, three different consumption cost states: a low cost, a high cost and a very high cost might occur during the consumption period. The abov e cost scenarios also represent a favorable, an unfavorable and a very bad consumption state, respectively. Let and ; these three s tates happen with probabilities and , respectively. In this paper, we also refer those different buyers in the consumption period as high, low and very low segments, in terms of their evaluations for the service. In addition, we assume to ensure that the very bad consumption state is a non empty set and at least, some buyers will fall into this st ate. This captures a situation under which the
28 h that the losses fully cover the gains from the consumption. control issues, we assume that consumers have present biased preferences. This approach is widely used to model self control problems (e.g., Jain 2009, Gilpatr ic 2009, Dellavigna and Malmendier 2004). In particular, the discount function for time when evaluated at period equals 1 for and equals for with .The present value of a flow of future utilities as of time is: ( 2 1) We can interpret the parameter as the short run discounting (or present biased parameter) and as the parameter of long run discounting. For more information about the explanations about differences in discounting factors, one can consult Jain (2009). The standard time consistent exponential model corresponds to the case where is equal to 1. If is smaller than 1, the individual exhibits time varying discounting. The discount factor between the present period and the next period is while the discount factor between any two adjacent periods in the future is simply . The difference, between the short run and the long run discount factors generates time inconsistent preferences. Following most of the marketing literature (e.g., Jain 2009 and Gilpatric 2009) , we assume that equal to 1. We allow partially naÃ¯ve consumers to overestimate their consistency. A partially naÃ¯ve consumer expects (erroneously) the discount function (1): ith in all future periods. The buyer, therefore, correctly anticipates that he will have hyperbolic preferences in the future, but he overestimates the future
29 parameter of short run discounting if . The difference between the perceived and actual future short run discount factor reflects the overconfidence about the future self control. The larger the value is, consumers are more like ly to overestimate their abilit ies to b e time consistent. Some special cases deserve mention. An exponential consumer has no biased preference and is aware of it ( ). A sophisticated consumer has present biased preferences and is aware of it ( ) . and he charges the AS price To reduce the advance purchase risk, the seller also promises a refund to early buyers if t hey choose to walk away during the second period. During the second period, the seller incurs the marginal service cost when the buyer consumes the product. The seller designs an optimal pricing strategy to maximize the total p present biased preference and overconfidence issue. We summarize the timeline of the selling process as follows: at time 1, the seller sells the product when buyers are uncertain about the actual co nsumption cost. Buyers decide whether to advance buy the product for the advance price announced by the seller or just walk away. At time 2, when the advance buyer learns his cost type, he decides whether to consume the final product or claim the refunds. If he chooses to consume, he incurs a consumption cost at time 2 and the seller incurs the service cost. Otherwise, the buyer will claim the refunds, from which he gets some compensations and the seller does not incur any service cost. Given the consumptio n at time 2, consumers will receive a direct benefit from the product at time 3. In our paper, we implicitly restrict the present biased parameter to satisfy the
30 following condition: such that, at least, some advance buyers will f inally consume the products. We summarize all the notations for the model in Table 2.2 . 2.2 .2 Model W e consider the following marketing pricing strategy: pure AS with refunds. Under this strategy, the seller not only adopts advance selling strategy but also particularly interested in the following research questions: a) what is the optimal pricing strategy (i.e., the optimal advance sell price and refunds level) under this strat egy? b) How does behavioral factor (i.e., overconfidence) affect the seller profits and buyer surpluses? Whether this disadvantage (e.g., the naÃ¯vetÃ© about future preference) from the buyer side benefits the seller and in turn hurts the buyers as well? c) Should offering refunds for service cancellations to be viewed as a more costly option than the no changes across periods? Moreover, does a refund policy directly suggest that consumers will have highe r surpluses? Let be the amount of refunds to obtain in the second period. Let expected surplus from ASR is: ( 2 2 ). Here, represents an indicator function of and is a value set that might fall into. For instance, is an indicator function about the second . Specifically,
31 In other words, th maximization problem for the seller. At time 1, the seller determines the advance selling price and refunds , and he will maximize the following profit function (2 3 ) Overall, we report the impact of behavioral factors on pricing strategies for both types of buyers: sophisticated and partially naÃ¯ve buyers. The former captures the situation when buyers are sophisticated ( ) that they could correctly predict biased preferences. This assumption of sophistication entails rational expectations of the future behavior. In addition to this case, we also present the results for partially naÃ¯ve buyers under the situation when the consumers fail to correctly anticipate their future present biased preference s (i.e., ), which captures an often observed human tendency to put off unpleasant tasks today in the belief that tomorrow we will have the will power or self discipline to do them (Gilpatric 2009, Jain 2009). The optimal pricing st rategies for the above two scenarios are reported in Lemma 1. In addition to the pricing strat egy, we also report the seller profits and consumer surpluses 2 , given different levels of present biased preference and overconfidence. 2 T follow the method most commonly used in self control literature (e.g., Jain 2012, DellaVigna an d Malmendier 2004): C onsumer welfare in models with hyperbolic discounting is measured using preference s of consumer s with a long run perspective, i.e., . Another way this can be justified is by asking the discount factor one would use to advise consumers
32 2. 2. 3 Optimal Strategies S olving the objective function in (2) leads to the following lemma about the optimal pricing and profits when consumers advance purchase the investment goods. Let the notation in the following table denote a sufficient small numbe r that Lemma 1 shows optimal pricing strategies vary according to different ranges of present biased preference and overconfidence. In the subsequent section, Proposition 1 demonstrates the impact of behavioral issues on the opti mal pricing structure; Proposition 2 shows conditions when the behavioral issue (e.g., overconfidence) benefits the seller; and Proposition 3 examines the impact of the refunds option on the seller profits as well as consumer surpluses. 2.2 . 3. 1 Impact of behavioral issues on the pricing strategy Proposition 1 (price refund s tructure) . When advance selling investment goods, the two buyer behavioral biases, present biased preference and overconfidence, impact the optimal pricing and refund policy. Specifical ly, compared to the benchmark case ( ), the existence of buyer behavior biases ( ) 1. lowers the optimal level of refunds; 2. enlarges the set of the optimal Price Refund Structure: High Price/Low Refund is subop timal in the absence of buyer present biased presence but can be optimal in its presence. Proposition 1 illustrates a significant change of pricing structure of AS strategy for the investment goods, compared to results from the standard model incorporate the behavioral issues. A few remarks on the results are in order.
33 First, for any optimal strategy, we found that the optimal refund level tends to be lower than those identified when This finding is intriguing since it suggests that the buyers might be willing to accept a lower level of refunds in the advance sell setting. Roles of refunds have been widely discussed in money back literature (e.g., Xie and Shugan 2001, Xie and Gerstner 2007, Moorthy and Srinivasan 199 5, Davis et al. 1995). For instance, Xie and Shugan (2001) show that a refunds policy can serve as insurances to buyers against the future risk. Xie and Gerstner (2007) find op portunity for the multiple selling in the capacity constrained service. Different from the existing literature, our results suggest that the refund role might go beyond the above functions. New features of refunds are endowed in the advance selling setting commitment device to solve his own self control problem across two refunds could be strategically used, considering those consumer behavioral issues. Our results highlight that buyers could use the future refunds as a pre commitment device in the ASR setting and, therefore, have a preference for a lower of refunds. The intuitions are as follows: Due to time inconsistent preference, the future self is more likely to give up consumption than the expectation of the first stage self. In order to be consistent, the first stage self prefers a price and refund structure with a less future refund walking away. Moreover, by placing the refunds at a certain level, the seller could buyer might fail to foresee his fut ure refunds claiming activity and advance purchase with the expectation to consume the product in the second stage.
34 Second, when advance selling to consumers with the following characteristics ( ) , we observe several pairs of the optimal refunds and advance sell prices. Contrary to previous results identified under our results reveal the following interesting finding: behavioral issues cause the potential optimal pricing s advance sell price. For instance, due to different levels of overconfidence, we identify the following optimal price and refunds pairs: a) and b) and When is relatively small compared to we learn that a higher price might be paired with a lower level of refunds, as shown by b). On the contrary, a lower price might be in company with a higher level of refunds, as shown by a). 2. 2. 3.2 Impact of refunds policy The above proposition illustrat es how the behavioral issues significantly change the pricing structure in the advance selling setting. Then, how does the the refund effect, we first model the advance se lling strategy under a non refund policy (e.g., ) as follows. For this strategy, the advance buyer expected surpluses from advance purchasing degenerates into: ( 2 4 ) In the first stage , consumers will advance purchase only when We solve the maximization problem for the seller. At time 1, the seller determines the advance selling price and his profit function degenerates into: ( 2 5 )
35 Lemma 2 reports the pricing strategy of the seller under the non refunds strategy in the AS setting. As a special case, we first present our finding in Corollary 1, which examines the condition under which the r efund policy is advantageous when Corollary 1 ( impact of refund policy when ) . When present biased ), a refund policy benefits the seller (weakly) but profit when the service cost is not too low (i.e., ). Corollary 1 shows that, given no present biased preference, offering refunds increases the seller profits as long as the service cost is sufficiently high, and do not influence the buyers aggregated surpluses since surpluses transfer between buyers. The re sults collide with the findings in Xie and Shugan (2001). Then, what is result when it comes to the case that Intuitively, offering refunds should help buyers since they have time inconsistent preference (e.g., lower preference for the investment goods during the consumption period) and the refunds to some degree allow buyers to make up those decision mistakes they have previously made. Hence, the seller in turn will become less profitable. As a result, compared with the case whe n , we expect that offering refunds is less likely to increase the seller profits; or the marginal cost threshold for the extra benefit tends to be higher. Proposition 2 directly examines the condition under which the refund poli cy is advantageous when . Proposition 2 ( impact of refund policy ) . When advance Selling investment goods, in the presence of buyer present biased preference (i.e., ), a refund policy benefits the seller (weakly), but can either benefits or harms buyers (weakly). Specifically, offering refunds
36 Increase the seller profit when buyer present biased preference is not extremely strong and cost is not too low (i.e., and ). I ncrease the buyer surpluses when buyer present biased preference is weak and the unit cost is sufficiently high (i.e., and ); Decrease the buyer surpluses when buyer present biased prefer ence is among the medium range and the unit cost is sufficiently high (i.e., and ) No impact: neither the seller nor the consumers will be affected if buyer present biased preference is either too stron g or the unit cost is too low. Our results in Proposition 2 show whether and when the refunds policy is refund policy, a refunds policy could help the seller achieve more profits as long as the unit service cost is sufficiently high (i.e., ) and present biased preference is sufficiently weak (i.e., ). Furthermore, such refund policy is more effective when consumers possess certain degree of overconfidence, as shown by a lower unit cost requirement for the profit impro vement (i.e., when and ). profits from the refunds strategy. However, the effect of refunds depends on the degree of present biased preference. When present biased preference is severe, sufficiently strong overconfidence alone is enough to induce some buyers in the future to have time inconsistent preference towards the consumption of the product (e.g., the future unfav orable segments will always self select to cancel the service even without the refunds; however, the present self will fail to anticipate this behavior). Hence, the seller has no incentive to use the refunds to drive up the profits under this condition. Ho wever, when present biased preference is quite weak, under a no refunds policy, partially naÃ¯ve consumers will never self select to give up consumption (i.e., except those end ing up with very bad states) during the consumption period, and consumers always behave consistently for any level of
37 overconfidence. Hence, for this situation , the seller could not benefit from inconsistent behavior across different decision making periods. However, by introducing a refunds policy and setting the refunds level between the following range (i.e., ), the seller could again own a device, which induces advance buyers to have time inconsistent behavior across two decision periods. In other words, for weak present biased preference, the refund promised breaks up the when present biased preference is weak . This profit mechanism from refunds is be neficial only when the unit service cost is also sufficiently high. For a relatively low service cost, it is more beneficial to serve a large volume of consumers and therefore, the seller has no incentive to offer refunds whi ing away. Mathematically, the optimal refund level is set at When , and the profit difference between a refunds policy and a non refunds policy is: if It is easy to deduce that the higher the chance of ending up in the unfavorable state, the lower probability of profit advantage from offering refunds. Second, compared to a non refund policy, we found that offering refunds might even lead to a (WIN WIN) situation when buyers have a higher chance to walk away than to stay for consumption during the future non favo rable states (e.g., ), present biased preference is very weak ( ) and the unit service cost is sufficiently high, which suggests that offering refunds could be more economically efficient. When the unit se rvice cost is sufficiently high, the seller has an incentive to use refunds to induce time inconsistent behavior from the future low
38 segments; by doing this, he/she shifts the consumer added committed value occurred from the advance purchasing of investmen t goods to his/her own side. Hence, the amount of losses from buyer side is: which linearly decreases to zero as grows to 1. We refer the above negative impact of refunds on buyers as the refunds manipu lating effect. However, to capture such benefits, the seller has also to return more money to future very bad state segments (e.g., with a chance of to appear in the second period) than he has extracted through the AS price when buyers possess present biased preference. As a result, the amount of benefits the advance buyer gains from refunds are: which is referred as the refunds compensation effect. It is easy to deduce that the gain achieved from is an inverted U shaped function as grows. As a benchmark for the comparison, without offering refunds, the seller designs an optimal AS price finally serving both high and low segments (since present biased preference is very weak). Note that, for this case, the amount that buyers gain when falling into unfavorable future state is: . The buyer surpluses for buyers fall into very bad state are close to zero. The buyer surpluses from favorable future states remain the same for both cases. As a result, after offering refunds, the future unfavorable segments incur losses while the future very bad segments achieve gains; as an aggregate, those gains could dominate the losses when is su fficiently high relative to , and is suff iciently large. Hence, refunds might still benefit buyers, and our resu lts id entify conditions that refunds could lead to benefits for both parties . Third, compared with the refund role as , Proposition 2 demonstrates a striking difference in profitability mechanism from refunds. First, the marginal cost
39 threshold for the profitability might decrease (e.g., i f ), suggesting the refunds policy might help the seller reap more profits (compared to ) under the market condition where and . The reason behind is th at refunds help the seller realize two folded benefits altogether. Due to refunds, the value for the commitment when AS purchase the investment goods; h owever, when the refunds policy is helpful only when the service cost is high. Second, consumer advance buying risk directly influences the marginal cost threshold for the extra profitability. Specifically, the marginal cost thre shold decreases as increases and increases as increases. As a contrast, when profitability from the refund policy since is not influenced by advance purchase risk. Third, present biased preference also influences the extra profita bility derived from the refund policy. We have shown that when , the refun d policy could be more effectiv e due to a lowered cost threshold; however, when , the refund is no more effective in driving up pro fits no matter how large the unit service cost is. The result is important as it suggests that refunds benefit the seller by weakening biased preference is not severe. For severe present bia sed preference, the refund is no more effective since overconfidence alone could lead to time inconsistent behavior. 2. 2. 3.3 Impact of overconfidence Proposition 3 ( Impact of overconfidence ) . When advance selling investment goods, buyer overconfidence benefits the seller (weakly) but can either benefit or harm the buyers (weakly). Specifically, compared with the case where
40 consumers are fully aware of their present biased preference ( ), buyer overconfidence ( ) can lead to either a Win Loss or a Win Win, as summarized in the Figure 2 4 Previously , we have discuss ed the impact of refunds of seller profits and buyers surpluses. Then, how could ov under the advance selling setting? Intuitively, higher overconfidence implies that the buyer tends to be more likely to overestimate his/her ability to be time consistent. er could design a pricing structure to take in profits out overconfidence effect, we compare our re sults to the those under the benchmark case where consumers are sophisticated (e.g., ). Proposition 3 shows detailed discussions on this issue. The results presented in Proposition 3 show the impact of buyer overconfidence on seller profits as well as consumer surpluses. First, our results confirm the conventional belief that buyer overconfidence could enhance the seller profitability in the pric ing game. Our findings in Proposition 3 show when and how the seller could make more benefits from buyer overconfidence. Overall, our findings identify two profitability avenues for the seller to explore. The profitability paths could depend on the firm l evel character (e.g., the unit service cost) and consumer level characteristics (e.g., degree of overconfidence and present biased preference). For instance, when present biased preference is severe ( ), the seller can profit from just exploring strong buyer overconfidence ( ). Under this range of present biased preference, a higher level of overconfidence induces buyers to accept a higher price, during the advance
41 purchase period, compared to the situation when buyers are sophisticated. Note that both the future high segments (e.g., with a chance of ) and very bad segments (e.g., with a chance of ) will always make time consistent decisions across two decision periods; and the corresponding first stage self makes the correct expectation about the behavior of future self; however, strong overconfidence induces the future low state segments (with a chance of ) to have time inconsistent behavior. In terms of the pricing strategy, since future low segments (due to a low evaluation of products during the consumption period for severe present biased preference) will self select to cancel the servi ce for any refund level, the seller will just set the refunds level which is at the close to zero level and take advantage of buyers overconfidence issue by only raising the AS price. To be specific, during the consumption period, given such severe present biased preference and the low spot period evaluation for the consumption, buyers with unfavorable states will always choose to claim the refunds (or walk away) when realizing their actual consumption costs; however, sufficiently high overconfidence could lead those early arrivals to falsely expect the consumption of the product given the unfavorable states in the future. Due to such false expectation, the early arrivals accept a much higher AS price than they would actually do under the sophisticated case. Hence, given a high price charged, the buyer surpluses drop and the seller profits improve. Second, the profit advantage due to overconfidence still prevails when present biased preference is weak ( ); and overconfidence could help the seller achieve more profits as long as the marginal service cost is sufficiently high. Under this range of present biased preference, for a pricing strategy with a close to zero refund level, higher overconfidence no lo
42 outcomes since early buyers will always consume the product (e.g., except falling into a very bad state in the second period); and they will always have correct expectations about future decision no matter what degree of ov erconfidence they possess. As a result, the seller could no longer benefit from buyer overconfidence by just adjusting the AS price alone. However, through a properly designed refunds policy, together with a sufficiently high service cost ( ), the seller can again profit from the buyer overconfidence. Specifically, by setting the refunds level a little higher than the spot period evaluation once buyers have unfavorable states (e.g., ), the seller could again induce such time inconsistent behavior from buyers even under weak present biased preference range. Given the above refund level, during the advance period, the early arrival will expect to purchase product when he has either an unfavorable or favorable s tate in the future. However, the actual self during the consumption period will always self select to claim refunds once falling into the unfavorable states under the similar refund level. In other words, buyer overconfidence enables the seller to capture two folded benefits all together by offering a medium level of refunds: 1) the consumer added surpluses placed (higher preference for consumption ) by early buyers when purchasing the investment good in advance; 2) the cost savings during the consumption p eriod when some buyers finally automatically cancel the service. Therefore, the seller refunds and adjusting the AS price to induce AS purchase accordingly. Compared to the sophi sticated case, the seller could make more profits as long as the unit service is sufficiently high. Third, contrary to the conventional belief, we show that overconfidence might not necessarily bring harms to buyers. Instead, for certain type of buyer
43 char acteristics (e.g., consumers are more likely to walk away than to stay for consumption under non favorable situations, mathematically, ), we find that overconfidence might result into the gaining in consumer surpluses, which happens when present biased is very weak and the unit service cost is among the medium range. Recall under the above scenario, due to overconfidence, the seller use r efunds (for a not low marginal cost) to induce time inconsistent behavior from the future unfavorable segments. We refer the above negative impact of refunds on buyers as the refunds manipulating effect. To capture such benefits, the seller has also to co mpensate more money to future bad state segments. We have already shown that buyer surpluses would increase once the refunds compensation effect dominates the refunds manipulating effect, which happens when is sufficiently large relative to and present biased preference is very weak. As a benchmark for the comparison, when buyers are sophisticated, given the similar range of other parameters, the seller will accordingly set a close to zero level of refun ds as the optimal strategy and design the optimal AS price finally serving both high and low segments. Comparing both sophisticated and overconfidence cases, we find that overconfidence could lead to higher consumer surpluses only when the marginal service cost is among the medium range. The service cost cannot be too low, under this situation, the seller has no incentive to offer refunds when overconfidence exists; hence, seller profits and buyer surpluses remain the same. The service cost cannot be too hi gh, under this situation, for both situations, the seller will offer refunds; however, for naÃ¯ve buyers, the seller could charge a much high advance price. Hence, the seller profits increase while buyer surpluses reduce due to overconfidence.
44 To summarize, Proposition 3 first confirms the conventional belief that overconfidence could help the seller achieve more profits. Specifically, Compared to the case when consumers are sophisticated ( ), we find two conditions under which overconfidence could help the seller. When present biased preference is severe, the degree of overconfidence matters to this extra profitability; accordingly, the seller could make more profits as long as overconfidenc e is sufficiently strong. However, when present biased preference is weak, the unit service cost plays an important role on reaping extra benefits due to overconfidence; and the seller could benefit more when the unit service cost is sufficiently high. Sec consumer surpluses would decrease. Instead, we find a situation that is suff iciently large relative to , present biased preference is very weak and the service cost is among the medium range. 2.3 Model Extension: Mixed Advance Selling Strategy In the preceding section, we investigate the pure AS strategy with refunds strategy where the seller could make sales from advance selling period and offers the refunds during the consumption period to reduce the advance buying risks. In this section, we consider a more common observed pricing practice within the in dustry: mixed AS with refunds. Under this strategy, we assume that the seller not offers the spot sales to the late arrivals. In sum, the seller could make sales from both a dvance and consumption period. This model best captures the current industry practices and firm strategies. We are particularly interested in the following research questions: a) what is the
45 optimal pricing strategy (i.e., optimal spot price, advance sell price and refunds level) under this strategy? b) How adding a spot sales option affects the impact of For this strategy, in addition to assumptions made in the last section, we allow that there are also spot period arrivals, and the seller not only makes sales from early arrivals but also caters the demands from later arrivals. The seller will determine the following pricing strategies: the optimal advance selling price and the spot price, simultaneously . Furthermore, we assume that the seller announces the price strategy in the first stage, and she honors (Shugan and Xie 2000) the announced prices that the seller does not announce a price strategy in the first period and then changes to another one in th e next period 3 Let and advance selling with refunds and waiting, respectively. Hence, the expecte d surplus from advance purchase, , which is equal to ( 2 4 ). The expected surplus from waiting, , is ( 2 6 ) Moreover, the seller will determine advance sell price, spot sell price and refunds in order to maximize the following profit function: 3 Since the buyer might be worried about the further drop of the spot price during the consumption could further relax this assumpti on that the seller could use the service cost to endogenously build up the credibility for the announced spot price. For instance, for partially naÃ¯ve consumers, when , buyers will trust the high spot pric e announced, i.e., the higher the overconfidence, the higher the service cost lower bound for the high spot price announcement; when buyers will always trust the high spot price.
