An Empirical Investigation on the Choices of Supply Chain and Operations Management. Vinod Singhal College of Management Georgia Tech

Transcription

1 An Empirical Investigation on the Choices of Supply Chain and Operations Management Vinod Singhal College of Management Georgia Tech Paper for University of Illinois Proseminar Presentation October 28,

2 An Empirical Investigation on the Choices of Supply Chain and Operations Management Executives Many firms are changing the composition of their top management team (TMT) to include a Supply Chain and Operations Management Executive (SCOME). The SCOME as part of the TMT has responsibility for all or a broad spectrum of the Supply Chain and Operations Management functions. This paper empirically examines three issues related to the appointment of SCOMEs. First, we examine the stock market reaction to announcements of appointments of SCOMEs. Second, we shed light on the timing of the decision to appoint a new SCOME by examining whether appointments of SCOME are preceded by poor stock price performance. Third, we identify and investigate the factors that can affect the likelihood of whether the SCOME is appointed from inside or outside the firm. Our empirical analysis is based on a sample of 496 SCOME appointments that are publicly announced during period. In examining the stock market reaction, we find that on the day of the announcement, the market reaction is positive. This reaction is more positive for newly created positions than appointments in existing positions. The stock market also reacts more positively when the SCOME is an outsider than an insider. Additionally, we find evidence of poor stock price performance during the year preceding the appointment of SCOMEs. Much of this poor stock price performance is observed in the subsample where the firm replaces the current SCOME with an outsider. We use probit models to examine the factors that drive the choice of SCOMEs and find that the likelihood of the SCOME being an outsider is higher for (i) smaller firms, (ii) for firms that operate in more concentrated industries, and (iii) for firms that have experienced poor prior performance. 2

3 1.0 Introduction Over the last decade, the complexity and challenges of managing global and outsourced supply chain networks in an intensely competitive environment has increased the prominence and importance of the Supply Chain and Operations Management (SCOM) function. The ability to develop SCOM strategies and execute them effectively can significantly influence the performance of firms. To ensure that the SCOM function is effectively executed, many firms are changing the composition of their top management team (TMT) by creating positions to appoint Supply Chain and Operations Management Executives (SCOMEs). The SCOME, as part of the TMT, has responsibility for all or a broad spectrum of the SCOM functions including operations, procurement, logistics, and distribution. SCOME positions have been created at the Senior Vice President and Executive Vice President levels and these senior executives are commonly referred as Chief Supply Chain Officers (CSCOs). In several firms, SCOMEs now report directly to the Chief Executive Officers (CEOs) or Chief Operating Officers (COOs), a sharp contrast from 15 to 20 years ago when such SCOME positions were not part of the TMT and rarely reported to CEOs or COOs (Grosyberg et al. (2011)). Articles by practitioners provide some evidence about the frequency of SCOMEs appointments to top management teams. An analysis of a survey by CSCO Insights (2008) of 125 manufacturers, retailers, wholesalers from the Fortune 500 list indicates that 25% of these 125 firms have an executive in the firm s TMT who is in charge of the entire supply chain. More recently, a survey by Aberdeen Group (2010) finds that 36% of the 215 respondents indicate that their firm has an executive who plays the role of the CSCO. While there is an extensive body of analytical and empirical literature in both academic and practitioner literature on design, development, and implementation of strategies in SCOM, we know little about executives who have the responsibility to manage all or a broad spectrum of the SCOM functions. For example, we know little about the background of individuals who are appointed as SCOMES, how investors react to the appointment of SCOMES, when firms choose to replace existing SCOMES or create new SCOME positions, and what factors influence the choices made by the firm to appoint SCOMEs by 1