46 ( 2 7 ) We illustrate our analysis as follows: in Lemma 3, we first investigate the result s obtained under the common assumption that consumers do not possess present biased tendency ( ), where we identify three potential optimal strategies. The notations denote a cost range where each strategy dominates. As a contrast, in Lemma 4, we demonstrate potential optimal results by incorporating those behavioral issues into the model. In Proposition 4, by using results from the sophisticated case as a benchmark, we provide the impact of overcon fidence on the Through the comparison with the Proposition 2, we can find out how adding the spot well as consumer surpluses. Results in Lemma 3 are consistent with those in (Xie and Shugan 2001). For instance, when the service cost is among the medium range, i.e., , there are an infinite number of optimal refunds levels, whic h fall into a range of . Accordingly, there are an infinite pairs of optimal advance sell prices and refunds. However, these potential optimal pricing strategies have a single structure: a high refund is always related to a high a dvance sell price. Then, we report the case for sophisticated and partially naÃ¯ve consumers. The following Lemma demonstrates the optimal results for five potential selling strategies: For each strategy, we report the optimal pricing:
47 The notation denotes the cut off point for the marginal cost when the seller chooses between and . Results in Lemma 4 illustrate that there might be two potential optimal spot prices: a low spot price (e.g., ) and a high spot price (e.g., ) the seller worth noting another optimal scenario for the optimal spot price is among the following range (e.g., ), which presents that the seller charges a very high spot price and, eventually, only serves the early arrivals. Such optimal str ategy happens when the service cost is very high. Since we consider the mixed AS strategy, we only discuss the+ optimal solutions from which the seller could makes sales from the late arrivals. Proposition 4 examine the overconfidence effect when adding a spot sales option into the AS model. Note that, given a spot sales option, conditions under which the refund policy is advantageous remain the same as that shown in Proposition 2. P roposition 4 (impact of buyer o verconfidence under mixed AS with r efunds) For mixed AS with refunds strategy, consumer overconfidence ( ) can benefit or harm both the seller and buyers. Specifically, compared with the case where consumers are fully aware of their present bias ed preference ( ), buyer overconfidence ( ) can lead the follow situations, as summarized in the Figure 2 6 . Proposition 4 highlights some important results under the mixed AS with refunds strategy. First, after adding a spot sales option, our findings in Proposition 2 still hold. For the Win Loss and Win Win case, we find a consistent pattern of overconfidence e ffect identified in Proposition 2. However, different from Proposition 2, we find two additional cases brought by overconfidence: Loss Win and Loss Loss.
48 These results directly imply that under the mixed advance selling strategy, consumer overconfidence co uld become a double Overconfidence might not necessarily always boost the seller profits, as shown in Proposition 2. biased preference is not very severe, a higher level of consume r overconfidence might lead to a lower seller profit. As discussed in the last section, when present biased preference is not very severe ( ) and cost is not sufficiently high, the seller will choose to offer a close to zero level of refunds as the optimal solution. Then, please recall some unique properties under the mixed AS strategy. First, adding the spot sales will influence the decision process for early arrivals. When deciding whether to advance buy, early arrivals will also evaluate the expected utility of the waiting option as an alternative to the walking away; Second, early arrivals have uncertainties about future consumption states. When evaluating waiting options, early arrivals will also take those uncertainties into co nsideration. Third, the seller will design the pricing structure aiming to balance the sales from both markets: profits from advance selling period and from spot selling period. However, the determinants on the optimal pricing strategies for the two market s are different. To be specific, the optimal advance selling price could be affected by the predicted value of the present biased value of the present biased parameter, as shown by Lemma 4. Since the advance selling has no profit advantage for the low spot price situation, we focus our discussion on the high spot price case when or the unit service cost is not too low.
49 Hence, we could find that ov erconfidence tends to influence the expected value of advance purchase and the waiting option differently. Note that the first stage action is based on the previous of second stage self sticks to consum e or not). As a result, in terms of evaluating the expected value of advance purchasing (i.e., ), since advance buyers will expect whether stick to consume or claim the refunds (i.e., close to zero) in the second period, the expected value relies only on the predicted value of present biased preference, namely This directly implies that overconfidence might positively influen ce the expected value of AS (hence, the AS price); however, this impact is conditional on the actual value of present biased preference, When the actual present biased preference is not very severe ( ), o verconfidence no longer influences the expected utility of AS since the advance buyers always have the correct expectation about future consumption during this present biased parameter range. In contrast, for the expected value of the waiting option, overc onfidence will directly influence the expected value of waiting option since the advance buyer future consumption utility and but also the spot price. The former is influenc ed by and the later is driven by . Hence, a very strong overconfidence could increase the expected value of future expectation (Hence, decrease the AS price) for a not very severe present biased preference . To be specific, when the present biased preference is not very severe, , a higher level of overconfidence (i.e., ) will no longer impact however, it will directly increase the value of , which in turn reduces the
50 AS price and the profits of seller. Hence, under this case, we found that strong overconfidence could lead to a lower AS price, compared to the sophistica ted case. Therefore, we could identify (LOSE WIN) scenario, which happens when present biased preference is not very severe, overconfidence is very strong and the unit service cost is not sufficiently high. Under this situation, the seller will adopt the s trategy ASR_3, in which a close to zero refund level is offered. However, when consumers are sophisticated and all the other parameters remain the same, the seller will adopt the strategy ASR_1, where the seller also offers a close to zero refund but charg es a higher AS price. Hence, overconfidence leads to a loss from the seller but a gain from buyers. Third, interestingly, we find that confidence could lead to a (Lose/Lose) situation, which occurs when overconfidence is very strong ( ), present biased preference is severe, and the service cost of the seller is among the medium range ( ).This result is explained as follows. Given the above present biased preference and overconfidence, when the service cost i s not low ( ), the seller will adopt the strategy , in which he has an incentive to use a achieve more profits than in the strategy ASR_3 (when no refunds is offered). Given the same conditions for other parameters and zero overconfidence level, we could identify as the optimal strategy. Comparing these two strategies, consumers under might end up having fewer surpluses. This is because advance buyers have been fooled to make inconsistent decisions induced by the refunds promised. refunds manipul ation effect is no longer linear in , and it instead becomes a U -
51 shaped function as increases due to the increase in the expected value of waiting. Moreover, as discussed above, consumer might also benef it from such surpluses is an inverted U shape function of the parameter and it has the strongest positive impact when is among the medium range. Given this two forces, consumers suffer when present biased preference is either severe (yet not too severe such as ) or quite weak (but this case is unlikely to happen when the overconfidence is also very strong). As to the seller, under this case, although he has the motivation to use refunds to induce decision mistakes from advance buyers, too much overconfidence, m eanwhile, forces him to reduce the AS price to medium low range of service cost, where the seller always has an incentive to manipulate the naÃ¯ve buyers; however, the cost saving is not high enough to fully cover the losses caused by the drop in the AS price caused by very words, too much overconfidence level might give rise to an interesting lose/lose scenario under ASR strategy. To sum up, incorporating the spot sales option, we found the overconfidence effect is greatly influenced. For the seller, buyer overconfidence could become a double edged sword and very strong when present biased preference is not very severe. Specifically, when the unit service cost is low, very high overconfidence reduces the AS price and meanwhile, the seller has no incentive to offer a higher level of refunds. Hence, compared to the
52 due to a lower AS price charged. More importantly, we identify a loss loss situation hen present biased is severe (but not too severe), overconfidence is very strong and the unit service cost is among the medium range. A not too low service cost range induces the seller to use refunds to manipulate advance buyers and benefit from such time inconsistent behavior; however, when overconfidence is also very strong, the seller also has to drop the AS price to induce the advance buy due to the great increase in the expected value of waiting option under this situation. Compared to the sophisticat ed case, the cost savings from using refunds cannot fully compensate the losses in AS price drop when the service cost is not high enough. As a result, the seller profits could suffer for a medium range of service cost. Moreover, the buyer could also suffe r when present biased preference is severe, under which the refunds manipulation effect overconfidence might lead to a (lose lose) situation under mixed AS with refunds strategy . 2.4 Summations This paper provides insights on how the seller should advance sell the investment services, such as smoking clinics, gym classes and conference etc., present b psychological issues might facilitate the seller to become more profitable in AS setting, especially when present biased preference is severe, overconfidence is enough strong and the service cos t is sufficiently high; however, adding a spot sell option to the later arrivals during the advance sell weakens such profit advantage derived from early arrivals. Furthermore, offering a partial refund for service
53 cancellation should not be viewed as a co stly option, but as a potential profit source. Such refunds allow the service provider to capture some of the consumer added surplus (i.e., the commitment value) that is created when consumers purchase investment services in advance as well as to save the service cost from not serving the low margin segments during the consumption period. Last but not least, we found that too much overconfidence might lead to a LOSE LOSE situation for both parties. 2.4.1 Research C ontribution First, our paper contributes to the service marketing literature by exploring the benefit from advance selling strategy. In contrast to the past literature which has Xie and Shugan 2001) and behavioral issues such as anticipated regrets (e.g., Nasiry and Popescu 2012) and the anecdote reasoning (e.g., _ _2013), we are the first paper to draw attention on two important but under explored behavioral factors (e.g., time inconsistency and overconfidence) whe n consumers advance purchase the investment goods. The separation of purchase from consumption process provides favorable environments for the occurrence of the above behavioral factors. e extra profit sources in the advance selling setting. The seller could achieve more benefits if present biased preference is severe, overconfidence is strong and the service cost is sufficiently high. This result implies that, in addition to profit driver s identified in the ignored in the AS setting by the industry practitioners. Moreover, we also warn the managers that adding the spot sales option, although helping the se ller expand the sales market, threatens such profit advantage.
54 Second, this research contributes to the money back guarantee literature by offering another unique but important role for refunds. In AS of investment goods, realizing the preference conflicts within advance purchase and consumption periods, individuals always have a preference for commitments in the decision making process. A partial refunds promised by the seller could be used as one of important pre lf control. Hence, the role of refunds might go beyond from those identified in the literature. For instance, refunds could offer an insurance against the future dissatisfaction and reduce current buying risks (Fruchter and Gerstner 1999); refunds effectiv product quality (Moorthy and Srinivasan 1995); refunds policy are viewed as more customers friendly than other means (i.e., overbooking) and offers opportunities for the multiple selling (Xie and Gersnter 2007). Diff erent from previous findings, our research suggests that buyers might have a tendency to accept a lower level of refunds given certain product characteristics and marketing strategies. More importantly, a refund policy could be better than a non refunds po licy in generating profits. The seller benefits when buyers have weak present biased preference, display some degree of overconfidence and the service cost from the seller is sufficiently high. (e.g., caused by psychological issues) might not necessarily benefit the seller. biased preferenc e is weak, very strong overconfidence could negatively influence profits achieved from AS. After adding spot sales, the seller has to balance the profits from both markets (periods); and advance buyers have another alternative (i.e., waiting) to evaluate w
55 overconfidence will positively influence the expected value for both AS purchase and waiting option. However, for the former, such positive impact is conditional on the present biased parameter. In other words, wh en present biased preference is weak, the buyer will always make time consistent decision (e.g., the product consumption) and, hence overconfidence no longer induces advance buyers have wrong expectations (i.e., overestimation) about future consumption act ivity, no matter what degree of overconfidence buyers possess. However, for the later, the positive impact might persist, especially when the opportunity cost difference between different consumption states is small relative to the consumption benefit. As a result, the waiting option, being a substitution to the advance purchase, could greatly reduce the attractiveness of advance selling and, therefore, force the seller to reduce AS price. 2.4.2 Managerial Insights and Future Research Due to vast market pot entials and significant social impacts, it is critical for the service providers to understand how they can succeed from advance selling the managerial implications. First, our findings remind managers of those untapped profit sources brought achieved could be much higher when the targeted consumers possess severe present biased preference, strong o is not sufficiently low. However, when deploying the AS strategy, the seller should also be cautious about adding the spot sales option in the selling process. Although the market expands significantly, the p rofits obtained from selling to early arrivals could be cannibalized if the seller tries to satisfy the demand from later arrivals.
56 Second, our results also warn managers that over interests as well. This situation usually occurs in the market when consumers display not very severe present biased preference, and possess very strong overconfidence; meanwhile, when the seller promises a refunds policy and her service cost is among the medium range. Hence, our results point out an interesting marketing implication: designing advertising slogans (or contents) with a purpose to mitigate the degree of naÃ¯vetÃ© from buyers. Such kind deed not only educates advanc e buyers about their inclinations on overconfidence (or subjectivity) but also helps the seller herself out to achieve more profits. Last but not the least, our paper raises caution to consumers to be alert to a higher level of refunds promised by firms. A lthough generous refunds appear to be friendly to consumers and could increase temporary buying satisfactions, such refunds could eventually become a sweet killer. For instance, we reveal that a medium level of refunds could bring harms to buyers in the ma rket of investment goods/services, where consumers possess some degree overconfidence and not very severe present biased preference. Several aspects of the advance selling strategy invite future research. First, how should the seller advance sell when deal ing with the vice goods? Similar to the investment goods, the vice goods, such as services from casinos and night clubs, will also cause the buyer s self control problems. Second, we only consider a monopoly case in this paper. Investigating a market with multiple sellers would also add significant work to this research. Will the competition effect influence all the conclusions in our paper? Third, since service providers sometimes have capacity
57 constraints issues, we could investigate how the limits on the production power affect the optimal pricing strategy and the final profits.
58 Table 2 1 . Industry Examples of Advance Selling the Investment Goods Online link Table 2 2 . Summary of Key Notations
59 Figure 2 1 . Optimal Pricing under Pure Advance Selling with Refunds Strategy and Lemma 1
60 Figure 2 2 . Optimal Pricing under Pure Advance Selling without Refunds Strategy and Lemma 2 Figure 2 3 . Optimal Pricing under Pure Advance Selling with Refunds When and Lemma 3
61 Proposition 3. When advance selling investment goods, buyer overconfidence benefits the seller (weakly) but can either benefit or harm the buyers (weakly). Specifically, compared with the case where consumers are fully aware of their present biased preference ( ), buyer overconfidence ( ) can lead to either a Win Loss or a Win Win, as summarized in the Figure 2 4 : Figure 2 4 . Impact of Overconfidence for Pure Advance Selling with Refund M odel and Proposition 3.
62 Figure 2 5 . Optimal P ricing under Mixed Advance Selling with Refunds S trategy and Lemma 4
63 Proposition 4 (Impact of Buyer Overconfidence under Mixed AS with Refunds) . For mixed AS with refunds strategy, consumer overconfidence ( ) can benefit or harm both the seller and buyers. Specifically, compared with the case where consumers are fully aware of their present biased preference ( ), buyer overconfidence ( ) can lead the follow situations, as summarized in the Figure 2 6 : a) b) Figure 2 6 . Impact of Buyer Overconfidence under M ixed AS with R efunds and P roposition 4
64 CHAPTER 3 G LOBAL EXPANSIONS O ERSUS ROM MERGING MARKETS : AN EMPIRICAL S TUDY OF THE C OMPLETION OF C ROSS B ORDER MERGERS & ACQUISITIONS While cross border mergers and acquisitions (M&As) involving emerging markets have been increasing in recent years, a high percentage of them collapse before completion. This study investigates how the predictors of cross border M&A completion involving emerging markets depend upon the direction of global expansion, i.e., investment inbound to a develo ping market or investment outbound to a developed market. Analysis based on 30 years of data from the two largest emerging markets, China and India, reveals fundamental differences in the predictors of inbound vs. outbound M&As completion. Inbound transact ions are found to be most influenced by country level risk factors reflecting differences in political, trade, and legal environments. By contrast, outbound transactions are most influenced by firm level risk ence. F urthermore, deal level factors (e.g., percentage of stake sought by the acquirer, whether the deal size is disclosed, and whether the deal is a cash transaction), have significantly different effects on inbound and outbound M&As. These findings have important managerial implications for enhancing the success of global expansions. 3 .1 Introduct ory Remarks While rapid economic growth has made emerging markets major battlefields for global expansion from developed economies, a growing number of firms f rom emerging markets are going overseas. In the last decade, global expansion by emerging economies has almost caught up to the speed of developed economies in emerging markets. For example, according to the Thomson Securities Data Corporation's (SDC) Merg ers and Acquisitions Database, the global expansions by firms in emerging economies (e.g., BRIC) have increased by 8% in the last decade,
65 a growth rate that approaches that of expansions by firms in developed economies in the same period. Some well known e xpansions by firms in emerging economies computer business in 2005; n 2008. Among alternative entry modes for global expansion such as joint venture and greenfield, cross border M&As have become a widely adopted entry mode for such as the k nowledge base, technology, and human resources (Shimizu et al. 2004). While much research has been conducted on cross border M&As, many studies have focused on either the selection of cross border M&As compared with other entry modes (e.g., Osland, Taylor and Zou 2001; Meyer et al. 2009) or the post M&A performance (e.g., Capron and Hulland 1999; Homburg and Bucerius 2005; Chakrabarti et al. 2007; Swaminathan, Murshed, and Hulland 2008; Aybar and Ficici 2009; Chari et al. 2010; Gubbi et al. 2010). In fact, an M&A procedure typically completion stage and a post M&A integration stage (Boone & Mulherin 2007; Dikova, Sahib, and Witteloostuijn 2010). While it is challenging to integrate companies from different countries during the post M&A stage, it is also very risky in the pre completion stage. In the past several decades, there have been high failure rates in completing cross border M&As. For instance, empirical evidence suggests that up to 25% of the M&A attempts were abandoned at some point in the pre completion stage (Holl and Kyriazis 1996; Dikova et al. 2010). Muehlfeld, Sahib, and Witteloostuijn (2007) reported a range from 14% to as high as 25% of domestic deals that were abandoned or suspended for long periods. In our samp le of cross border M&As, this number reaches as high as 38% due to more
66 complicated international environments. However, little research has explored the factors that influence the completion success and how firms can enhance their chances of successfully completing a cross border M&A during the pre completion stage. Withdrawing an announced M&A deal can be costly to acquirers, entailing substantial costs (e.g., penalties), which in some cases can be as high as over 6% of the purchase value (Rosenkranz and Weitzel 2005). In addition to the costs involved in the M&A pre completion stage (e.g., payments to lawyers and M&A agents, as well as resources and time invested) and substantial penalties on the break of a contract, a substantial proprietary cost could also be incurred (Luo 2005). term deployment based on M&A announcement information. Furthermore, the termination ity (Luo 2005). Although an M&A deal termination may be more beneficial to acquirers than proceeding with a transaction that could cause dramatic future losses, firms essentially aim at a successful deal completion (Dokiva et al. 2010). Hence, it is crucia l for global marketers to understand the influential factors of a successful cross border M&A completion. Although a few recent works have started to focus on this issue (Muehlfeld et al. 2007, 2012; Dokiva et al. 2010), none of them have investigated the completion of the cross border M&As that are particularly related to rapidly growing emerging markets. Given that emerging markets may have some inherent characteristics that are essentially different from developed countries at the country level (e.g., so cial, cultural, legal, and regulatory environments) as well as the firm level (e.g., capital strength and past M&A experience), the impact of these factors on an M&A
67 completion can be fundamentally different for global expansions via cross border increasingly important roles in the success of many international firms (Burgess & Steenkamp 2006), there is a pressing need to understand whether and how cross border M&As initiated from emerging markets differ from those from developed countries. In this paper, we study an intermediary period during the completion of a cross border M&A: the public takeover process, which starts when the acquirer announces its intention of acquiring or me rging with a target company and ends with a completion announcement (or a withdrawn announcement). We develop a conceptual framework in which to investigate the influential factors on this particular period of cross border M&A completion and compare two ty pes: (1) an inbound M&A in which a firm from a developed economy acquires a firm in an emerging economy (i.e., global expansions to emerging markets), and (2) an outbound M&A, in which a firm from an emerging economy acquires a firm in a developed economy (i.e., global expansions from emerging markets). We use a sample of 1,666 cross border M&As related to the two largest emerging markets, China and India, from 1980 to 2009 to empirically examine the potential differences in terms of the influential factors on the completion of two types of M&As. Our empirical analysis reveals some fundamental differences between the two types of cross border M&As. Specifically, we find that country level factors, which measure the differences between the two countries invo lved (e.g., differences in political, trade, and legal environments), are the most influential factors for the successful completion of inbound M&As. However, firm level factors, which measure the most influential
68 factors for the completion of outbound M&As. Furthermore, we also find an asymmetric effect of past M&A experience on the completion of cross border M&As. In particular , we find that while past M&A experience may hinder the completing of inbound M&As, it helps to improve the completion of outbound M&As. These findings contribute to the literature on market entry and international marketing and provide useful managerial implications to global marketers on enhancing the success of global expansions. To our knowledge, this is the first study in market entry literature to investigate an im portant but underexplored issue the completion of M& As. We provide a theoretical framework regarding the impacts of country and firm level factors on the completion of cross border M&As and empirically demonstrate two opposite impacts: the negative impacts of country level factors (i.e., country law, regul ation distance, and country risk distance) and the positive impacts of firm level factors (i.e., size and past M&A completion experience). Furthermore, this work adds to a fast growing stream of research examining marketing strategy in emerging markets and how they may differ from those in developed markets (e.g., Yin and Choi 2005, Johnson, Yin, and Prasanna, 2012, Sudhir and Talukdar 2012, Kushwaha and Petersen 2012, Shen and Xiao 2012, Tellis and Chandrasekaran 2012). Our research explicitly compares global expansions to and from emerging markets. Our findings about the fundamental differences of the completion of the two types of cross border M&As imply that multinationals from developed countries and newcomers from emerging markets should focus on di fferent risk related factors. For example, in the case of developed multinationals entering fast changing emerging markets (i.e., inbound M&As), more attention should be given to country level factors to avoid potential pitfalls that may
69 lead to the abando nment of M&As. However, for newcomers from emerging markets that are ente ring developed countries (i.e., outbound M&As), more consideration should focus on enhancing their own capabilities in completing cross border M&As. The rest of the paper is organized into three sections. We first discuss the theoretical background and develop a conceptual framework concerning the successful driving forces of cross border M&As and the differences between the two types of cross border M&As. We then introduce the empiric al model, present our data, and discuss our empirical analyses and findings. We conclude with a discussion of contributions, managerial implications, and opportunities for future research. 3 .2 Theory and Hypothesis 3.2.1 M&A Procedure and Deal Completion A completed stage and a post merger integration stage (Boone & Mulherin 2007; Dikova et al. 2010), as shown in Fig. 3 1. The pre completed stage begins with the takeover proc bidders (i.e., the target company) and asks them to submit a preliminary indication of interest (Muehlfeld et al. 2012), or vice versa (Boone and Mulherin, 2007). After signing a confidentiality a greement, the interested bidders will receive private information of the acquiring company and engage in initial negotiations until one bidder has been selected (Boone and Mulherin, 2007; Dikova et al. 2010; Muehlfeld et al. 2012). The acquiring company wi ll then perform a detailed analysis (i.e., due diligence) to assess the organizational fit with the selected target based on a range of criteria such as relative size, type of business, capital structure, organizational strengths, core competencies, market channels, etc. The private takeover process
70 ends when the acquiring company reaches a preliminary agreement with the target company and announces it in the financial press the date of which is referred to as 3 1. Accordingly , the M&A procedure enters the completion stage. During this period, the acquirers will make further objective, independent examinations (due diligence) of targets, focusing on financial issues, asset s and business valuation, foreign government regulations, risk expropriation, etc. Finally, both parties will decide whether to close up the deal or not (i.e., complete or abandon it). If the M&A can be completed, then the second period ends with the compl etion announcement, the date of which is referred to 3 1. The second period in the pre completion stage, the public takeover process, can take several months (Dikova et al. 2010) or may end up in failure. As a detailed ana lysis has already been performed during the private takeover process, what might cause the termination of an announced M&A? We argue that withdrawing the announced M&A is caused by unanticipated new information/knowledge available to firms after entering t he public takeover stage, which can be the results of both unexpected changes occurred in the public takeover stage, and misunderstanding of existing information or overlooking some important information in the private takeover process. First, the busi ness environment in a country might change during the public takeover process; for example, new regulations and/or new economic policy could be imposed. In response, the two firms, the acquirer and target, may have to re evaluate the risk and returns of an announced M&A and renegotiate the deal with their partner. Subsequently, one or more of the parties might be determined to withdraw if the returns of the intended M&A are adversely affected under the new
71 business conditions. As most of the unexpected chan ges are beyond the two firms' control, the likelihood of a deal completion can be significantly affected by these changes during the public takeover process. Second, in addition to the unexpected changes, the assessment performed during the private period and the public takeover period have different focuses: the former is primarily on strategic decisions of organizational fit and strategic compatibility, the latter, however, also on both strategic and administrative activities related to compliance with re gulations, and final negotiations on future strategies and implementation of announcement strategies (Muehlfeld et a l . 2012). Hence, as the investigation and negotiation process proceeds to administration and implementat . ion issues, the two firms may iden tify some information that has been overlooked in the private takeover process (Hotchkiss et al. 2005), or perhaps there is a misinterpretation or misunderstanding of each other's strategic goals, financial evaluations, etc. A new understanding and newly i dentified information may reveal some potential conflicts between two parties, thus making them reconsider and renegotiate. As a result, they may have to abandon the deal if the conflicts cannot be resolved. Thus, to evaluate the completion likelihood of the two types of announced M&As (inbound and outbound, as previously defined), one needs to examine the likelihood of these causes occurring given the characteristics of the two countries, the two firms, and the deal. In the following section, we first id entify influential factors at the country, firm, and deal level under which these causes are likely to happen and then develop hypotheses concerning the impact of these factors on the completion likelihood of cross border M&As and their differential impact s on the two types of cross border M&As.
72 3.2.2 . Country Level Factors According to the extant literature on cross border M&As, the post M&A performance can be largely affected by the distance between the home and host countries in their cultural, eco nomic, regulatory, and legal environments (Shimizu et al. 2004). Among variables that measure country level distances, we expect that the two most widely examined and relevant factors to fledgling emerging countries country law and regulation distance as w ell as country risk distance can affect the likelihood of announced M&A completion. Country law and regulation d istance. T differences in laws and regulations that are related to business and is also known as the differen differences in law and regulation distance might adversely affect a deal completion because of the complexity of these systems within individual countries, which greatly increases the l ikelihood of misinterpretation, poor understanding, and important information negligence, and in turn might reduce the probability of an announced deal completion. According to institutional theory, institutions, which are the rules of 1990), are nation specific (Dikova et al. 2010). That is, the rules of the game vary across nations. Some rules may be unique to a jurisdiction or incompatible. In addition, nearly every jurisdicti on has its own stock exchange rules, securities laws, and corporate law statutes (Keegan and Green 2011). When two countries differ greatly in terms of legal and regulatory environments, firms involved in cross border M&As may encounter legal and regulator y complexities that cannot be fully interpreted (and comprehended) based on their native knowledge and skills (Dikova et al. 2010). As a result, firms may misjudge the chance for an M&A success or
73 overlook some important aspects related to the deal. If the parties realize such misjudgments or negligence after the M&A is announced, they may have to abandon the deal during the public takeover period. A recent example, the failure of China oil company, Unocal, illustrates the underestimate of political risk. Soon after CNOOC made an acquisition bid of $18.5 billion in cash for Unocal on June 23, 2005, there was strong opposition in Washington, which continued to grow and eventually forced CN OOC to environment has made it very difficult for us to accurately assess our chance of success, creating a level of uncertainty that presents an unacceptable risk to our abilit becomes significantly large, complexity can increase substantially. It then becomes difficult for acquirers to fully understand and adjust to an extensively different legal and regulatory environment, which can seriously increase the likelihood of negligence and poor understanding, and may eventually impede the deal completion. In the CNOOC example, while the Chin ese executives expected opposition and a thorough review by The Committee on Foreign Investment in the United States, they were still shocked by the intensely negative reaction from Congress (The Washington Post, August 2005). Country risk d istance. According to international business and international marketing literature, country risk refers to the risk of investing in a country in which drastic changes in the business environment may adversely affect profits or the value of assets (e.g., Rothaerme l, Kotha, and Steensma 2006; Johson and Tellis 2008). The overall country risk in an international market can come from two
74 sources: political and economic (Rothaermel, Kotha, and Steensma 2006). Political s political environment or government policy that would adversely affect a company's ability to operate effectively and of expropriation and civil disorder or ethnic conflict t o less extreme forms of tax increases, exchange rate control, imposition of tariffs and restrictions on foreign investment (Keegan and Green 2011). Similarly, economic risks, including financial hanges in the business Steensma 2006, p.59). Some economic risks that foreign businesses often encounter could be in the form of recessions or market downturns, currency crises, or sudden bursts of inflation (Johnson and Tellis 2008). Accordingly, different than country law and regulation distance that measures the differences in legal and regulatory environment between the home and host countries, country risk distance measures the difference between two countries in the likelihood of dramatic changes that may be caused by political and economic forces in the business environment and adversely affect the firms involved in M&As. Country risk distance might affect the deal complet ion in three ways. First, the higher the country risk distance, the more likely that the country with higher risk (either home or host country) may encounter adverse changes in its political or economic environment during the public takeover period. Such a dverse changes may either reduce the potential of the announced M&A or entail losses to firms involved, which in turn reduces the likelihood of deal completion. Second, even adverse changes may not significantly affect the M&A potentials; such changes coul d create new information for firms involved in the M&A to comprehend in the
75 short period of time that comprises the public takeover period. Due to time pressure after the M&A is announced, the likelihood of misunderstanding and misjudging the potentials of the announced M&A increases and in turn the likelihood for firms to abandon the M&A. Third, drastic changes can also create great uncertainty to firms involved in the M&A concerning future market stability and hence reduce their confidence in doing busine ss in such a changing business environment. Overall, as country risk distance enlarges, the likelihood of adverse changes can significantly increase in the high risk country, which in turn, decreases the likelihood of deal completion. In sum, we expect tha t both the country law and regulation distance and country risk distance can negatively affect the completion of cross border M&A and propose the following hypothesis: H1 (country level f a ctors: impact on M&A c ompletion) : In cross border M& As, the country differences between the home and the host country can affect the likelihood that the announced deal will be completed. Specifically, the deal is less likely to be completed as: The distance in country law and regulation increases, The di stance in country risk distance increases. Inbound Vs. o utbound. Considering the two types of M&As: inbound M&As (i.e., global expansions to emerging markets) and outbound M&As (i.e., global expansions from emerging markets), we expect that the direction of cross border M&As (i.e., inbound vs. outbound) moderates the impact of country level factors such as country law and regulation distance and country risk distance on an announced M&A completion. Specifically, we expect that the negative impacts of two c ountry level factors are stronger on the deal completion of the inbound M&As than on outbound M&As.
76 regulatory environment is large, firms involved in cross border M&As may encounter difficulties in fully understanding the legal and regulatory requirements of the host country based on their native knowledge and skills and, as a result, may misinterpret or neglect some important aspects in completing the M&A deal. Such misinterpretatio n and/or negligence can severely impede the announced M&A completion during the public takeover period. Compared with outbound M&As, it is more likely that firms involved in inbound M&As would encounter such misinterpretation and/or negligence and consequ ently withdraw than in outbound M&As. Note that inbound M&As involve acquirers from developed countries merging with or acquiring firms from emerging markets. Unlike the developed countries that have established relatively concrete, comprehensive, and tran sparent legislation and regulations for doing business, emerging markets still lack adequate and transparent legislation and regulations for doing business (Hitt et al. 2000). In emerging markets, business may be conducted on the interpretation of individu al officials who are in charge of starting and operating a business (Henisz and Zelner 2010). In such a business environment, with murky laws and poorly defined legislation and regulations, acquirers from developed countries often find it hard to fully und erstand how to do business (The Washington Post, April 02, 2011). This significantly increases the likelihood of misunderstanding and negligence and in turn reduces the chances of success of the announced inbound M&As. In contrast, while acquirers from eme rging markets still have to exert great effort in understanding the legal and regulatory requirements for their intended M&As in developed countries, it is relatively easier for them to comprehend as the legal and regulatory systems in developed markets ar e usually complete and clear. Thus, the chance for
77 misunderstanding and negligence would be relatively lower for outbound M&As than for inbound M&As, and, as a result, the distance in the legal and regulatory environment between developed and emerging coun tries negatively impacts the deal completion of the announced outbound M&As to a lesser extent than that of inbound M&As. Similarly, we expect the negative impact of country risk distance on the deal completion to be higher for inbound M&As than for outbo und M&As. In general, the development (Keegan and Green 2004). Compared with developed economies, emerging markets have a relatively higher level of country risk, with a higher pro bability of changes in their political and economic environment. Such changes can be a sudden government regime change due to a history of political strife, tax increases, exchange rate control, or imposition of tariff and restrictions of foreign investmen t because of rapid economic reforms (Arnold and Quelch 1998). As reported in The Washington Post (April 02, 2011), many foreign investors have been telecommunications giant, for example , was slapped with a $2.5 billion capital gains Consequently, the U.S. and British ambassadors, the European Commission, and ngton Post, April 02, 2011). A World Bank study in 2004 revealed that 15% 30% of the contracts covering $371 billion of private infrastructure investment in the 1990s were subject to
78 government initiated renegotiations or disputes in emerging markets (Hen isz and Zelner 2010). As many of these political and economic changes are difficult to foresee in emerging markets, acquirers from developed countries may be subject to one of these changes that proves to be detrimental after they announce an inbound M& A deal in the emerging market. Consequently, they may be forced to withdraw from the deal. Conversely, in a developed country, the host country in the outbound M&A tends to have more entrepreneur friendly regulations, better protection of intellectual prop erty rights, less corruption, and more transparent and better functioning capital markets, all of which makes the outcome easier to forecast. Thus, due to the fast changing and highly uncertain environment in emerging markets, the negative impact of countr y risk distance on deal completion might be higher for inbound M&As relative to outbound M&As Based on the above arguments, we develop the following hypothesis: H2 (country level factors: inbound vs. o utbound M&As) : The negative impact of country level dis tances on deal completion is stronger for inbound than outbound M&As. Specifically, The distance in country law and regulation has a stronger negative effect on inbound than outbound M&As, The distance in country risk distance has a stronger negative eff ect on inbound than outbound M&As. 3.2.3 . Firm Level Factors Compared with country level factors that are mostly beyond a firm's control and represent main challenges that cross border M&As encounter, firm specific advantages (i.e., financial capability, human resources, and cross border M&A knowledge) can significantly enhance the success of global entry (Johnson & Tellis
79 2008). Built on the resource based view literature (e.g., Barney 1991), we expect that two firm level factors, firm size and experienc e on cross border M&As, which reflect the tangible and intangible resources at the firm's disposal, can significantly enhance the completion of cross border M&As. Firm s ize. There are three reasons why larger firms might have a greater completion rate than smaller firms for the announced cross border M&As. First, larger firms have more resources or can commandeer more resources than smaller firms (Bonaccorsi 1992; Johnson & Tellis 2008), which in turn greatly reduces the possibility of misunderstanding and negligence during the private takeover stage. For example, a larger firm can hire an experienced local consulting firm and lawyers to help them better prepare and negotiate for the intended cross border M&A. Second, because the opportunity costs of incompl ete M&A deals tend to be higher for larger firms as they have more business partners and stakeholders, they are more motivated to avoid a deal failure. Also, as it is more detrimental to their reputation if the announced M&A fails, larger firms are more mo tivated to try all means to eliminate potential misunderstanding or negligence during the private takeover process. Third, during the public takeover stage, large firms can take advantage of financial resources to finalize/secure the deal, especially when they encounter misunderstandings, negligence, or adverse changes. Overall, we argue that as firm size represents tangible resources that a firm possesses, the larger the acquirers, the more likely they can complete the announced completed M&A deals. M&A e x perience. Literature has suggested that organizations can learn from their past experience of post M&A management and apply the knowledge and skills accumulated to the integration of new M& As (Shimizu et al. 2004). Two recent works have also shown that acquirers can benefit from their past M&A completion
80 experience by transferring some of the acquired skills and knowledge to the completion of a new M&A (Dikova et al. 2010; Muehlfeld et al. 2 012). While these two works primarily focus on M&A completion without specifically related to emerging markets, we expect that past completion experience can also help improve the success of completing both inbound and outbound M&As that are related to eme rging markets. For example, a previous M&A completion experience enables a company to know what important information should be assessed and anticipated (e.g., what might be regulatory barriers in the host country) and to better negotiate and prepare respe ctive strategies when encountering obstacles during the public takeover process. Furthermore, firms that have successfully completed cross border acquisitions in the past may have developed organizational routines for the takeover process, and how to acces s outside financial, legal, or other resources (Shimizu et al. 2004). Hence, we propose the following two hypotheses regarding the impact of two firm level factors on the completion of inbound and outbound M&As. H3 (firm l evel f actors : impact on M&As c omp letion): In cross border M&As, the firm level factors can affect the likelihood that the announced deal will be completed. Specifically, the deal is more likely to be completed as The size of the acquirer increases, bo rder M&As increases. Inbound vs. o utbound. In contrast to the two country level factors that have more negative impacts on the deal completion of inbound M&As than on outbound M&As, we expect that the firm level factors will have more positive impacts on the deal completion of outbound M&As than on inbound M&As. There are three reasons for this. First, compared with acquirers from developed countries, large size acquirers from emerging markets might enjoy more resources because of the
81 important roles they play in their home country economy. Most emerging markets impart special favors on large firms (The Economist, 2010). For instance, large firms have been found to have enduring advantages in attracting and training talents in rapidly growing emerging mark ets (The Economist, 2010). Because of the size may signal its credibility and capability, which further helps large firms secure more external resources for globalization , such as generous financial aid, favored policies from the government, or financing privileges from the local banks. Thus, large acquirers from emerging markets, the acquirers in outbound M&As, have more resources to better prepare and negotiate during th e private takeover stage and to renegotiate, if necessary, when new information emerges (e.g., policy changes) or misunderstanding and negligence is found during this stage. In contrast, because of the highly competitive and relatively efficient markets in developed countries, large acquirers from developed countries, the acquirers in the inbound M&As, have no such advantages in securing more resources as acquirers in the outbound M&As. As a result, firm size plays a more positive role in the completion of M&A deals for outbound M&As than for inbound M&As. Second, in regard to the impact of past completion experience on inbound M&As vs. outbound M&As, we expect it is much easier for acquirers to learn from this experience in a developed country (the host c ountry of outbound M&As) than in an emerging market (the host country of inbound M&As). In examining the limitations in which firms apply their resources, the resource view literature has suggested that remains relatively fixed (Kraaijenbrink, Spender, and Groen 2010). In regards to our application of past M&A completion experience (i.e., an intangible resource),
82 because the business environment (i.e., legal and regulatory environment) in developed countries is well defined, transparent, and relatively stable over time, the acquirers from emerging markets can easily summarize their experience from their past M&A completions and apply to a new one. Finally, because the acquirers from emerging markets are newcomers in global expansion, they have considerably limited knowledge about the process of cross border M&As. According to the resource based view (e.g., Barney 1991), resources in general can be t ransformed into a competitive advantage only if they are valuable, rare, and hard to imitate. For young players on the globalization stage, because of their limited knowledge of cross border M&As, learning from past experience is extremely valuable and ten ds to be more effective in helping to manage the completion of a new cross border M&A. As reported in McKinsey Quarterly (2008), previous cross border experience and talent in this kind of experience are shown to greatly affect the success of global expans ion from emerging markets. The lengthy negotiation between Lenovo and IBM (about 18 months) demonstrates that even the best Chinese firms have still not developed the killer instinct or are skillful enough to close a deal quickly (Financial Times, 2006). I n contrast, for acquirers from developed countries, those who have substantial experience in processing a cross border M&A, learning from the past tends to be less effective as organizations may develop inertia and become used to their existing routines. Based on the above analysis, we have the following hypothesis: H4 (firm level factors: inbound vs. o utbound M&As): The positive impact of firm level factors on deal completion is stronger for outbound than inbound M&As. Specifically,
83 T he size of the acqui rer has a stronger positive effect on outbound than inbound M&As. The acquirer's experience has a stronger positive effect on outbound than inbound M&As. 3 .3 E mpirical A nalysis 3.3.1 Model We apply a logistic regression model with M&A deal completion as a dependent variable to estimate how the completion likelihood of a cross border M&A is affected by the country and firm level factors (e.g., Muehlfeld et al. 2007, 2012; Dokiva et al. 2010). Specifically, the probability of deal completion is assumed to be a logistic function of exploratory variables such as country level, firm level variab les and other control variables ( 3 1) where X i is a vector of explanatory variables, with coefficient being a vector of parameter estimat es. i is an indicator of M&A deals. The empirical specification for function X i is given by , ( 3 2) Where law_reg_dis i and country_risk_dis i denote the distance in law and regulation and the distance in country risk between the home country and host country at the time when an M&A deal i is announced, respectively. The variables SIZE i and EXP i denote the acquirer's size and past M&A experien ce of the M&A deal i , and out i is set to 1 for the outbound M&A i , and 0 for the inbound M&A i . The vector control i includes all other country , firm and deal level variables, which we explain in detail in the following empirical section. Thus, the coeff 1 4 capture the main effect of country and firm level factors on the likelihood of deal
84 completion, which are hypothesized in H1 and H3; whereas the coefficients of the interaction terms between country or firm level variables and variable out i 5 8 , capture the moderating effects of M&A direction on the cross border M&A 9 represents the main 10 18 is a vector of coefficient s of control variables. 3.3.2 Data We collected all cross border M&As related to emerging markets from the Thomson Securities Data Corporation's (SDC) Mergers and Acquisitions Database. This database provides comprehensive information about M&A deals world wide on the date of a cross border M&A's announcement, its completion status, acquirer and target's information, and deal specific information. It has been extensively used in academic research on M&As in management, finance, international business, and ma rketing (e.g., Dikova et al. 2010; Swaminathan et al. 2008; Beckman and Haunschild 2002). To empirically examine and compare the impact of country and firm level factors on the completion of two types of cross border M&As, we constructed two samples: inb ound M&As and outbound M&As. The inbound sample consists of all publicly disclosed M&As whose acquirers are from developed countries and targets are firms in one of two emerging markets China and India. The outbound sample consists of all publicly disclose d M&As whose acquirers are either Chinese or Indian firms and targets are firms in developed countries. We focus on the cross border M&As that expanded to or from these two leading emerging markets, as China and India have become major players in the world economy and cross border M&As related to China and India accounted for 68% of the total cross border M&As that
85 are related to emerging markets (i.e., BRIC) in the last decade, respectively. Following Burgess and Steenkamp (2006), we used the FTSE Group's classification samples, including the United States, the United Kingdom, Canada, Japan, France, Australia, Germany, Netherlands, Switzerland, Sweden, Italy, Belgium, Finland, Spain, Norway, Denmark, New Zealand, Austria, Luxembourg, Greece, Portugal, and Singapore. The cross border M&As in our two samples span from 1980 to 2009, a period that includes the time when the Chinese and Indian governments first opened their doors to foreign investors in the early 1980s, and when many of the leading Chinese and Indian firms started going global in the l990s. To identify reliable and meaningful data, we further selected the cross border M&As in which acquirers are public firms (targe t firms might either be public or private firms). Focusing on publicly traded acquirers allowed us to collect rich financial information of acquirers in our samples. Overall, our inbound and outbound samples comprise 1,212 and 454 cross border M&As, respec tively. Table 3 1 provides the distribution statistics of the cross border M&As in our two samples. 3.3.3 Variables Deal c ompletion. The dependent variable in our empirical analysis is the completion status of announced M&As. Following literature (e.g., B ao and Edmans, 2009; Muehlfeld et al. 2007, 2012; Dokiva et al. 2010), we define the deal completion to be 1 if the deal is completed and 0 otherwise. As stated earlier, the Thomson SDC's M&A database provides information about the announced M&As on the da tes of announcement and completion as well as the completion status (e.g., withdraw or pending). According to the literature, the median number of days to
86 completion is about 62 with 94% of all deals completed within a year (e.g., Muehlfeld et al. 2012). W ithin our sample related to emerging markets, the mean number of days to completion is about 69 for all deals, with 79 days for the inbound sample and 44 days for the outbound sample, respectively. Since we examine the completion status till the April of 2 013, four more years after 2009, for the cross border M&As that were pending till the April of 2013, we considered them to be withdrawn. Country law and regulation d istance. Following management and international business literature (e.g., Meyer et al. 20 09; Gubbi et al. 2010), we apply the economic freedom index developed by the Heritage Foundation to construct the variable for country law and regulation distance between the home and host countries. As property right protection plays an important role in cross border M&As (e.g., Chari et al. 2010), we use a country's freedom score in property rights, developed by the Heritage Foundation, to proxy a country's law and regulation environment. We then adopt the formula in Tsang and Yip (2007) and measure the distance in law and regulation between two countries as the freedom score of the host country divided by that of the home country for inbound M&As. For outbound M&As, the distance in law and regulation is measured in the reverse way: i.e., the freedom scor e of the home country is divided by that of the host country. Accordingly, the higher the variable value, the smaller the distance in law and regulation between the two countries. Country risk d istance. In line with the literature (e.g., Johnson & Tellis, 2008; Dikova et al. 2010), we derive the country risk distance using the country risk score from the PRS Group's International Risk Guide, which provides a three dimensional measure for each country on political, financial, and economic risk. Specifically , following Johnson & Tellis (2008), we first develop a composite country -
87 risk measure for each country (see the explanation in the PRS Group's International Risk Guide and Johnson & Tellis 2008). Similar to the measurement for country law and regulation d istance, we then measure the country risk distance as the country risk score of the host country divided by that of the home country for inbound M&As and measure it in the reverse way for outbound M&As. Thus, the higher the variable value, the smaller the distance in country risk between the two countries. Firm size and M&A e xperience. Similar to the previous literature (e.g., Gubbi et al. 2010; Johnson & Tellis 2008), we measure the size of the acquirer as the natural logarithm of the acquirer's average t otal assets two years prior to the date when the acquirer announced the deal. The total assets are collected from Datastream and Thomson Research. The past M&A experience is measured as the number of completed deals by acquirers in cross border M&As prior to the announced M&A deal. Control v ariables. Following Aguilera & Decker (2008), we measure Industry Relatedness to be 1 if the two SIC codes of the target and acquirer in a cross border M&A are the same and 0 otherwise. The SIC codes are available from t he Thomson M&A Database. We adopt Hofstede's four cultural dimensions to measure the cultural distance between the home and host country in a cross border M&A (Geert 1991). Specifically, we compute the cultural distance as , wher e S A,i and S T,i dimension i. We incorporate the target firm's ownership status, Public Target, in our empirical analysis such as public target = 1 when the target is a public firm and 0 otherwise. Further more, we also control some deal specific variables as provided in the Thomson M&A Database. Specifically, we use Stake Sought to measure the
88 percentage of stake that the acquirer seeks from the target, and Cash to capture the payment method of the deals su ch that Cash = 1 if the deal is paid totally through cash or 0 otherwise. Two other dummy variables are also incorporated such that Attitude = 1 if the manager perceived the deal as a friendly M&A and 0 otherwise, and Disclose = 1 if the transaction value of the deal is disclosed in the public announcement or 0 otherwise. Finally, we use Year to capture the time trend and define a dummy variable, India, to distinguish cross border M&As related to India from those related to China, such that India = 1 if eit her the acquirer or target is an Indian firm and 0 otherwise. The descriptive statistics of all variables are reported in Table 3 2. 3.3.4 Results We estimate the log istic regression model in Eq uations 3 1 and 3 2 by a full sample including the inbound an d outbound sample together. Specifically, we first estimate the logistic regression model without interaction terms to show the main effects of the country and firm level factors on the deal completion and then estimate the full model with interaction ter ms included. The estimation results are presented in the columns of Model 1 and Model 2 in Table 3 3, respectively. Because the test for heteroskedasticity using both the Goldfeld Quandt test (p < .01) tcity might be a potential issue for our data, we use the corrected white co variance matrix in both logistic regression estimations to correct the heteroskedastcity issue. Overall, as shown in Table 3 3, Model 2 (i.e., the full model) fits better than Mod el 1 (i.e., the main effect model). Impacts of country level f actors. As shown in Model 1 of Table 3 3, both the coefficients of country law and regulation distance and the coefficient of country -
89 1 = 2 = 2.080, p < .01), suggesting that the larger country distance in law and regulation and in country risk, the less likelihood of an M&A deal completion. Thus, our empirical evidence shows the sign as predicted in H1(a) and supports H1(b). When considerin g the moderating effect of M&A direction, as shown in Model 2 of Table 3 3, the coefficients of the interaction terms between country law and regulation distance and Out and between country 5 = .385, p < .01 and 6 = 5.079, p < .01), suggesting that the negative impacts of country distance in law and regulation and in country risk on the deal completion are weaker for outbound M&As. That is, the negative impacts of country distance in law and regulation a nd in country risk on the deal completion are stronger for inbound M&As than for outbound M&As. Thus, H2(a) and H2(b) are supported. Furthermore, as shown in Model 2 of Table 3 3, the net impacts of country law and regulation distance and country risk dist ance on the deal completion for inbound M&As (i.e., when Out = 0) are significantly 1 = 2 = 3.694, p < .01), whereas the net impacts of these two country level factors on the deal completion for outbound M&As (i.e., when 1 5 2 6 =1.385, p > .1). This further suggests that country level distance in law and regulation and in country risk between the home and host country are most influential in inbound M&As (i.e., global expansion from developed economies to emerging markets) than in outbound M&As. Impact of firm level f actors. As shown in Model 1 of Table 3 3, the 3 = .138, p < .01), suggesting that the size of an acquirer can significantly increases the likelihood of deal completion. Thus, H3(a) is supported. To see the difference with regard to the
90 impact of acquirer size on the deal completion between inbound M&As and outboun d M&As, we found in the results of Model 2 in Table 3 that the coefficient of 7 = .100, p < .05). This result implies that the positive impact of firm size on the deal com pletion is stronger for outbound M&As than for inbound M&As. Therefore, H4(a) is also supported. However, we found that the coefficient of a firm's past M&A experience is 4 = .010, p < .01), implying that a firm's past M&A expe rience could negatively impact the deal completion. Our further examination of the results of Model 2 in Table 3 3 shows that the coefficient of the M&A experience 4 = .008, p < .05) but the interaction term between M&A exper ience and Out is significantly positive ( 8 = .276, p < .01). These two results imply that while past M&A experience reduces the likelihood of the deal completion of inbound M&As, it increases the likelihood of the deal completion of outbound 4 + 8 = .268, p < .05). Hence, H3(b) is only supported for outbound M&As and H4(b) are empirical supported. The negative impact of past M&A experience on the deal completion for inbound M&As suggests that the skills and knowledge that an acquirer developed thro ugh past cross border M&As may hinder a new cross border M&A completion. A possible reason for this negative impact of M&A experience may be because the accumulated skills and knowledge are market specific. For example, the skills and knowledge that an acq uirer has attained in developed countries for cross border M&As may not be transferable to emerging markets. Thus, to investigate whether the learning effect from past M&A experience can be market specific, we measure the past M&A experience using emerging market experiences for inbound M&As
91 and using developed market experience for outbound M&As. Specifically, instead of measuring the past M&A experience as a total number of past completed cross border M&As regardless of where the host country is (i.e., a t a global level), we measure the emerging market M&A experience as a total number of past completed inbound M&As particularly in China and India and measure the developed market M&A experience as a total number of past completed outbound M&As where the ho st countries are developed countries as defined previously. The estimation results by incorporating such a market specific experience are presented in Table 3 4. The results in the column of Model 1 and Model 2 in Table 3 4 correspond to the estimation res ults of main effect model (without interaction terms) and of full model (with interaction terms), respectively, where the M&A experience is measured as market specific. As shown in Model 1 of Table 3 4, the coefficient of M&A 4 = .063) though insignificant (p < .1). When we incorporate the moderating effect of M&A direction, we find that while the impact of past M&A experience is still significantly positive on the deal completion of outbound 4 + 8 = .290, p < .05), it is positive but insignificant on the deal completion of inbound M&As ( 4 = .024, p > .1). These results, combined with those in Table 3 3 show that while the acquirers from emerging markets can benefit from past M&A experience in completing a new cross border M&A in developed markets, the counterpart from developed economies may not benefit or may even suffer if the parties simply apply their skills and knowledge accumulated from the past to a new cross border M&A in emerging markets. Such asymmet ric learning effects from past experience reflect the fundamental differences of emerging economies from developed economies as emerging economies are still transitional and their political regimes, economic structure, and regulations may change over time; thus, past
92 information acquirers have learned may not be applicable to the present. In contrast, as developed economies are relatively stable, during the private takeover stage, acquirers can better prepare for what might happen in the intended cross bord er M&A based on their past experience (especially from their past experience in developed countries), which helps acquirers enhance the completion of present cross border M&As in developed countries. In sum, our empirical analyses reveal three key finding s concerning the fundamental differences on the deal completion of two types of cross border M&As. First, our results show that while country level distance (i.e., the distance in country law and regulation and in country risk) can decrease the likelihood of cross border M&A completion, firm level factors (i.e., acquirer size) in general increase the likelihood of cross border M&A completion. The former results concerning the impact of country level distance are consistent with Dikova et al. (2010), which s hows that the institutional distance (measured only by country risk distance) reduces the likelihood of cross border M&A completion. However, different from Dikova et al. (2010), which studied the deal completion of cross border M&As in the service industr y in developed countries, we examine the deal completion of cross border M&As related to emerging markets and explicitly compare the deal completion of inbound and outbound M&As in emerging markets. Second, in comparing the deal completion between inbound and outbound M&As, we found that the negative impact of country distance (i.e., the distance in country law and regulation and in country risk) on the deal completion is stronger for inbound M&As than for outbound M&As, whereas the reverse pattern holds fo r the impact of firm level factors. Specifically, the positive impact of firm level factors (i.e.,
93 inbound M&As. Third, we found different learning effects from past M&A ex perience on the deal completion of inbound and outbound M&As. Specifically, we found that acquirers from developed countries may be hurt when applying their past M&A experience to the cross border M&As in emerging markets, as emerging markets are transitio nal and what acquirers learned may not be applicable to the present. On the contrary, we found that acquirers from emerging markets can actually benefit from their past experience in processing new cross border M&As in developed countries, as the latter ma rkets are stable and thus learnable. Robustness and validity of r esults . To examine the robustness and validity of our results, we performed several additional analyses. First, we tested for the multi collinearity issue. As the variance inflation factors ( VIF) of all our variables are within acceptable levels (<10), multi collinearity is not a major issue among the variables incorporated in the logistic regression model. Second, as our sample of cross border M&As is collected from various countries related to various industries and ranging for almost 30 years, there may exist certain correlations within countries, within industries, and/or within time periods. To control for such potential correlations across observations, we allow the covariance matrix of e rror term in our model to be non identical (i.e., assuming the off diagonal terms in the covariance matrix to be non zero for observations within same countries, within same industries, and/or same years). The estimation results by controlling such correla tions in error terms are presented in the column of Model 1 and Model 2 in Table 3 5 . Specifically, Model 1 and 2 demonstrate the estimation results and adjusted standard errors by controlling two levels of correlations: within country and within industry correlations
94 and three levels of correlations: within country, within industry, and within time, respectively. As can be seen from the results of Model 1 and 2, the significance of our key results still hold when standard errors are adjusted. Third, to fu rther control cross industry and cross country effects, we estimate the logistic regression model in Eqs. ( 3 1) a nd ( 3 2 ) by incorporating some industry dummies and country region dummies (see Table 3 6 for the distribution of cross border M&As in our samp le by industry and by developed countries and the definitions of industry and region dummies). The estimation results by incorporating cross industry and cross region effects are presented in the column of Model 3 and Model 4 in Table 3 5 . As shown in Mode l 3 and 4 of Table 3 5 , our key results concerning the main effects of two country and firm level factors and the interaction effects between the two country and firm level factors and the M&A direction still hold when cross industry and cross region eff ects are incorporated. Finally, we also estimated our logistic regression model by using several sub samples separately. Specifically, we separate our full sample into: (1) inbound and outbound samples, as we introduced earlier; and (2) India and China sam ples. The India sample includes inbound and outbound M&As, which are either to or from India, whereas the China sample includes inbound and outbound M&As which are either to or from China. The estimation results are reported in Table 3 8 , where the first a nd second column provide the estimation results using the inbound and outbound sample separately, and the third and fourth column provide the estimation results using the India and China samples separately. As shown in the first and second column of Table 3 8 , our results in Table 3 3 still hold when we estimate the logistic regression using the inbound and outbound sample separately. For example, the impact of two country distance variables,
95 country law and regulation distance, and country risk distance, are mainly negative (except the insignificant impact of country risk distance on the deal completion of outbound M&As); whereas the impact of two firm level variables are mainly positive (except the impact of M&A experience is significantly negative, as we explained earlier). Also, when comparing the impact of the two country and firm level variables on the deal completion between inbound and outbound M&As, the results demonstrate a consistent pattern as that in Table 3 3. Lastly, as shown in the third and fo u rth column of Table 3 8 , the estimation results by using the India and China sample separately also show a mainly consistent pattern as the full sample. The 2 = 20.52, p >.1) further suggests that there is no systematic difference between the India and China sample, thus it is valid to combine the two samples together. 3 .4 Summations With rapid economic development in recent decades, emerging markets have become not only global centers attracting thousands of foreign direct investment via cross border M&As , but also global contenders with many of their companies using M&As as their main globalization strategy to expand aggressively worldwide. Clearly, it is challenging for both multinationals from developed countries to expand into emerging markets and upstart companies from emerging markets to expand into global markets. It is also difficult in processing both cross border M&As and managing the post M&A integration . While past literature has extensively studied what factors impact the post M&A performance and focus mainly on the cross border M&As as related to developed markets, the impact of success in processing a cross border M&A and how global expansion via cross border M&As relate to emerging markets differ in their M&A directions (i.e. , to vs. from emerging markets)
96 has been largely ignored. This paper examines an important period in the process of a cross border M&A the public takeover period (i.e., from when an intended cross border M&A is announced to the completion of such an announ ced M&A) and investigates (1) what influential factors affect the completion of cross border M&As that are related to emerging markets; and (2) how the impact of these factors differs in global expansions via cross border M&As to emerging markets (i.e., i n bound M&As ) and from emerging markets (i.e., o utbound M&As ). 3.4.1 Research Contribution s First, this paper contributes to the marketing research on the success of foreign market entry with a special focus on the success of completing cross border M&As. In contrast to prior literature which has primarily focused on the performance of post M&As, we draw attention to an important but underexplored intermediary stage of completing cross border M&As the public takeover stage and provide conceptual mechanisms to show how an announced M&A may end up failing. Specifically, we argue that three reasons could lead to an M&A failure: (1) unexpected changes; (2) misunderstanding of existing information; and (3) neglecting important existing information. We accordingl y identify country , firm and deal level factors and empirically demonstrate the opposite impacts of country and firm level factors. While the country level distance (i.e., law and regulation distance and country risk distance) decreases the likelihood o f a cross border M&A completion, firm level factors (i.e., acquirer size and previous M&A experience) can enhance the likelihood of a cross border M&A completion. Second, our findings suggest that multinationals from developed countries and upstart compan ies from emerging markets should recognize the fundamental differences of global expansions to and from emerging markets in completing their
97 cross border M&As and pay attention to country and firm level influential factors. Specifically, our results imply that for multinationals from developed countries, the most significant risk, which is sometimes unforeseeable, is the possibility of changes in the investment environment in the host country when entering the emerging markets. The lack of understanding/preparation of such possibilities could potentially raise unexpected barriers that in turn cause the termination of an announced M&A. Finally, this paper also extends the learning theory in ma rketing and organizational behavior literature by underscoring a new contingency in learning from past M&A completion experience. Although previous literature has shown some contingencies that limit the transferability of past skills and knowledge to man aging a new completed cross border M&A, we add to the literature by showing a new contingency the target market's economic development in transferring past M&A completion experience to successfully completing a new cross border M&A. In addition to two rece nt works ( Dokiva et al. 2010; Mue hlfeld et al. 2012), which have demonstrated the contingencies in learning from past completion experience such as institutional distance (i.e., country risk distance) and industry similarity; we suggest that target market development could be another contingency and, as such, learning from past completion experience may be asymmetric. For example, when expanding to a target market that is still undergoing economic development and thus transitional, our results show that the skills and knowledge accumulated from this market or similar markets may not be helpful or may even be harmful to the completion of a new cross border M&A in such markets. In contrast, when expanding to a target market that is well established and conside rably stable, past M&A experience can help enhance the success of completing a new cross border M&A.
98 3.4.2 Managerial Implications Despite the vast opportunities, engaging in and completing cross border M&As related to emerging markets is usually highly u ncertain. Given the substantive costs from the failure of a cross border M&A (i.e., monetary cost, reputational cost, and information cost), it is critical for global companies to understand how they can enhance the chance of success of completing a cross border M&As. O ur findings provide some managerial implications for multinationals. First , we raise the attention for multinationals to the intermediary stage in completing a cross border M&A the public take over stage. Despite the large amount of attention that has been paid to post M&A management, we show the high failure rate in processing a cross border M&A and identify three causes that can lead to the failure of an announced M&A. Thus, to enhance the chance of success in completing an M&A, multinationa l and upstart companies should assess how they can avoid the three possible pitfalls. For example, to avoid any unexpected change that is detrimental to the intended M&A deal, global companies may have to evaluate the host country's political, legal, and r egulatory environment or consult expertise to better anticipate any possible regulatory change and prepare respective strategies before announcing their deals (i.e., during the private takeover stage). To avoid misunderstanding or negligence, a thorough an alysis/research of the target company and its host country is also needed . Second, our findings suggest that multinationals from developed countries and upstart companies from emerging markets should recognize the fundamental differences of global expansi ons to and from emerging markets in completing their cross border M&As and pay attention to country and firm level influential factors. Specifically, our results imply that for multinationals from developed countries, the
99 most significant risk, which is s ometimes unforeseeable, is the possibility of changes in the investment environment in the host country when entering the emerging markets. The lack of understanding/preparation of such possibilities could potentially raise unexpected barriers that in turn cause the termination of an announced M&A. On the contrary, our results that firm level factors (i.e., size and experience) are more influential in global expansions from emerging markets imply that the strongest impacts on the success of a M&A deal come from the acquirers themselves. While understanding that country level differences are still important to newcomers from emerging markets, it is less of a problem as the relatively reliable and predictable financial, legal, and economic systems in develope d nations make it considerably easier for newcomers to learn and prepare during the very early stage of a M&A and are less likely to incur large unexpected changes after the deal is announced. Instead, the challenge lies in whether the newcomers have capit al assets and past M&A experience in selecting the right acquisition target , can cobble together a network of experienced investment banker and lawyers, and manag e often skeptical regulators, unions , and stakeholders in the developed countries during the d eal renegotiation stage . Third, our results on past experience learning also raise caution to multinationals in completing an inbound cross border M&A in emerging markets . Without understanding the potential limitations in transferring past skills and know ledge, multinationals could make mistakes in basing any decisions on past experience. For upstart companies from emerging markets, we demonstrate the benefits of learning from past completion experience. New players from emerging markets could learn from t heir past, summarize key lessons, and accumulate skills in order to improve the success chance in engaging in global expansions.
100 Finally, our results also suggest some differences in the impact of deal level factors on the completion of two types of cross border M&As. For example, we find that paying in cash and owning a large percentage of stakes have stronger impacts on the deal completion of inbound M&As than that of outbound M&As. This result underlines the importance of satisfying the target's financi al needs by paying cash immediately and maintaining control over the global expansion in completing the cross border M&As to emerging markets. In contrast, our results that disclosing the M&A transaction value tends to have a stronger impact on the complet ion of global expansions from emerging markets imply that signaling their capability and confidence in buying a foreign company in a developed country can be a success factor as well for new contenders from emerging markets. 3.4.3 Limitations and Further R esearch T his paper is subject to several limitations that provide opportunities for future research . First, as an initial study in examining the completion of global expansions to and from emerging markets, we focus only on cross border M&As related to the two largest and fastest growing emerging markets: China and India . I t w ould be more insightful if future studies could examine the completion of global expansions related to other developing countries such as Russia and Brazil. Second , due to data availab ility, we focus on only public acquirers in our study. Further research can examine the generalizability of our results by including both public and private acquirers when more data on private firms become available. Using data from both public and private acquirers , future study can also explore how the completion of global expansion differs in acquirers' public status as well as in the M&A direction (i.e., to and from emerging markets). Finally, it would also be interesting to investigate how the failure of completing a cross border M&A impacts the acquirer.
101 For example, how would the announcement of withdrawing from an announced M&A impact the acquirers in the different (asymmetric) way as the impact of the announcement of completion? Future study could a pply the event study methodology to examine such a potentially asymmetric impact of completion vs. a withdrawn announcement of a cross border M&A.
102 Table 3 1. Inbound and Outbound M&As Distribution Inbound M&A (N=1212) Outbound M&A (N=454) Number Percentage Number Percentage Emerging markets China 750 62% 81 18% India 462 38% 373 82% Developed Countries North America 531 44% 225 50% Europe 395 33% 154 34% Asia * 286 23% 75 16% Year 1980 1995 20 2% 2 0.4% 1995 1999 125 10% 14 3.1% 2000 2004 359 29% 96 21% 2005 2009 708 58% 342 75% Completion Status Completed 757 62% 303 67% Uncompleted 455 38% 151 33% * : Including developed countries in Asia, plus Australia and New Zealand. Table 3 2. Descriptive Statistics Variable Inbound M&A (N=1212) Outbound M&A (N=454) Mean SD Mean SD Law and regulation distance * 2.839 1.063 2.1 3 5 . 897 Country risk distance 1.11 3 . 1 15 1.1 30 .0 9 7 Acquirer size 13.3 43 3.9 86 11. 8 90 2.02 9 M&A experience 9.593 16.594 .62 2 1.11 8 Culture Distance 17. 097 4.2 46 15.0 69 2. 897 Industry Relatedness .48 7 . 500 .60 0 .490 Target Status .20 4 .40 3 .16 5 .37 1 Stake Sought 58. 191 36. 205 79. 364 31. 765 Cash .15 5 .3 62 .19 3 .39 5 Disclosed .52 5 . 500 .56 0 .49 7 Attitude .89 9 .30 2 .96 5 .184 India .3 81 .48 6 .82 2 .38 3 Year 2004. 366 3 .732 2005. 571 2. 978 * : S ample size for law and regulation distance N = 1188 for inbound M&As and = 452 for outbound M&A due to missing data.
103 Table 3 3. Impact of Country and Firm Level Factors on Cross Border M&A Completion Variables Model 1 Model 2 Law and Regulation Distance .071(. 10 2) .157* (. 109 ) Country Risk Distance 2.080***(.833) 3.694 ***( .954 ) Acquirer Size .138***(.020) .12 2 ***(.02 1 ) M&A Experience .010***(.004) .008**(.004) Law_Reg_Dis Out .385 ** * (.1 62 ) Country_Risk_Dis Out 5.079 ***(1. 539 ) Size Out .10 0 **(.06 0 ) M&A Experience Out .2 76 ** * (.1 09 ) Out 7.935 ** * (2. 082 ) Culture distance .039**(.017) .034**(.017) Industry Relatedness .010(.109) .050(.110) Target Status .177(.178) .2 29 (.18 5 ) Stake Sought .005***(.002) .005***(.002) Cash .280*(.1 79 ) .2 56 *(.1 79 ) Disclosed .605***(.122) .64 3 ***(.12 3 ) Attitude .070(.2 0 9) .0 57 (.2 07 ) Year .062***(.021) .0 75 ***(.02 2 ) India .563***(.235) . 672 ***(.2 44 ) Log likelyhood Value 1011.587 999.161 Pesudo R2 .0 64 .075 Note: *** : p <0.01, ** : p<0.05, * : p<0.1, one tail test. Numbers in parentheses are standard errors. Due to missing data in variables of country law and regulation distance and acquirer size, sample size N = 1640.
104 Table 3 4 . Impact of Country and Firm Level Factors on Cross Border M&A Completion Variable Model 1 ` Model 2 Law and Regulation Distance .07 8 (. 101 ) .165* (. 109 ) Country Risk Distance 2. 2 8 1 *** (.8 28 ) 3.787 *** ( .946 ) Acquirer Size .1 02 *** (.0 17 ) . 093 *** (.0 18 ) M&A Experience 0.063 (.0 6 4) 0.024 (.0 67 ) Law_ Reg_ Dis Out .367 ** ( .1 63 ) Country_ Risk_ Dis Out 5.126 *** (1. 544 ) Size Out .1 27 ** (.0 59 ) M&A Experience Out .2 66 ** (.1 43 ) Out 8.225 ** * (2. 084 ) Culture distance .03 1 ** (.01 6 ) .0 28 ** ( .017) Industry Relatedness .01 6 (.10 8 ) .0 42 (.1 09 ) Target Status .1 24 (.17 5 ) . 197 ( .18 2 ) Stake Sought .00 5 *** ( .002) .005 *** ( .002) Cash .28 9 * (.1 77 ) .2 72 * (.1 78 ) Disclosed .6 14 *** ( .12 1 ) .6 37 *** ( .12 2 ) Attitude . 164 (.2 07 ) . 135 (.2 04 ) Year .06 9 *** ( .02 1 ) .0 79 *** ( .02 1 ) India .56 7 *** (.23 3 ) . 663 *** (.2 44 ) Log likelyhood Value 10 22 . 311 1011.317 Pesudo R 2 .059 .069 Note: *** : p <0.01, ** : p<0.05, * : p<0.1, one tail test. Numbers in parentheses are standard errors. Due to missing data in variables of country law and regulation distance and acquirer size, sample size N = 1640 . M&A experience is measured as market specific experience.
105 Figure 3 1. M&A Procedure Figure 3 2. Conceptual Framework
106 Table 3 5. Robustness of Results Variables Model 1 Model 2 Model 3 Model 4 Law and Regulation Distance 0.157 (0. 126 ) 0.157* (0. 118 ) 0.159* (0. 110 ) 0.144* (0. 111 ) Country Risk Distance 3.694 *** ( 1.174 ) 3.694 *** ( 1.115 ) 3.794 *** ( 0.954 ) 4.076 *** ( 1.062 ) Acquirer Size 0.12 2 *** (0.02 3 ) 0.12 2 *** (0.02 4 ) 0.12 5 *** (0.02 1 ) 0.12 2 *** (0.02 2 ) M&A Experience 0.008 ** (0.004) 0.008 ** (0.004) 0.008 ** (0.004) 0.00 7 ** (0.004) Law_ Reg_ Dis Out 0.385 ** ( 0 .1 67 ) 0.385 ** ( 0 .1 77 ) 0.393 ** * ( 0 .1 64 ) 0.385 ** * ( 0 .1 62 ) Country_ Risk_ Dis Out 5.079 *** (1. 640 ) 5.079 *** (1. 618 ) 5.156 *** (1. 542 ) 5.190 *** (1. 558 ) Size Out 0.10 0 ** (0.0 56 ) 0.10 0 ** (0.0 60 ) 0. 094 * (0.06 0 ) 0. 094 * (0.06 1 ) M&A Experience Out 0.2 76 ** * (0.1 09 ) 0.2 76 ** * (0.1 06 ) 0.2 78 ** * (0.1 11 ) 0.2 78 ** * (0.1 11 ) Out 7.935 ** * (2. 158 ) 7.935 ** * (2. 240 ) 7.962 ** * (2. 085 ) 7.983 ** * (2. 174 ) Culture distance 0.03 3 ** (0.01 8 ) 0.03 3 ** (0.01 8 ) 0.03 3 ** (0.017) 0.03 5 ** (0.01 8 ) Industry Relatedness 0.0 47 (0.1 05 ) 0.0 47 (0.1 15 ) 0.0 27 (0.11 3 ) 0.0 31 (0.11 3 ) Target Status 0.2 29 (0.18 7 ) 0.2 29 (0.18 9 ) 0.2 12 (0.18 7 ) 0.2 04 (0.18 7 ) Stake Sought 0.005 *** (0.002) 0.005 *** (0.002) 0.005 *** (0.002) 0.005 *** (0.002) Cash 0.2 56 * (0.1 85 ) 0.2 56 * (0.1 87 ) 0.2 43 * (0.1 80 ) 0.2 41 * (0.1 81 ) Disclosed 0.64 3 *** (0.1 16 ) 0.64 3 *** (0.1 26 ) 0.6 38 *** (0.12 3 ) 0.6 35 *** (0.12 3 ) Attitude 0.0 57 (0.2 17 ) 0.0 57 (0.2 29 ) 0.0 57 (0.2 07 ) 0.0 77 (0.2 08 ) Year 0.0 75 *** (0.02 3 ) 0.0 74 *** (0.02 2 ) 0.0 75 *** (0.02 2 ) 0.0 80 *** (0.02 2 ) India 0. 672 *** (0.2 82 ) 0. 672 *** (0.2 72 ) 0. 669 *** (0.2 44 ) 0. 743 ** (0.2 44 ) Cross Industry Effects Yes Yes Cross Country Effects Yes Log likel i hood Value 999.161 999.160 99 7.122 99 6.936 Pesudo R 2 0.07 5 0.075 0.076 0.076 Note: *** : p <0.01, ** : p<0.05, * : p<0.1, one tail test. Numbers in parentheses are standard errors. N = 1640 due to some missing data in variables of country law and regulation distance and the size of acquirer. Model 1: The covariance matrix of error term in this model is assumed to be correlated within the same country and within the same industry. Accordingly, the standard errors reported in Model 1 are adjusted standard errors. Model 2: The covariance matrix of error term in this model is assumed to be correlated within the same country, within the same in dustry, and within the same year. Accordingly, the standard errors reported in Model 2 are adjusted standard errors. M&A experience is measure as global experience.
107 Table 3 6. Distribution of Cross border M&As by Developed Countries Inbound M&A (N = 1212) Outbound M&A (N = 454) Number Percentage Number Percentag e United States 430 35.479% United States 204 44.934% Singapore 125 10.314% United Kingdom 59 12.996% United Kingdom 114 9.406% Singapore 34 7.489% Canada 101 8.333% Australia 31 6.828% Japan 99 8.168% Germany 27 5.947% France 64 5.281% Canada 21 4.626% Australia 57 4.703% France 24 5.286% Germany 53 4.373% Italy 13 2.863% Netherlands 41 3.383% Belgium 9 1.982% Switzerland 35 2.888% Japan 7 1.542% Sweden 34 2.805% Spain 7 1.542% Italy 15 1.238% Netherlands 7 1.542% Belgium 8 .660% Switzerland 5 1.101% Finland 7 .578% Finland 5 1.101% Spain 6 .495% Denmark 4 .881% Norway 5 .413% New Zealand 3 .661% Denmark 5 .413% Norway 3 .661% New Zealand 5 .413% Portugal 2 .441% Austria 4 .330% Sweden 1 .220% Luxembourg 2 .165% Greece 1 .220% Greece 1 .083% Luxembourg 1 .220% Portugal 1 .083% Austria 1 .220% Table 3 7. Distribution of Cross border M&A by Industry Inbound M&A (N = 1212) Outbound M&A (N = 454) Number Percentage Number Percentage Acquirer Industry Agr. and cons. product 80 7% 1 5 3% manufacturing 548 45% 209 46% Utilities and transportation 81 7% 14 3% Wholesale and retail trade 54 4% 3 1% Financial services 175 14% 20 4% Tourism and misc. service 274 23% 193 43% Target Industry Agr. and cons. product 93 8% 37 8% manufacturing 539 44% 153 34% Utilities and transportation 85 7% 19 4% Wholesale and retail trade 67 6% 24 5% Financial services 130 11% 20 4% Tourism and misc. service 298 25% 201 44% Note: According to the distribution of cross border M&As in our sample, we create region dummies for North America and Asia ; Europe is the omitted region . Also we create industry dummies for manufacturing (SIC 20 39), utilities (SIC 40 49), wholesale and retail (SIC 50 59), financial (SIC 60 69) and service (SIC 70 99) sector ; agriculture and construction (SIC 01 19) is the omitted industry .
108 Table 3 8 . Estimation Results Using Several Sub Samples Separate Samples by M&A Direction Separate Samples by Emerging Countries Inbound M&As Outbound M&As India Only China Only Law and Regulation Distance 0.104 (0. 114 ) 0.173 ( 0.306 ) 1 . 494 ** (0. 695 ) 0.168 ( 0.152 ) Country Risk Distance 4.118 ** * (1. 1 0 4 ) 2.289 * (1. 570 ) 3 . 593 ** (1. 573 ) 5. 186 *** (1.6 29 ) Acquirer Size 0.1 25 *** (0.02 2 ) 0. 203 *** (0.0 59 ) 0.05 9 (0.04 8 ) 0.15 3 *** ( 0.028 ) M&A Experience 0.0 077 ** (0.004 4 ) 0.32 *** (0.12) 0.00 2 ( 0.006 ) 0.01 0 * ( 0. 00 7 ) Law_Reg_Dis Out 0.895 ( 1.085 ) 0.181 ( 0.414 ) Country_Risk_Dis Out 5.602 *** (2. 009 ) 5.296 * ( 3.906 ) Size Out 0.1 25 * ( 0 .08 4 ) 0.20 1 ** ( 0 .1 02 ) M&A Experience Out 0. 297 *** ( 0 .11 1 ) 0.2 35 ( 0.388 ) Out 6.664 * * ( 2.976 ) 6.962 * ( 4.629 ) Culture distance 0.0 25 * (0.018) 0.07 6 * (0.04 7 ) 0.1 08 ** ( 0 .05 1 ) 0.0 15 (0.022 ) Industry Relatedness 0. 094 (0.12 8 ) 0. 105 (0.22 1 ) 0. 011 ( 0 .16 8 ) 0.08 1 ( 0.167 ) Target Status 0. 232 (0.21 6 ) 0.5 01 * (0.36) 0.0 63 ( 0 .25 5 ) 0. 588 ** ( 0. 3 12 ) Stake Sought 0.008 *** (0.002) 0.000 6 (0.004) 0.0035 (.00 28 ) 0.007 *** ( 0 .003) Cash 0.6 63 *** (0.22) 0. 490 * (0.3 12 ) 0.25 1 ( 0 .2 48 ) 0.2 37 ( 0. 2 73 ) Disclosed 0. 578 *** (0.14 7 ) 0.9 85 *** (0.2 37 ) 0. 499 *** ( 0.170 ) 0.8 40 *** ( 0. 1 80 ) Attitude 0.1 55 (0.2 27 ) 0. 927 (0.7 47 ) 0.0 26 ( 0.321 ) 0.13 5 ( 0 .3 13 ) Year 0. 094 *** (0. 025 ) 0. 022 (0. 048 ) 0.05 7 ** ( 0.027 ) 0. 097 ** ( 0 .04 3 ) India 0. 809 *** (0. 268 ) 0. 043 (0. 733 ) Log likelihood Value 7 22.320 26 5.021 48 3.876 50 4.630 Pesudo R 2 0.0 86 0.07 9 0.05 2 0.09 9 Sample Size 1188 452 817 823 Note: *** : p <0.01, ** : p <0.05, * : p <0.1, one tail test. Numbers in parentheses are standard errors ; M&A experience is measured as global e xperience
109 CHAPTER 4 STOP VOLUNTARY DISCLOSURES? Retailers routinely disclose strategic information regarding their businesses, stores, and products to analysts, investors, and the general public. Interestingly, a signif icant portion of such information is not mandated by regulations. Literatures in marketing, finance, and accounting propose several reasons why firms voluntarily disclose such information. Unfortunately, empirical support is scant and cannot disentangle th ese hypotheses. We contribute to this literature by focusing on why retailers stop making one of such disclosure that helps distinguish organic growth and expansion growth comparative store sales (CSS). The CSS disclosure stoppage by some retailers and c ontinuance by others provides a natural experiment setting that helps empirically tease apart four alternate theories of information and strategy, not possible otherwise. Using a unique dataset collected over 17 years for 164 retailers, we competitive spillover as opposed to information postponement, loss of metric relevance, or long term focus. Moreover, in contrast with extant research , we provide evidence that investors respond favorably towards disclosure stoppage when they can rationally attribute it to the reduced competitive spillover effect . We discuss implications of our findings and opportunities for further research. 4 .1 Intro duct ory Remarks existed for at least one year. Since the 1970s, retailers have been announcing monthly CSS in the first Thursday of each month (Bird and McGee 1997). The m onthly CSS d
110 aggregate also serves as a barometer of the retail industry. The strength of the retail industry is closely watched by economists to decipher consumer confidence and ther efore signs of economic bubble or recovery. The decades old tradition of monthly CSS disclosure, however, had a crack in the late 1990s following household names of Best Buy and Bed Bath & Beyond announcing a break from the tradition. In recent years, cessation of monthly CSS has become a symphony, including 39 apparel and accessory stores, 1 5 home goods companies and 16 department stores withdrawn from the reporting routine . The phenomenon is intriguing because the cessation occurs at a time when inves demand for more frequent financial reporting (e.g., real time reporting) is on the rise and when advanced technology is available for firms to quickly close their books and aggregate information from various retail locations. 1 This study examines t he reasons for the cessation as well as changes in firm performance from the period before the cessation to the period after it to better distinguish the alternative explanations for the phenomenon. Based on disclosure theories, extant empirical research, and the media, we propose four alternative but not mutually exclusive explanations: (1) poor performance, (2) loss of relevance, (3) short termism, and (4) the proprietary disclosure costs. 2 E conomic theory shows that when disclosure is voluntary, the good news is typically disclosed and bad news is withheld (Dye 1985). Houston, Lev, and Tucker (2010) examine the hotly debated phenomenon 1 Section 409 of the Sarban es 2 rivals increase production to chip away the f
111 to provide earnings predictions for future quarters and find that poor performance is a major reason f or the stoppage. Retailers may stop monthly CSS announcements after lackluster monthly CSS numbers . Some critics argue that CSS has lost its relevance because of a lack of industry standards in calculating monthly CSS and because a quarter rather than a mo nth might be a more meaningful period to evaluate a business (Plitch 1997). In addition, s ome critics of monthly CSS disclosure argue that the disclosure forces retailers to manage the business on a monthly basis, for example, shifting the timing of promot ions to pump up CSS in one month at the expense of future months (Plitch 1997; Bird and McGee 1997). Breaking from the term strategies. Moreover, economic disclosure theories demonstrat e that proprietary disclosure costs may deter firms from disclosing information that is useful to capital market investors (Verrecchia 1983; Wagenhofer 1990). We examine the above explanations by using a sample of 79 public U.S. retailers , who ceased to v oluntar ily provide monthly CSS information during the period between 1995 and 201 2 . Together, w e address three broad research questions: first, what determines the stoppage of monthly CSS announcement? Second, how do investors respond to such stoppage? Spe cifically, an event study is conducted to estimate the abnormal change in the stock price for our sample of retailers around the date when the stoppage information was released to the public. Third, how does the product characteristics (e.g., trendy vs. staple) influence the m otive for the stoppage decision?
112 Issues addressed in this paper are important for a number of reasons. First, there is little empirical evidence (if any) on the magnitud e of the economic impact relating to the redu ced internal information to the market. This paper attempts to fill this gap by estimating the impact of stoppage on the stock price, a common measure of firm value. The evidence of the change of the stock price associated with less frequent communication would give the retailer top management an idea on how it affects their wealth, and therefore, the importance of managing disclosure policy effectively. Second, the information content studied is unique in the literature. Compared to the other types of mark eting information (e.g., R&D, advertising, sales forces, new products etc.), the internal information studied her e (monthly CSS) is more formal and institutionalized, as manifested by a consistent reporting routine and a persistent reporting format. Severa l recent studies (e.g., Houston et al. 2010 and Chen et al. 2011) in the financial literature indicate that a poor performance is one of the primary reasons for the cessation of information such as managerial forecasts of quarterly earnings. Differently, the monthly CSS serves as the most important metric in the retailing industry to gauge each retailer s operating health condition (e.g., Cole and Jones 2004). The different information content, demand side information with much higher disclosure frequency and accuracy, might give rise to new reasons for the stoppage. Furthermore, our study within the single industry (e.g., retailing) can get rid of many unobserved industry level factors confounding previous results. Third, our paper also contributes to the stream of literature related to the new product preannouncement (e.g., Eliashberg and Robertson 1988, Robertson, Eliashberg and Rymon 1995, Lilly and Walters 1997, Sorescu et al. 2007). Previous results, largely based on the survey and laboratory data , ar e subject to
113 the response bias and potenti ally omitted variable biases. The monthly CSS announcement can be seen as firms preannounce overall operating ability in the established order prior to the required quarterly earning s announcement. Hence, by invest igating the firms shifting from pre announc ing to silencing, this paper adds to the existing l iterature by offering much clear er datasets for the theory testing purposes. In addition, complementing previous work, we document how the shareholder characteris tics (e.g., the percentage of dedicated institutional investors), stock performance (e.g., the volatility of stock returns) and information characteristics (e.g., the volatile of information and value relevance) affect the preannouncement decision. The r est of the paper is organized as follows. We first discuss the theoretical background and develop a conceptual framework concerning those driving forces behind the monthly CSS cessation. We then introduce the empirical model, present our data and empirica l analyses. As a further test of our theory, we present an event study to test the impact of the monthly CSS cessation announcement on the stock market. A regression analysis is also used to identify factors that influence the magnitude of change in market value. We conclude with discussions of contributions, managerial implications and opportunities for the future research. 4 .2 Theory a nd Hypotheses 4.2.1 Comparable Store Sales in Retailing Industry Retailing is the most prominent industry that reports monthly sales. As a commonly used measure of overall performance for the retailer, the comparable store sales (CSS) considers sales growth in stores that have been open for at least one year. Often disclosed on the first Thursday every month, this number not only puts almost every major retailer s performance under the spotlight but also serves as a barometer
114 for the economy by showing changes in consumer s expenditure 3 . T he idea of announcing monthly CSS on a preset day was the brainchild of a then Donal dson, Lufkin & Jenrette analyst named Jerry Gallagher in the early 70s (Bird & McGee, Wall Street Journal 1997). Since then, most of the major retailers have been sticking to report ing monthly figures. The news is eagerly awaited by the financial community , the media and many others. The performance results are clearly mov ing retail stocks and analysts characterize companies as winners and losers . An example of monthly CSS voluntary disclosure through news press is provided in Figure 4 1. The CSS metric h as been shown to be the most important measure to gauge retailer performance (Cole and Jones 2004). The importance of this measure could be manifested as follows: this metric provides additional information to investors through helping them to distinguish between sales growth that arises from the growth in the asset base (e.g., opening new stores) and store growth that arises from the existing assets. The sales growth that arises from the new store openings might be dangerous since this way of driving up sales is more costly and would not last long (Cole 2004). The SEC believes the CSS in retailing industry is helpful in giving the investor an opportunity to look at the company through the eyes of management (SEC ). But meanwhile, retailers have a great deal of discretions in deciding how to provide this type of information. This metric has gain ed increasing attention from researchers (Cole and Jones 2003, Lundholm, McVay and Randall 2010, B uskirk 2011) and most of the related 3 The strength of the retail industry is closely watched by economists to decipher consumer confidence and therefore signs of economic bubble or recovery.
115 studies have focus ed on the quarterly disclosure of this metric. For instance, Cole and Jones (2003) examine the value relevance of the disclosure of quarterly CSS information and find that revenue changes (i.e., CSS g rowth) contribute to incremental explanatory power in future revenue , earnings and contemporaneous stock returns. Lundholm, McVay and Randall (2010) present a sales forecasting model and show how to use the historical series of sales, store information and comparable store growth rates to estimate the sales rates of new stores and of existing stores. Results show that forecast errors generated by the model are comparable to analyst revenue forecasts. By comparing the firms who provide monthly disclosure to those who provide only quarterly disclosure in 1996, Buskirk (2011) finds that the practice of regularly providing monthly revenue disclosures is not associated with reduced information asymmetry. In contrast, he finds that the frequent disclosure is ass ociated with higher information asymmetry among investors. The results suggest that more frequent disclosure attracts the sophisticated investors to acquire information , which in turn increases the information asymmetry among investors. A recent trend on t he breaking up of American retailing monthly CSS reporting intrigues this empirical research on why firms discontinue previous reporting routine and how investors respond to such discontinuity on the cooperate practice. Notably, it also provides much clean er dataset to test the impact of voluntarily disclosing monthly CSS within the retailing industry on the managerial relevance and value for retailing managers and shareholders. Different from Buskirk (2011), we focus on the disclosure policy change; and th is shift of disclosure policy within a single group the monthly disclosure group helps to remove those unobserved group effects in the empirical test
116 between those who report monthly a nd quarterly. Hence, this data offers good opportunities to address t he following related managerial dilemma : monthly CSS disclosure routine encourages investors and analysts to emphasize on the information transparency, which also fosters myopia managerial behavior of meeting short term sales goal that is detrimental to fi rms long term growth and value creation (e.g., A.Steigrad, 2011, http://www.retailgeeks.com/ ). In the following part, we conceptually develop four potential drivers, which are most likely to influence the monthly CSS disclosure b ased on disclosure theories, extant empirical research, and the media . We also propose several variables which could directly test those e xplanations. Table 4 1 lists various reasons for the cessation decision by retailers. 4.2.2 T h e Poor Performance Explanation A large body of literature investigates incentives for managers to disclose internal information to the outsiders. Based on advers e selection (Grossman 1981), theory predicts that managers disclose all value relevant information. However, the above full information disclosure situation does not happen either when the disclosure is costly or when investors do not know whether managers receive such private information (Dye 1985). In other words, under the above case, managers tend to release good news and suppress bad news. Under the multi period game situation, firms may even develop a reputation for reticence (Grub 2006). Assuming th at investors believe a firm to be more likely to receive private signals after it made disclosures in prior periods, Grubb shows that firms would conceal all news, either good or bad, if their disclosure not only directly inform investors of the firm valu e but also indirectly hurt firm value. An example of the latter is disclosure of poor performance, which could generate excessive speculation,
117 anxiety, and negative publicity in both the media and the capital market (Houston et al. 2010). Therefore, we con jecture that the past poor CSS performance might increase incentives to curtail the monthly disclosure of monthly CSS. Several empirical studies on voluntary disclosure literature also strongly indicate an increasing tendency to disclose in good times and , by implication, a decreasing tendency to disclose when performance deteriorates. For instan ce, Houston et al. (2010) find that poorly performed firms tend to stop the quarterly earning guidance announcement. Miller (2002) finds that the frequency of volu ntary disclosures increases when firms perform well and the manager shifts to disclose less as the firm approaches earning declines. Miller and Piotroski (2000) reveal that the frequency of voluntary forward looking disclosure is much higher for firms wit h stronger and more persistent earnings during turnaround periods. In terms of CSS information, as pointed out by industry practitioners: monthly CSS not only makes the retailers remain transparent to the investor communities but also signals that retaile rs are on the right pace of doing business; only poorly performed firms tend to hide information. , we have the following conjecture on the association between retailers monthly CSS performance and the cessation decision . Prediction 1: Retailers with poor monthly CSS performance are more likely to stop monthly sales disclosure. 4.2.3 The Loss of Relevance Explanation Based on the demand side of internal information, Dye (2001) argues that managers have incentives to make voluntary dis closures which investors find useful. However, the practice of monthly CSS disclosure might lead to fewer investors or analysts demands for this type of information.
118 Fi rst, lack of industry standards might potentially lead to the reporting biases. On this note, s ome stock analysts complain: retailing managers will do all kinds of things to dress up the store sales such as moving around promotional events. They also doubt: due to the lacking of standards, financially, some retailers will consider st ores at the first anniversary, while others wait until the 14th month. Some retailers even exclude renovated, expanded or relocated stores while others do not. In short, the practice of reporting monthly CSS information might be viewed as haphazard and i nconsistent, which might directly lead investors to regard this type of information as useless which cause the stoppage decision. Second, even without such reporting bias (or those retailers report truthfully), the monthly sales metrics might still be a l ess meaningful way to reflect the retailers true performance and tell little of a retailer s overall performance trend. "I think we will see a big shift toward retailers reporting quarterly comps instead of monthly," predicted Eric Beder, a specialty reta il analyst at Brean Murray Carret & Co. " there is too much noise associated with reporting monthly." Beder and his peers mention that calendar shifts , inconsistent weather and changes in receipt flow can all skew monthly comps results and create stock vola tility that has little correlation to a company's actual vitality. Instead, " q uarterly comps are usually reported at the same time as earnings, which gives investors a lot more information to digest and [with which to] evaluate the business. I think this eliminates some of the irrational weight that can be placed on a monthly comp," Poole said . "If investors don't understand how components behind impact a company's monthly comp, they do not have a true meaning of the business and could be very
119 surprised when quarterly earnings are reported," said Ann Poole, retail analyst at Nolenberger Capital Partners . frequent update on business and reflects customer loyalty, it also creates great volatility in th said. This volatility has led many analysts to advocate shifting to quarterly comps. Hence, based on the above arguments, we expect that a higher stock volatility could lead to the stoppage decision. Moreover, the difference in the coefficient of variation between the quarterly and monthly CSS figure can also reflect the quality of the monthly CSS figure to information users such as investors and analysts. If monthly CSS itself is t oo volatile as a metrics, it has a lower quality and hence not that useful to the external users. For instance, the predictability and persistence, defined as the ability of earnings to predict itself, as a measure of earning quality is based on the view t hat an earning number that tends to repeat itself is of a high quality. This view is similar to the idea, implied by Dechow and Schrand (2004), that a high quality earning 4 number is representative, that is, a good indicator of future earnings. Hence, we e xpect, that the monthly CSS is of poor quality as a performance measur e if the coefficient of varia tion of monthly CSS is too large; or the difference number 5 between quarterly and monthly disclosure is lower for the stoppers than for the maintainers befor e the stoppage. 4 The high quality earning number, as defined (Dechow and Schrand (2004)), will: 1) ref lect the current operating performance; 2) it will be good indicator of future operation performance ; 3) it will accurately annuitize the intrinsic value of the firm. 5 For the persistence argument , it assumes that persistent earning improves the ability of earnings to capture value relevant information. But consider a company that has reported $5 of earnings in each of the past few years. The earning number is persistent and has low volatility, but it may not be relevant for measuring anticipated growth , which reduces its usefulness for valuation; however, since we also have quarterly CSS numbers , and difference between this two metrics could help to mitigate t he issue caused by the above explanation .
120 Finally, the association between stock price and the monthly CSS update could directly reflect the information value of monthly CSS to the users. Overall, because the market would require more monthly CSS information when this number has mo re value relevance (Hutton (2005)), we expect that the higher the value relevance of this number, the less likely the retailer is going to stop the monthly disclosure. Prediction 2: Retailers are more likely to stop monthly CSS disclosure when the figure l oses its relevance. Specifically, retailers will be more likely to stop if the stock volatility is higher; the difference between quarterly CSS and monthly CSS coefficient of variance is lower; the association between stock price and such information is lower . 4.2.4 The Short Termism Explanation As mentioned in the previous section, much of the current debate related to monthly CSS figures centers on whether this practice engenders managers with a short term focus. Moreover, in justifying their decisions to give up monthly CSS disclosure, retailers have widely cited desires to focus on the long term performance as the key reason behind their stoppage decisions Hence, we expect that the long run orientation could be another explanation for the cessation . The reasons are as follows: first, one month period is a relatively short period to judge a health condition for a firm. The pressure to meet short term sales targets precipitates actions that destroy long te rm shareholder value, such as over emphasizing on the price promotions (e.g., to boost the sluggish sales at the end of the month) and the cancellation of marketing campaigns due to the short term budget. Furthermore, the monthly CSS disclosure not only f acilitates the retailing managers to
121 have short term orientations but also makes the customers become more strategic and price sensitive due to more frequent and timely price promotion before the monthly reporting period. Second, retailer managers thems elves might face loss of reputation , credibility, and even employment issues if they are unable to deliver a good performance, especially under the reporting routine of publishing th ose figures with all industrial peers. Third, the CSS disclosure practice also induces investors to place too much emphasis on the figure of CSS, resulting in extreme stock price drops when retailers publish poor CSS performance ; thereby shifting the manager focus from the retailer s fundamentals to bottom line sales f igures. For instance, some industry analysts (Plitch 1997) mention that in retailing, where same store sales determine stock price to a greater degree than earnings, retailers are forced to manage the business on a monthly basis essentially. We document manager s short termism tendency in two aspects. This myopia behavior refers to sub optimal underinvestment in long term projects for the purpose of meeting short term goals. Hence, we expect the retailers with lower brand equities , caused by underinvest ment in information system, channel relationship , consumer loyalty program and private brands etc. , are also more likely to focus less on the long run performance, hence being less likely to stop the monthly sales disclosure. Second, the investors short t erm mentality of the companies will also influence the manager s short term behavior. Previous studies show that institutions that invest in firms with the intention of holding substantial ownership blocks over a long horizon have
122 strong incentive to incur the cost of explicitly monitoring managers and ensure the firm does not cut profitable R&D to meet short term earning goals (Dobrzynski 1993). Hence, we expect that the percentage of shares held by long term institutional investors will mitigate the mana ger s myopia behavior and motivate the firms to stop the monthly disclosure if the long run orientation explanation holds. Compared to other types of institutional investors such as transitional and quasi indexers investors , the dedicated institutional inv estors, usually taking larger stakes in firms and having a low portfolio turnover, are also found to trade the stock based on the long term value (Bushee and Noe 2000) and serve as an active monitoring role in corpo rate governance. On the contrary, the tra nsient institution s are characterized as having high levels of portfolio turnover and diversification. These characteristics reflect the fact that the transient institution tend s to be short term focused investors with little interest in long term capital appreciation (Porter 1992). Fi nally, the quasi indexes group is characterized as having a low portfolio turnover a nd highly diversified holding. Porter (1992) claims that the presence of quasi indexes exacerbates incentives for myopic investment behavior because their passive, fragmented ownership leads them to gather a little information on the company and provides little incentive to monitor managers. Based on the above argument, we have the following hypothesis: Prediction 3 : Retailers focusing on the long run performance are more likely to stop the monthly CSS disclosure. Specifically, retailers are more likely to stop if brand equities are higher; the ownership percentage by dedicated institutional investors is higher.
123 4.2.5 The Information Spillover Explanation Finally, we propose the potential information spillover effect brought by the monthly sales disclosure as another alternative reason for the stoppage decision. According to Verrecchia (1983), the release of a variety of accounting statistics about a firm (e.g., sales, net income, cost of operation) may be useful to competitors, shareholders or employees in a way which is harmful to firm s prospects even if the information is favorable . In the retailing industry, which is often crowded with ho mogeneous firms, we expect that this propriety cost tends to be high. For example, all retail ing stores are under intense pressure to create item assortments that cannot be C Penny says (Harvard Business Review, 2012 May) . Furthermore, retailing is the industry that heavily uses the sales promotions as major marketing tools. Various innovative promotion strategies (i.e., consumer loyalty programs, time limited coupons) to b oost sales have been developed . However, many invests in variou s sales promotions (e.g., sweep stakes, coupons, time limited price discounts, free gifts or samples, special events, displays, membership rewards, consumer directed promotions and so on) are fo und to be even easier to imitate than the simplest new products (Gelb et al., MIT Sloan Management Review, 2007). In recent years, technology and the internet have also greatly reshaped the landscape of retailing industry. For instance, regarding to thos e fashion stores, statistics (Time magazine, August 6, 2012, P18) show that the idea to shelf time for a new product is now 3 4 weeks compared with the past 6 9 months. A p ercentage of merchandise that is very trendy was 20 30% but now is 100%. Stores are swapping out big orders of staples like T shirts for smaller, more frequent batches. Hence, the
124 monthly CSS information, to some extent, could greatly complement the extensive new information of innovative designs and promotions, and even serves as a poten tial marketing test tool to inform the acceptance of new marketing strategies and products, thus facilitating the quick imitation process. For example, based on more accurate demand information , rivals right reactions (i.e., supplying the right amount ) towards the newly launched products might increase the future competition. Moreover, timely disclosure will also induce timely imitation, which further reduces the monopoly rent of new products. We expect that such information spillover tends to be more severe for retailers with the following characteristics. First, the retailers carrying more trendy products face higher risks of being imitated. Since most of the trendy products demands are with greater uncertai nties and are harder to predict , the deman d information has a greater value for rivals to redesign their own marketing strategies. Second, retailers , who operate in a highly dynamic market face more competitions and are with greater risks of being imitated . For example, sales outcomes related to the profitability of certain products can result in the entry of new competitors, hence lowering the economic rents for the existing retailers in the future. Third, however, retailers with higher brand equit ies are fearless of being copied since some of th e intangible element , such as channel relationship, branding of private labels and consumer loyalty could impede the copy process. Finally, due to the rivals strategic use of firm s new promotion or new product information based on those timely sales repo rt s , the retailers with higher expenditures on sales and marketing campaigns of new products bear a higher development cost (the sunk cost) in the new product designs and innovati ve selling, which encourage
125 them to stop the monthly CSS sales disclosure. Mo reover, retailers with more sales promotion are more likely to face more competitive environment. Hence, we have the following hypothesis: Prediction 4 : Retailers facing more severe information spill over are more likely to stop the monthly CSS disclosure. Specifically, retailers are more likely to stop if products are trendier; more new entrants to the market; branding equities are lower; more sales efforts are made. 4 .3 Empirical Analysis 4.3.1 Model We apply a Probit Regression model with the monthly CSS stoppage decision as a dependen t variable to estimate how the stoppage decision is affected by each variable. Specifically, the probability of stoppage decision is assumed to be a normal function of e xploratory variables such as variables related to each explanation and other control variables: where ( 4 1) Here, is a vector of explanatory variables, with being a vector of parameter estimates ; represents each retailer within our sample and is the cessation month. The empirical specification for the function is given by ( 4 1). Notation denotes probability and is the cumulative distribution function of standard normal distribution.
126 In (4 1), the variable represents the monthly CSS figures prior to the stoppage month for each retailer. The variables are the stock return volatility, the investors reaction towards the monthly CSS information, and the difference between the CV ( coe fficient of variation ) of quarterly and monthly CSS information prior to the month The variables represent the brand equity, the percentage of block holders and the percentage of dedicated institutiona l investors for the retailer in months prior to the stoppage month And denote the product characteristics , the selling efforts and competition environments prior to the stoppa ge. The vector controls include all other variables, which we will explain in detail in the next section. 4.3.2 Data In order to have complete data, we collect all the monthly CSS information between 1995 and 2012 by hand; the quarterly CSS data before 2002 has also been hand collected; and Compustat offers detailed quarterly CSS data after the year 2002. There are sever al characteristics of this dataset: f irst, it covers complete CSS information (both monthly and quarterly) for all the public retailers in the US between the year 1995 and 2012 ; s econd, it contains extensive retailing formats (e.g., SIC code 5000 6000), and the retailing categorical information are based on SIC code information from Compustat . We exclude those firms (restaurant chains) between 5800 and 5900 for SIC code. Table 4 2 summarizes the distributions of retailers across each sub retailing forma ts (i.e., in terms of SIC 2 code) within our sample. We use the following approach to identify retailers that maintained and those that stopped providing monthly CSS. We refer to each quarter during the sample period
127 between 1994Q1 6 and 2012Q4 as an event quarter and to the preceding three quarters and the event quarter as the pre event period while the subsequent four quarters as the post event period. We define the stoppers as the retailers that consistently issue monthly CSS for the first three p re event quarters, provide last monthly disclosures in the fourth pre event quarter, but offer no monthly CSS for any of the four post event quarters. Those that provide monthly CSS for all the 4 pre event quarters and continuously provide the monthly CSS for 4 post event quarters are defined as maintainers . Following the above identification method, we totally identify 116 stoppers across 7 sub industries. To obtain much cleaner data, we apply the following rules to further screen our sample. First, we exclude those stoppers who went bankruptcy immediately after the stoppage; s econd, we exclude those stoppers who either went to private or had been acquired right after the stoppage. Based on the above process, we totally exclude 35 stoppage events and fi nally we are able to identify 81 stoppers between the year 1995 and 2012 as the sample of the stoppage group, which are distributed across the following sub industry: warehousing home improvement (52), department stores (SIC 53), convenience food stores (5 4), car stores (SIC 55), apparel and accessory stores (SIC 56), home product stores ( SIC 57) , and miscellaneous stores (SIC 59). Table 4 3 summarizes the sample selection procedure. Due to missing variables related to some independent variables, the final stoppage sample size of the empirical testing would be between 77 and 81. 6 The earliest stoppage we identify is finish line In c, who stopped monthly CSS disclosure on March 1995.
128 In addition , for the control group, we have identified 1655 main taining cases for monthly CSS disclosure while the stoppage event happened. Those firms would continuously report monthly information for over four quarters after the stoppage event. Table 4 4 summarizes the distribution of stoppage events across the time, and the total number of firms who report on the monthly basis at the beginning of each calendar year. 4.3.3 Variables Month st op . The dependent variable in our empirical analysis is the stoppage decision of monthly CSS disclosure, which is coded as eithe r 1 or 0. The number one stands for the stoppage decision , and zero otherwise. CSS_performance. To measure the CSS performance of retailers, w e employ the averaged monthly CSS figure across 24 months prior to the stoppage month. The sales information is ha nd collected via Factiva. Stock _v ol. Following (Krishnaswami and Subramaniam 1999), we calculate the stock volatility as the standard deviation of the market adjusted daily stock returns in the one year period prior to the stoppage month. Data is collected from CRSP. Diff_ CV. W e measure this as the standard deviation of quarterly CSS subtracts the standard deviation of monthly CSS in the past 24 months prior to the cessation. Here, e ach standard deviation is also divided by the absolute value of the means over the calculation period so that the two standard deviation m easures are comparable. Coeff. Follow ing the literature (Dechow, Ge and Schrand 2009), w e estimate the following regression: , where the stock r eturn on the day of and the day after the monthly CSS announcement minus the retail industry stock index over these two days . Moreover, is the CSS figure in
129 the same month for the previous year to allow the seasonality. Overall, we use the monthly sales observations in the past 24 months. T he CSS measure is also divided by The coefficient, , finally capture s how info rmative the monthly CSS numbers are to the More informative component s will have a higher Brand _ e quity. We adopt the two years averaged Tobin s Q value prior to the stoppage event as the proxy for the brand equity, which usually includes the brand image, consumer loyalty, customer service, information and supply chain systems, exclusive partnerships with suppliers, expert system etc.. Marketing literature (i.e., Mizik and Jacobson 2009, Srinivasan and Hanssen s (2009), Rao, Agarwal and Dahlhoff (2004), Simon and Sullivan (1993)) uses market to book ratio or Tobin s Q to measure the unmeasured contribution of knowledge, goodwill, technology, branding and other intangible assets that a company might possess. Fol lowing Rao et al. (2004), we use the following formula: Tobing s Q=(MVE+PS+DEBT)/TA, where MVE=(share price) (number of common stock outstandings), PS=liquidating value of the firm s preferred stock, DEBT=(short term liabilities short term assets)+book value of long term debt, and TA=book value of total assets. Dedicated_ i nvestor. The variable is measured as the averaged percentage ownership by dedicated institutional ownership relative to total shares outstanding across four q uarters prior to the stoppage quarter. We use the data from Thomson Reuters Institutional (13F) Holdings database and the Bushee s website, in which lists categor ical information of institution al investors in the following three categories: dedicated inves tors, transitional investors and quasi index er investors.
130 Trendy _ p roduct. We code 1 for those stores with trendy products, which include apparel stores (SIC 56), shoe stores (SIC 56), department store (SIC 533) and consumer electronic stores (573); while we code 0 for other types of retailing format. The store classification information (SIC code) is based on Compustat. New _ entries . The average number of new entries, divided by the total number of retailers, within sub industries in four quarters prior to the stoppage event captures the intensities of industry competition . The sub industry classification is based on SIC 2 code. Selling_ Effort . We use the averaged four quarters of SG&A expenditure, divided by total assets, prior to the stoppage event as a pr oxy for the promotion effort exercised by retailers. S everal prior studies in the literature have used SG&A to measure the stock of marketing spendings (e.g., Dutta, Narasimhan, and Rajiv 1999; Mizik and Jacobson 2007). For instance , Dutta, Narasimhan, and Rajiv (1999, p. 556) good proxy for the amount the firm spends on its market research, sales effort, trade promotion Other controls . Information asymmetry is also one determinant for voluntary disclosure in the literature (e.g., Healy and Palepu 1993, 1995). To control for this effect, we use the number of analyst s , which is the maximal number of analyst s estimating the quarterly earning s per share in one year prior to the stoppage event. This measure captures the information asymmetry from the analyst information and the data provided by FirstCall. Following the literature (i.e., Diamond 1985, Harris 1998), the size of the retailer is also included in our model as a control variable and it migh t have a positive or negative association with disclosure. On the one hand, larger firms will have
131 the resources to produce information more easily. On the other hand, more information for large r the firms are generally available to the public, thus substi tuting for some information the management would otherwise release. I use the variable size, measured as . The retailers disclosure might also be influenced by peers practice . We measure this variable, peer, as the percentage o f retailers who report monthly within the stopper s sub industry prior to the stoppage event. Finally, we control for the fixed industry effect and year effect. 4.3.4 Empirical Results We estimate the results of the Probit Regression model. Table 4 5 shows detailed results on determinants of the monthly CSS stoppage decision. For the poor performance explanation , we found that there is a strong negative relationship between the stoppage decision and firm s prior monthly CSS performance ( ). Hence, prediction 1 is supported. This result is also consistent with those from Chen (2011) and Houston (2010), which show that the poor firm performance is often the main rea son for the stoppage decision on voluntary information (e.g., qua rterly earning s forecasts). For the loss of relevance explanation, we found , although that the stock return volatility positively influences on the stoppage decision ( ), the other two testing variables do not significant ly influenc e the stoppage decision. For the stock return volatility, another possible explanation for the positive sign is that the monthly disclosure might attract more transient investors trading the stock around the disclosure date eac h month, thus contributing to higher volatilities of the stock price. Overall, we identify very weak evidence show ing monthly CSS losing value relevance.
132 In terms of the short termism explanation, none of our test variables are statistically significant. This result implies that the focus on the lo ng run is not the major driver i n the reduction of reporting frequency of CSS information. Hence , prediction 3 is not supported. As to the information spillover explanation , we found that there is a strong positive relationship between the number of new entries and the stoppage decision ( ), implying that the competitive concern drives the stoppage decision. The coefficient of selling effort is of the predicted sign and is also close to significant ( ). However, the coefficient of the trendy products is not positively related to the stoppage decision and is not significant. As a result, we f ound some evidence supporting prediction 4 that the demand information spillover due to the fact that more frequent disclosure s could be one potential reason for the cessation decision, although this reason is often unstated as the excuses for the stoppage decision by retailers. For other variables, interestingly, we f ind th at retailer s decision to stoppage is not positively related to the peers decision. On the contrary, an opposite relationship has been found, suggesting that the retaile rs are more likely to disclose monthly CSS when fewer peers are doing so. This findi ng is consistent with the evidence that firms might adopt more frequent disclosure as a way to catch the investors attention . 4 .4 F urther E mpirical E vidence 4.4.1 Market Reaction to the Announcement of Stoppage In this section, we investigate whether retailers that stop providing monthly CSS experience abnormal returns when they reveal their intentions. On the one hand, firms
133 often claim that the reasons for giving up the monthly figure are to avoid useless infor mation as well as managerial attention on the short run performance of the company. If monthly CSS is truly a value decreasing proposition ( e.g., forcing managers to the short run performance ) and the market realizes this, we predict that the announcement period returns should be positive. On the other hand, prior research also suggests a negative relation between fewer disclosure and firm value, which implies that a firm stopping monthly CSS would experience a negative stock price reaction since the invest ors take the stoppage as a signal of deteriorating future performance. To document the market response to stoppage announcements, we analyze 5 day event windows [ 2, 2] where 0 represents the day of the announcement itself and record the associated cumul ative abnormal return (CAR) of the firm s stock. The basic idea is to test for the statistical significance of the average abnormal returns around an even date for a sample of firms experienc ing the same type of firm specific event: the stoppage of monthly CSS disclosure . To be specific, w e use a four factor model estimated over the time horizon [ 150, 10]. In particular , for each firm, to statistically test the abnormal returns during the st oppage period, parameters in ( 4 2) for each stock we re estimated over an estimated period, with durations of 141 trading days. Each estimation period ended two weeks (10 trading days) prior to the date of the announcement. These estimates, in turn, were us ed to determine the daily abnormal returns during the event period: ( 4 2) , w here is the excess return of the retailer at time . is the excess return on the market; (SMB) is the return difference between a portfolio of small and
134 big stocks; (HML) is the return difference between a portfolio of high and low book to market stocks, and (UMD) is a m omentum factor from Carhart (1997), which is the return difference between a portfolio of stocks with high returns in the past year and a portfolio of stocks with low returns in the past year. Based on the coefficients estimated by ( 4 2), we could obtain the abnormal return of the retailers experiencing the stoppage event. We also control the impact of other information during the stoppage announcement period. Within our sample, totally 30 retailers announce d the stoppage decision together with their an nouncement of quarterly earnings. Hence, the announcement period return cou l d capture both the impact of stoppage announcement as well as the earning s surprises, which are also shown to influence abnormal returns. To control this confound impact, we run the following regression: ( 4 3). Here, is the a bnormal returns derived from ( 4 3 ) for each retailer during the decision announcement period , indicates whether the retailer disclose this information together with the quarterly earning s announcement; is measured as the difference between the actual earning s per share and the most recent forecasted number, divided by the stock price prior to the date. And captures t he abnormal return experienced by the stoppers. Table 4 6 shows the results of the event study. Specifically, we found that a significantly positive sign on , suggesting that investors reward the decision of monthly CSS stoppage. This result contradicts to the poor performance
135 explanation. The coefficient of intercept indicates that the average market reaction, after controlling for the surprises in earnings , is also a statistically significant positive result of 0.022 (p<.05). What is the economic implication behind this result? By multiplying the average capital capitalization (one date prior to the stoppage date) with the averaged abnormal returns, we fou nd an economic impact of 114 million dollars bro ught by the stoppage decision. To sum up, different from previous findings in the literature related to quarterly earning s guidance, our results have shown that the market, in fact, rewards the firms stoppag e decision on the monthly CSS. However, this increase in the firm value might be caused either by the investors belief i n firms desire to focus on the long run performance or the reduced information spi llover effect, thus less threat from rivals taking advantage of monthly CSS information . In order to formally distinguish the above two beliefs, we directly test the relationship between abnormal returns and those variables capturing the above two concerns . 4.4.2 Drivers of Abnormal Returns To further find out the reasons behind the increased value after the stoppage, a regression model is developed to explain the abnormal returns. The model is given by: (4 4 ) Base d on (4 4 ), if the long run focus is the main reasons believed by the investor for the stoppage, we should expect that the coefficients be positive ; h owever, if the information spillover is the main concern for the stoppage, we should expect that
136 be positive and be negative. In addition, we also control for the firm size effect ( ), the time effect ( ) and other firm characteristics such as the reporting history ( ) (e.g., how many months the retailers have been continuously reporting the monthly sales prior to the stoppage) a nd whether the firm adopts a proactive or passive strategy ( ) when revealing the stoppage decision . Within our sample, totally 78 retailers reveal the stoppage decision around the stoppage month through either a preannouncement of their decisions or a silent stoppage . Chen et al. (2011) documented the above practice as either a proactive or a passive strategy for the information disclosure in the product recall context. Among those 78 firms, there are totally 57 retailers preannounce the stoppage decision before the actual stoppage month. The regression results are shown in Table 4 7 . We find a systematic pattern supporting the information spillover explanation . Specifically, we found that the abnormal return is higher for the trendy retailing format ( ) , for the subcategories with more new entries ( ) and for the retailers with more selling efforts ( ). This result further supports that the informat ion spillover concern is the main driver of the stoppage decision o f the monthly CSS figure. For the short termism explanation , we found that the abnormal return is not positively correlated with all three testing variables (i.e., brand equity, and the percentage of institutional investors). On the contrary, we found that brand equity negatively influence the abnormal returns This result, which in fact is the consistent with information spillover explanation, shows that a lower brand equity
137 reduces the imitation cost of the competitors and, hence, facilitates the easiness of copy. Therefore, the market rewards the stoppage decision more when the brand e quity is lower. To sum it up, our results derived from the model (4 4 ) show why the market rewards the stoppage decision. Totally, the retailers with more trendy products, experiencing more new entries within the sub industry , putting more selling efforts and having lower brand equities achieve more rewards for their stoppage decision from the market. 4.4.3 The Moderating Role of Product Format By far, we have identified the key drivers behind the stoppage decision, from both the inves tor and the firm perspective. In order to further verify our results, we run the Probit model ( 4 1) for the stoppage decision separately according to the retailing category: trendy vs. staple format. We expect that the product characteristics could influence the motive for the stoppage decision. Specifically, for those trendy retailers, since fashion product enables more industry dynamics, we expect that th e information spillover effect be higher among this category. Overall, we predict stronger impact s from all those varia bles capturing the competitive information spillover effect. In Table 4 8 , we find a result pattern that is consistent with the above conjecture. Specifically, we found, within the trendy format , the number of new entries and the selling efforts have a str onger positive effect. In addition, the coefficient of brand equity becomes more negative. For the staple format, we find that the information spillover effect becomes a less significant force for the stoppage decision. However, hiding of bad performance e merges as a more significant driver behind the stoppage decision. The reason is as
138 follows ; f or the staple format, due to fewer uncertainties in the future sales, the retailer sales trend s to be more persistent. Since the poor performance tends to have a continuous pattern and induces persistent indirect disclosure costs (e.g., the market responses differently towards good and bad news), those retailers within the staple category are more likely to hide poor performance by adopting a non monthly disclosure practice. For other variables, we find, the coefficient of peers within the trendy format becomes more positive, suggesting that the retailer within the trendy category is willing to differentiate himself more via the information disclosure practice. To sum it up, we find a moderating role played by the retailing format on the motives behind the monthly CSS stoppage decision. Striking differences have been found for the motive of stoppage between the trendy and staple retailing format. Specifically, for the trendy format, the information spillover concern becomes the key force behind the stoppage decision; however, for the staple format, the hiding poor performance becomes the major driver behind the stoppage decision . 4 .5 Summations As a commonly used measure of overall performance within the retailing industry, comparable store sales growth (or the same store sales growth) considers growth in stores that have been open for at least one y ear. This number puts almost every major retailer s performance under the spotlight and serves as a barometer for the economy. However, recently, many major retailers choose to withdraw from the previous monthly disclosure and only report quarterly on the CSS information. This new trend intrigues a hot debate between the retailer managers and investors from W all S treet on the usefulness of monthly same store sales data . We join in the debate and
139 conceptually build up four alternative explanations behind the stoppage decision. Using a unique dataset collected from the retailing industry covering all monthly same store sales information between the year 1995 and 2012, we directly test the determinants and consequence of the stoppage decision empirically . Speci fically, we find that the retailers stoppage decision is largely driven by the serious information spill over effect brought by frequent sales data disclosure. Moreover, we find that investors respond favorably towards the stoppage announcement. Specifica lly, the retailers with trendier products , higher selling expenditures, l ower brand equities and among the category with more new entries experience higher abnormal returns after the stoppage. Last but not the least, retailing category characteristics gre atly moderate those motives for the stoppage decision. The stoppage within the trendy category is largely due to the information spillover effect; however, the stoppage in the staple category is mainly driven by the sluggish performance. 4.5.1 Research Co ntributions First, this paper contributes to the marketing and finance interface literature with a special focus on the value of the monthly CSS announcement . This pape r responds to the call (Hanssen s , Rust and Srivastava 2009) of investigating what kinds of financial information should be disclosed by firms. Since t he CSS figure serves as an important metric gauging the retailer s performance and has been used by investors and retailing managers consistently , the importance of understanding the impact of voluntary disclosure of this figure can never be ignored . As the first paper in this field, we conceptualize totally four potential explanations for drivers of monthly CSS figure stoppage and propose various measures to test those explanations. Furthermor e, we directly document how investors evaluate such stoppage decision. On average, we find
140 that the stoppage increases the market value of the retailers by 2.2%. The average dollar change in the market value is as high as 114 million dollars, demonstrating a significant economic impact from the stoppage decision. Second, this paper also contributes to the new product preannouncement literature. To some extent, the monthly CSS disclosure could be viewed as the firm pre announces his operating ability before the required date (e.g., the date of the quarterly earning s report). Hence, e mpirically , we contribute a unique sample of firms who shift from pre announcing to sil ence . There are several advantages of this dataset. First, by identifying the shifting date , we can more accurately document the impact of the preannouncement. Previous results, largely based on the survey, are subject to the response bias and potentially omitted variable bias. Second, we further broaden the scope of the previous research by in vestigating how those variables at the capital market level such as the shareholder characteristics (e.g., institutional investors), market performance (e.g., stock return volatility) and information characteristics (e.g., the value and volatile of the inf ormation) affect the preannouncement decision. Ignoring these variables might lead to the omitted variables bias and influence the manager s decision making process. Third, we provide further insights on the motivation behind the recent upsurge of firms d iscontinuing the monthly CSS report. Our results shred insights related to the completely different motives according to the retailing products. Overall, although many of these firms argue that monthly CSS forces a short term orientation and impedes long t erm value creation, and many of investors complain the firms hiding bad performance and reducing the transparency to the community through a non disclosure,
141 our tests suggest that, although unstated, information spillover concern is, in fact, the primar y driver behind this stoppage within retailing industry. The recent development of information technology and advances in the inventory replenishment such as quick response systems greatly facilitate such process and make retailers bear huge proprietary co sts when disclose frequent information related to consumer s taste and preferences. Furthermore, the retailing products also greatly influence the manager s decision on how to communicat e the internal information to the capital market. Specifically, for th e retailers with trendy products, we find that the competitive information spill over is the major concern for the frequent information CSS; however, for those staple ones, hiding the poor operating performance strikingly influences the disclosure decisio n. 4.5.2 Managerial Implications Our study should be of interest to CFOs and investor relation s professionals. For those worried about the negative impact from stoppage of providing public information, our findings suggest the reduced corporate social responsibility (i.e., such as the reduction in the frequency of voluntary information disclosed to the public) or discontinuing an existing cooperate practice might not necessarily incur the severe punishment from the investors. Instead, the market rewards the decision on providing less information under certain situations. This implies, in a more dynamic environment , that the CFOs should be less worried about the negative impact brought by providing less frequent information but , should be more discerned about how to release demand side information . Hence, they could design the disclosure policy according to factors , interacting with the competition intense , such as information contents, timing and
142 communication channel; and based on the product characteristics, company s market position and the rivals potential responses towards such information. Our studies help the industry analysts and practitioners to have a better understanding of monthly CSS practice. The firms who stop monthly di sclosure might not necessarily be those wanting to hide poor performance or in a wrong pace of doing business. The product format plays a vital role in discerning the bad one and detecting the real motives behind the stoppa ge decision. Special attention s hould be given to stoppers within the staple category. Further more , although there is a lack of industry standards to regulate the disclosure, quite weak evidence is found showing that monthly CSS figure lose s its value relevance. 4.5.3 Limitations and Fut ure Research This paper subject s to several limitations that provide opportunities for the future research. First, as an initial study in the examination the impact of disclosing monthly same store sales data, we focus only on the retailing industry. It wo uld be more insightful if future studies could examine the restaurant industry, which also provides monthly same store sales information and has its own industry characteristics . Second, for study on the consequence of stoppage, we focus largely on the mar ket reaction towards stoppage. Future research could examine how the information environment is affected by the stoppage, for instance, whether the analysts will have larger quarterly earning s forecast error due to less disclosure frequency. Last but not l east, it would be interesting to investigate how the starting of announcing the monthly CSS information affects the firm and investors. Fo r instance, whether the start of the monthly CSS disclosure would induce the investor s reaction in the different (asy mmetric) way as the impa ct of the stoppage? Future studies could apply the event study methodology to
143 examine the potentially asymmetric impact of stoppage vs. starting announcement of monthly CSS information.
144 Abercrombie & Fitch Reports July Sales Results NEW ALBANY, Ohio, Aug 05, 2010 /PRNewswire via COMTEX/ -Abercrombie & Fitch (NYSE: ANF) today reported net sales of $270.9 million for the four week period ended July 31, 2010, a 17% increase from net sales of $232.3 million for the four week p eriod ended August 1, 2009. July comparable store sales increased 7%. For the fiscal month, total Company direct to consumer net merchandise sales increased 51% to $23.6 million. For the fiscal month, total Company international net sales, including direct to consumer net sales, increased 81% to $47.0 million. For the fiscal quarter ended July 31, 2010, the Company reported net sales of $745.8 million, a 17% increase from net sales of $637.2 million last year. Comparable store sales increased 5% for the q uarter. For the quarter, total Company direct to consumer net merchandise sales increased 50% to $69.0 million. For the quarter, total Company international net sales, including direct to consumer net sales, increased 85% to $133.2 million. Year to date, the Company reported net sales of $1.434 billion, a 16% increase from net sales of $1.239 billion last year. Comparable store sales increased 3% for the year to date period. Year to date, total Company direct to consumer net merchandise sales increased 46 % to $137.7 million. Year to date, total Company international net sales, including direct to consumer net sales, increased 93% to $252.2 million Figure 4 1. An E xample of Monthly Same Store Sales Disclosure
145 Table 4 1 . Cited R easons for the M onthly CSS S toppage Table 4 2 . Distribution of Retailing Format According to SIC Code SIC code Store Format Stoppers Total Number in the Sample 52 Building Materials 3 14 53 Department Store 1 4 635 54 Convenience Store 2 6 55 Car Store 2 40 56 Apparel and Accessory Store 3 6 1004 57 Home Goods Store 1 3 154 59 Miscellaneous Store 1 1 231 Company Cited Reasons Pier 1 Imports shareholders to do meaningful comparisons in sales, Bed Bath and Beyond of large Hancock Fabrics results of changes in the timing of promotion, holidays and West Marine as it removes some of the short term effects that weather Footstar Jennifer Convertable Wilson Leathers As a result of the previous announcement of converting the existing stores into a new studio concept Circuit City variations rather than on the overall business trends that are critical to Talbots brand Walmart term view we take to build shareholder value. We feel this also will reduce the intra period volatility related to events such as calendar shifts. Reporting sales quarterly also places us in line with
146 Table 4 3 . Sample Selection Procedures Sample Selection Process Total Stoppage Identified 116 Those went to bankrupt 13 Those went to private 10 Those being acquired 12 The final sample used for testing the determinants 77 81 Table 4 4 . Distr ibution of Stoppage Events a cross Time Number of Stoppage Event Number of Firms Reporting on Monthly Basis at the Beginning of Calendar Year 1995 2 55 1996 2 67 1997 3 69 1998 2 71 1999 8 79 2000 7 70 2001 1 75 2002 3 82 2003 4 82 2004 1 77 2005 5 71 2006 8 68 2007 9 54 2008 7 44 2009 4 36 2010 3 33 2011 6 29 2012 6 25
147 Table 4 5 . Determinants of Giving Up Monthly CSS Announcement Independent variables Coefficient Parameters Of Interest Predicted Sign Explanation s CSS_perform 0.03 1 *** (0.01 0 ) Poor performance Stock_vol 20 . 053 *** (4. 478 ) + Loss of relevance Diff_CV 0.0002 ( 0.003 ) Coeff 5. 530 (3.62 0 ) Brand_equity 0.000 1 (0.0 08 ) + Shortism Dedicated_invest 0. 645 ( 0.926 ) + Trendy_product 0.0 7 9 (0.24 0 ) Information Spill over New_entry 15 . 305 *** (7.87 0 ) + Selling_effort 1. 180 * (0.660) + Firm Size 0.0 29 (0.056) ? Analyst following 0.00 2 (0.01 0 ) _ Peers disclosure 2. 029 * (1.1 70 ) _ Fixed industry effect YES Fixed time effect YES N 1743 Pseudo R square 0.16 Log likelihood 26 5.416 Note: *** p<0.01, ** p<0.05, *p<0.1, two tailed test Table 4 6 . Consequences of Giving Up Monthly CSS: Stock Reactions towards Stoppage Parameters Of Interest Predicted Sign Model (2) Model (3) Intercept (+) or ( ) 0.014* (0.008) 0.022** (0.01) Quarterly Earning Announce ? .026 (0.020) Quarterly Earning Announce*Surprise + 2.08 (2.44) N 79 79 R Square N/A 0.025 Note: *** p<0.01, ** p<0.05, *p<0.1, two tailed test
148 Table 4 7 . Drivers of Abnormal Returns Coefficient Parameters Of Interest Predicted Sign Explanation Intercept 8.50 (5.90) Quarterly Earning Announce 0.025 (0.023) Quarterly Earning Announce*Surprise 2.50 (2.38) + Brand_Equity 0.013** (0.005) + Shortism Dedicated Investors 0.092 (0.15) + CSS_perform 0.032 (0.021) + Poor performance Trendy Format 0.032* (0.018) + Information Spill over New_entries 2.70* (1.48) + Selling efforts 0.22** (0.11) + Firm Size 0.013** (0.006) Proactive 0.022 (0.022) Reporting history 0.0004** (0.0001) Time trend 0.0042 (0.0029) N 7 7 R square 0.30 Note : *** p<0.01, ** p<0.05, *p<0.1, two tailed test
149 Table 4 8 . Additional Tests: Determinants of Giving Up Monthly CSS in Sub Category Independent variables Trendy Category Staple Category Parameter Of Interest Explanations CSS_perform 0.01 5 (0.01 4 ) 0. 10 *** (0.032) Poor performance Stock_vol 2 3 . 28 *** ( 6 . 78 ) 1 7 . 73 ** (8. 08 ) Loss of relevance Diff_CV 0.0013 (0. 0009 ) 0.22 (0.31) Coeff 3 . 85 * (2. 33 ) 0.491 ( 0.358 ) Brand_equity 0.0 16 (0.0 13 ) 0. 026 (0.0 24 ) Shortism Dedicated_invest 0 . 442 (1. 161 ) 1.98 ( 1.97 ) Trendy_product Information Spill over New_entry 6 0 . 82 ** ( 29 . 67 ) 13.08 (9. 60 ) Selling_effort 2 . 412 ** (1.00 0 ) 0. 306 (1.4 15 ) Firm Size 0. 09 (0.1 0 ) 0.021 (0.08 5 ) Analyst following .01 4 (0.015) 0.0 06 (0.019) Peers disclosure 5. 41 ** (2.4 6 ) 0.6 0 (1. 57 ) Fixed industry effect YES YES Fixed time effect YES YES N 751 487 Pseudo R2 0.22 0.24 Log likelihood 134 89 Note: *** p<0.01, ** p<0.05, *p<0.1, two tailed test
150 CHAPTER 5 CONCLUSION Environ mental uncertainties create new opportunities and challenges for marketers to design suitable marketing strategies. This dissertation consists of three separate essays related to marketing decision making under uncertainties. The first essay investigate s optimal advance selling strategie s when incorporating the buyers behavioral factors (e.g., present biased preference and overconfidence) into the classic advance selling model. When buyer uncertainties (e.g., uncertain about future consumption states) interplay with the above psychological issues, the optimal pricing structure significan tly change s. Moreover, the refund , often a tool to reduce the advance purchase risk, has a new role under the advance selling setting. Lastly, overconfidence might not necessarily benefit the buyer at the expense of buyers. I find a scenario that educating buyers about their tendency on overestimating their ability to be consistent during the multiple decision process might lead to a WIN WIN situation. The second essay exams marketers global market expansion strategy . Specifically, I empirically investiga te what drives the completion of cross border M&As related to emerging markets. Interestingly, I found that different risk sources interact with M&A directions on influencing the completion of cross border M&As. Specifically, I found that firm level risks such as fi rm size, past M&A experience have a stronger positive effect on the completion for the outbound M&As (e.g., companies from emerging marketing to developed market s to acquire firms); h owever, country level risks such as the difference on the regu lation and trade policy have a stronger negative impact on the completion for the inbound M&As (e.g., companies from developed
151 markets to emerging markets to acquirer firms). This result highlight s the importance of understanding different risk sources in managing the firm s global expansion. The third essay investigates how retailers voluntary disclosure policy affects their performance and what determines the retailers choice of such policy. We found that retailer stops the monthly same store sales ann ouncement mainly due to the competitive informati on spillover effect. Second, investors respond positively towards such stoppage decision. Third, the product characteristic moderates the above effect. Specifically , compared to those mainly selling staple products, retailers selling trendy products suffer a stronger information spillover effect and experience higher positive abnormal returns after the stoppage. Our results suggest the product level uncertainties (e.g., fashion vs. sta ple) strongly influence the firm s marketing communication strategy and affect how investors evaluate such communication strategy .
152 LIST OF REFERENCES R. Aguilera and C. Completion: A Relational Champaign. D. Arnold and Q. , Sloan Management Review , 40 (1), 7 20. B. Aybar and F. Aysun ( 2009 ), Border Acquisitions and Firm Value: An Analysis of Emerging Market Multinationals , Journal of International Business Studies , 40 , 1317 1338. J. Bao and E. Alex (2009), Do I nvestment B anks H ave S kills? Performance Persistence of M&A A dvisors, Wor king Paper , Fisher College of Business, Ohio State University . J. Barney (1991), Firm R esource and S ustained C ompetitive A dvantage, Journal of Management , 17, 99 120. C. Beckman, and H. Administrative Science Quarterly , 47, 92 124. D. Bell ( 1982 ), Regret in Decision Marking under Uncertainty , Operation Research. 30(5) , 961 981. D. Bell ( 1985 ), Disappointment in Decision Making under Uncertainty , Operation Research. 33(1) , 1 27. L. Bird, S. McGee (1997) , Heard on the S treet: M onthly S ales D ata L ose V alue for R etailers , Dow Jones Business News, March. A. Bonaccorsi (1992), On the R elationship b etween F irm S ize and E xport I ntensity, Journal of International Business Studies, 23 (4) , 603 635. B. Boone and J. Mulherin (2007), How A re F irms S old? The Journal of Finance , 112(2), 847 875. P. Bordalo , N. Gennaioli and A. Shieifer ( 2012 ), Salien ce Theory of Choice under Risk , The Quarterly Journal of Economics , 1243 1285. S. Burgess and S. Jan Benedict ( 2006 ), Emerging Markets Advances Marketing Science and Pra c tice , International Journal of Research in Marketing , 23 , 337 356. B. Bushee an d C. Noe 202.
153 cy and Inf Working Paper. Ohio State University. L. Capron and H. J ohn (1999), Redeployment of B rands, S ales F orces, and G eneral M arketing M anagement E xpertise F ollowing H orizontal A cquisitions: A R esource B ased V iew, Journal of Marketing , 63 (April), 41 54. Journal of Finance , 52(1), 57 82. R. Chakrabarti, G. S wasti and J. N arayanan (2009), Mars Venus M arriages: Culture and Cross B order M&A, Journal of International Business Studies , 40, 216 236. A. Chari, Q. Paige and T. Linda ( 2010 ), , The Review of Financial Studies , 23(4) , 1741 1770. S. Chen, D. Matsumoto and S. Rajgopal (2011) , Is S ilence G olden? An E mpirical A nalysis of F irms T hat S top G iving Q uarterly E arning G uidance , Journal of Accounting and Economics , 51(1 2), 134 150. Y. Chen and S. Moorthy and Z. Zhang ( 2005 ), Rebates as State Dependent Discounts , Management Science , 51(1) , 1131 1140. M. Cheng, K. Subramanyam, and Y. Zhang (200 7 ) , Earning G uidance and M anagerial M yopia , Working P aper , Columbia University. W. Chu, E. Gerstner and J. Hess ( 1998 ), Managing Dissatisfaction: How to Decrease Customer Opportunism by Partial Refunds , Journal of Service Research , 1(2) , 140 154. C. Cole and C. Jones (2004) , The U sefulness of MD&A D isclosures in T he R etailing I ndustry , Journal of Accounting, Auditing and Finance , 361 388. J. Core (2001) , A R ev iew of t he E mpirical D isclosure L iterature: D iscussion , Journal of Accounting and Economics, 31, 441 456. M. Darrough (1993) , Disclosure Policy and Competition: Cournot vs. Bertrand , The Accounting Review , 68(3), 534 561. D. Davis, E. Gerstner, M. Hagerty ( 1995 ), Money Back Guarantees in Retailing: Matching Products to Consumer Tastes , Journal of Retailing , 71 , 7 22. P. Dechow, Institute. Revi ew
154 S . DellaVigna and U. Malmendier ( 2004 ), Contract Design and Se lf Control: Theory and Evidence , The Quarterly Journal of Economics. May, 353 402. D. Dikova, S. Padma and W. Arjen ( 2010 ), Border Acquisition Abandonment and Completion: The Effect of Institutional Differences and Organizational Learning in the International Business Service Industry, 1981 2001 , Journal of International Business Studies , 41 , 223 245. J. Dobrzynski ( 1993 ), Relationship I nvesting . Business Week , 33 ( 09 ), 68 75. P. Dobson and E. Gerstner ( 2010 ), For a Few Cents More: Why Supersize Unhealthy Food? Marketing Science. 29(4) , 770 778. T. Donoghue and M. Rabin ( 1999 ), Incentives for Procrast inators , The Quarterly Journal of Economics , 114 (4) , 769 816. S. Dutta, O. Narasimhan, S. Rajiv (1999) , H igh technology M arkets: I s M ark eting C apability C ritical? Marketing Science, 18(4), 547 586. Research, 23(1), 123 145. 236. J. Eliashberg and T. Robertson (1988) , P roduct P reannouncement B ehavior: A M arket S ignaling S tudy , 292. J. Emmert, 2011. Transforming the Masses: Why Weight Loss Products Are Thriving In a Bad Economy. Direct Selling News, January, [avail able at http://directsellingnews.com/index.php/view/transforming_the_masses_why_weig ht_loss_products_are_thriving_in_a_ bad_econl ] . S. Fay, and J. Xie ( 2010 ), The Economics of Buyer Uncertainty: Advance Selling vs. Probabilistic Selling , Marketing Science , 29(6) , 1040 1057. Financial Times (2006), China G oes G lobal: Chinese C ompanies T arget E xpansion and Acquisition , 23. G. Fruchter, E. Gerstner ( 1999 ), , Journal of Service Research , 1(4) , 313 323. D. Fudenbe rg and D. Levine ( 2006 ), A Dua l Self Model of Impulse Control , The American Economics Review , 96(5), 1449 1476. W. Henisz and Z. Bennet Business Review, April, 88 96.
155 M. Hitt, M . Dacin, L. Edward, A. Jean Luc, and B. Emerging and Developed Market Contexts: Resource Based and Organizational The Academy of Management Journal , 43 (3), 449 467. P. Holl and K. International Journal of the Economics Business , 3(2), 165 184. C. Homburg and B. Matthias ( 2005 ), acquisitions: How Marketing Integration Affects Post Merger Performance , Journal of Marketing , 69 (January ) , 95 113. E. Hotchkiss, Q. J un and S. W eihong (2005), Holdups, Renegotiation, and Deal Protection in Mergers, Working Paper, Finance Department, Boston College. H. Geert (1991), Culture and O rganization: S oftware of the M ind. New York: McGraw Hill. A. Gelb, D. Andrews and S. Lam (2007) , A S trategic P erspective on S ales P romotions , MIT Sloan Management Review , 48(4), 1 7. S. Gilpatric ( 2009 ), Slippage in Rebate Programs and Present Biased Prefer ences , Marketing Science , 28(2) , 229 238. K. Girald (201 2) , M akeover t he F uture of R etailing? Harvard Business Review , May. S. Gubbi, A. Preet, R. Sougata Ray, M . Sarkar, and C. Raveendra ( 2010 ), International Acquisitions by Emerging Economy Firms Create Shareholder Value? The Case of India Firms , , 41 , 397 418. C. Hassens, R. Roland, S. Rajendra (2009 ), Marketing S trategy and W al l str ee t : N ailing D own M I mpact , Journal of Marketing , 73 (November), 115 118. P. Healy and K. Palepu (2001) , Information A symmetry, C orporate D isclosure, a nd t he C apital M arkets: A R eview of t he E mpirical D isclosure L iterature , Journal of Accounting and Economi cs , 31 , 405 440. P. Healy and K. Palepu (1993) , The E ffect of F F inancial D isclosure S trategies on S tock P rices , Accounting Horizons , 7, 1 11. P. Healy and K. Palepu (1995) , The C hallenge of I nvestor C ommunications: T he C ase of CUC International, Inc , Journal of Financial Economics , 38, 111 141. J. Hou ston, B. Lev, J. Tucker (2010) , To G uide or N ot to G uide? Causes a nd C onsequences o f S topping Q uarterly E arning G uidance , Contemporary Accounting Research , 27(1), 143 185.
156 M. Grubb (2006), University. Determinant of Managerial Earning Guidance Prior to Regulation Fair Research, 22(4), 86 7 914. S. Jain ( 2012 ), Marketing of Vice Goods: A Strategic Analysis of t he Package Size Decision , Marketing Science , 31(1) 36 51. S. Jain ( 2009 ), Self control and Optimal Goals: A Theoretical Analysis , Marketing Science , 28(6) , 1027 1045. A. Jeuland, C. Narasimhan ( 1995 ), Dealing Temporary Price Cuts by Seller as A Buyer Discrimination Mechanism , Journal of Business , 58(3) , 295 308. J. Johnson and T. G erard (2008), Drivers of S uccess for M arket E ntry into China and India, Journal of Marketing , 72 (May) , 1 13. -----, Y. Eden, and P. Krishna (2012) , broad or Paper , http://ciip.som.yale.edu/conferences/msem/agenda . W. Keegan an d G. Mark (2011), Global Marketing , Sixth Edition, Prentice Hall. J. Kraaijenbrink, J. Spender and G. Aard (2010), The Resource Based View: A Review and Assessment of Its Critiques, Journal of Management , 36 (1), 349 372. S. Krishnaswami, V. Subramaniam Corporate Spin 112. T. Kushwaha and P. Relationship between Marketing Efforts, Customer Behavior and Profitability: Emerging vs. Developed http://ciip.som.yale.edu/conferences/msem/agenda . D. Laibson ( 1997 ), A nd Hyperbolic Discounting , The Quarterly Journal of Economics , May, 443 477. D. Laibson ( 1998 ), Li fe C ycle Consumption and Hy perbolic Discounting Functions , European Economic Review , 42 , 861 871. L. Larwood and W. Whittaker (1977), Managerial Myopia: Self Serving Bi ases in Organizational Planning , Journal of Applied Psychology , 62(2), 194 198. A. Lilly and R. Walters (1997) , Toward A M odel of N ew P roduct P reannouncement T iming , Journal of Production Innovation Management , 14, 4 20.
157 Q. Liu and S. Moorthy ( 2007 ), ersus Rebate , Marketing Science. 26(1) , 67 82 . F. Machado and R. Sinha ( 2007 ), Smoking Cessation: A Model of Planned vs. Actual Behavior for Time inconsistent Consumers , Marketing Science , 26(6) , 834 850. Phenomenon and Its Long Term Performance Consequences in the SEO Marketing Science , 26 (3), 361 379. J. Nasiry and I. Popescu ( 2012 ), , Management Science , 58(6) , 1160 1177. The Economist (2010), The W orld T urned U pside D own: A S pecial R eport on I nnovation in E merging M arkets, ( April ) 1 14. The McKinsey Quarterly (2008), 81. The McKinsey Quarterly (2008), How Chinese C ompanies C an S uccess A broad, May , 1 9. K. Meyer, E. Saul , B. Sumon and P. M ike (2009), Institutions, R eso urces, and E ntry S trategies in E merging E conomics, Strategic Management Journal , 30 (1), 61 80. 391. Accoun ting Research, 40(1), 173 204. to Working Paper, Harvard University. S. M oorthy, K. Srinivasan ( 1995 ), Signaling Quality wit h Money Back Guarantees: The Role of Transaction Cost , Marketing Science , 14(4), 442 466. K. Mu e hlfeld, S. P adma and W. A rjen ( 2007 ), and Acquisitions: E vidence from the Newspaper I ndustry, 1981 2000 , Journal of Media Economics , 20 (2), 107 137 . D. North (1990), Institutions, I nstitutional C hange and E conomic P erformance . Cambridge: Cambridge University Press.
158 G. Osland, T. Charles, and Z. Shaoming g International Modes of Entry Ma rketing Intelligence & Planning , 19 (3), 153 161. -----, S. P adma and W. A rjen ( 2012 ), from Failures and Successes: A Study of Acquisition Completion in the Global Newspaper I ndustry, 1981 2000 , Strategic Management Journal , 33, 938 964 . G. Potter ( 1992 Accounting E arnings A nnouncements, I nstitutional I nvestor C oncentration, and C ommon S tock R eturns Journal of Accounting Research , 30 , 146 155. D. Prelec, G. Loewenstein ( 1991 ), Decision Making Over Time and under Uncertainty: A Common Approach , Management Science , 37(7) , 770 786. V. Rao, M. Agarwal and D. Dahlhoff (2004) , I s M anifest B randing S trategy R elated to T he I ntangible V alue of A C orporation? Journal of Marketing , 68, 126 141. T. Robe rtson, J. Eliashberg and T. Rymon (1995) , New P roduct A nnouncement S ignals a nd I ncumbent R eactions , Journal of Marketing , 59(3), 1 15. S. Rosenkranz and W. Working Paper, School of Economics, Universiteit Utrecht . F. Rothaermel, K. Suresh and S . Kevin (2006), "Int ernational Market Entry by U.S. Internet Firms: An Empirical Analysis of Country Risk, National Culture, and Market Size," Journal of Management , 32 (1), 56 82. M. Schweitzer and G. C achon ( 2000 ), t he Newsvendor Problem with A Known Demand Distribution: Experimental Evidence , Management Science , 46(3) 404 420. Q. Shen and X. Ping (2012) "McDonald's and KFC in China: Competitors or Companions?" Working Paper, the Wharton School, University of Pennsylvania. K. Shimizu, H. M ichael , V. D eepa and P. V incenzo ( 2004 ), Cross B ord er Mergers and Acquisitions: A Review of Current Research and R ecommendations for the future , Journal of Internati onal Management , 10 , 307 353. S. Shuba and H. Dominique (2009) , Marketing and F irm V alue: M etrics, M ethods, F indings and F uture D irections , Journal of Marketing Research, 46 (June), 293 312.
159 S. Shugan and J. Xie ( 2000 ), Advance Pricing of Services and Other Implications of Separ ating Purchase and Consumption , Journal of Service Research , 2(3) , 227 239. C. Simon and M. Sullivan (1993), The M easurement and D eterminants of Brand Equity: A Financial Approach, Marketing Science, 12 (1) , 28 52. A . Sorescu, A . Shankar and T. Kushwaha (2007) , P roduct P reannouncements and S hareholder V alue: D M ake P romises Y ou C , Journal of Marketing Research , XLIV, 468 489. Marketing and Firm Value: Metrics, Methods, Findings, and Future Directions 312. A. Steigrad (2011) , Teen Retailers to Opt Out of Monthly Same Store Sales R eports , February. http://www.retailgeeks.com/wpcontent/uploads/2011/02/2011_0202_Te enRetailersToOpt OutOfMonthlySame StoreSalesReports.pdf . K. Sudhir, and T. Debabrata (2012) "Productivity V ersus Transparency: Information O. Svenson ( 1981 ), A nd More Skillful Than Our Fellow Drivers? Acta Psychologica , 47 , 143 148. V. Swaminathan, M. F eisal and H. J ohn (2008), Value C reation F ollowing M erger and A cquisition A nnouncements: T he R ole of S trategic E mphasis A lignment, Journal of Marketing Research, 45 (Feb), 33 47. R. Thaler ( 1999 ), Mental Accounting Matters . Journal of Behavioral Decision Making , 12, 183 206. R Thaler and M. Shefrin ( 1981 ), An Economic Theory of Self Cont rol , Journal of Political Economy , 89(2) , 392 406. H. Tingliang and C. Ying Ju (2012), dvance Selling Without Showing the Price? The Role of Anecdotal Reasoning and Capacity , Working Paper . E. Tsang , and Y. P aul ( 2007 ), I nvestments , Academy of Management Journal , 50 (5) , 1156 1168. The Washington Post (2005), Chinese Drop Bid to Buy U.S. Oil Firm, August 03. The Washington Post (2011), India Frustrates Foreign Investors with Its Unpredictable Tax Policies, April 02. R. Verrechia (1983) , Discre tion ary D isclosure , Journal of Accounti ng Ec o nomics , 5, 179 194.
160 A. W genhofer (1990), "Voluntary Disclosure with a Strategic Opponent," Journal of Accounting and Economics , 12, 341 363. J. Xie and S. Shugan ( 2001 ), Electronic Tickets, Smart Cards, and Online Prepayment s: When and How to Advance Sell , Marketing Science , 20(3) , 219 243. J. Xie and E. Gerstner ( 2007 ), Service Escape: Profiti ng from Customer Cancellations , Marketing Science , 26(2) , 18 30. L . Y uanzhi (2005), Do I nsiders L earn from O utsiders? Evidence from M erger and A cquisitions, Journal of Finance , 60 (4) , 1951 1982.
161 BIOGRAPHICAL SKETCH C henxi Zhou received his Ph.D. in marketing from the University of Florida in 2014. He also holds a bachelor s degree in computer information system and a master s degree in management science and engineering in Tianjin University, China. Chenxi s research interest lies in the marketing/finance interface, behavioral economics. He is also interested in marketing strategies related to online socia l network and big data.