4 promoting internally within the firm or by hiring from outside the firm. This gap is surprising given that SCOM strategy is a key part of corporate strategy and the TMT is the main driver of corporate strategy. Furthermore, the effectiveness of a CEO is associated with the executives who are part of the TMT (Marcel (2009)). Appointing individuals to the TMT is among the most critical decisions made by a CEO. As an initial step towards developing more systematic knowledge about SCOMEs, this paper empirically examines three issues. First, we examine the stock market (or investor) reaction to the announcements of appointments of SCOMEs. Our sample consists of announcements of replacements of existing SCOME positions as well as appointments of newly created SCOME positions. We hypothesize that newly-created SCOME positions will be viewed positively by investors and test whether the stock market reaction is consistent with this prediction. We also examine whether the stock market reacts differently to hiring SCOMEs from outside (referred to as outsider) or promoting internally within the firm (referred to as insider). Hiring an outsider or insider has their own advantages and disadvantages and the stock market reaction provides an indication of which strategy is valued more by investors. Our motivation to examine the stock market reaction to appointments of SCOMEs is to provide an assessment of the value creation potential of the change in composition and governance structure of the TMT by the inclusion of SCOMEs. More importantly, it also provides an assessment of whether the stock market values the top-level focus placed on the SCOM function. Furthermore, our work complements research that examines the stock market reaction to appointments in TMT of other senior executives. Much of this research has focused on CEO appointments (see Kind and Schlapher (2010) and Huson et al. (2004) for a review of the literature). However, studies on stock market reaction to non-ceo appointments are limited and include the study of Chief Financial Officers (CFOs) (Mian (2001)), Chief Information Officers (CIOs) (Chatterjee et al. (2001), and Chief Marketing Officers (CMOs) (Nath and Mahajan (2008), and Boyd et al. (2010)). We contribute to this limited literature and compare and contrast the stock market reaction to appointments of SCOMEs with appointments of non-ceo top management appointments in other functional areas. 2

5 Second, by examining whether appointments of SCOME are preceded by poor stock price performance this paper sheds light on the timing of the decision to appoint a new SCOME. Several studies have shown that CEOs are replaced following weak financial performance (see for example Denis and Denis (1995), Parrino (1997), Furtado and Rozeff (1987), Warner et al (1988)). Collectively, these studies suggest that performance related disciplinary mechanisms are at work in replacing CEOs. A natural question that arises is whether similar performance related disciplinary mechanisms pertain to appointment of executives to non-ceo top management positions. Third, we examine the factors that drive the choice of SCOMEs. In appointing SCOMEs, firm could choose to promote an executive from within the firm or hire an executive from outside the firm. We provide evidence on factors that affect the likelihood of whether the SCOME is an insider or outsider. More specifically, we hypothesize why factors such as firm size, homogeneity of the firm s industry, level of concentration of the firm s industry, and performance of the firm in the period prior to the appointment of the SCOME affect the choice of appointments from inside or outside the firm. This analysis provides empirical evidence on the practices followed by firms with respect to appointment of SCOMEs. These findings also provide guidance and suggestions to firms on some factors to consider in making SCOME appointments. Our empirical analyses are based on a sample of public announcements of 496 SCOME appointments over a ten year period from 2000 to Our overall results on the stock market reaction show that on the day of the announcement, the mean (median) reaction is 0.25% (0.05%). The stock market reaction is more positive for newly created positions. Specifically, the mean (median) market reaction is 0.68% (0.22) for newly created positions and 0.01% (0.01%) for old positions. Further, the stock market reacts more positively when the SCOME is an outsider than an insider. The mean (median) market reaction for is 0.41% (0.05) for outsider appointments and -0.11% (0.03%) for insider appointments. We find evidence of poor stock price performance during the year preceding the appointment of SCOMEs. The mean (median) stock price performance is -8.61% (-3.74%) during the year preceding the appointment of SCOMEs by outsiders. We use probit models to analyze the factors that influence the choice of an insider 3

6 versus outsider SCOME. We find that the likelihood of the SCOME being an outsider is greater for firms that are smaller, operate in more concentrated industries, and have experienced poor prior performance. The rest of the paper is organized as follows. The next section develops our hypotheses on the stock market reaction and factors that affect the likelihood of hiring from inside or outside the firm. Section 3 describes the sample collection. Section 4 presents the empirical results on the stock market reaction to appointments of SCOMEs. Section 5 presents the findings from the probit analysis on the factors that affect the choice of appointments from inside the firm or outside the firm. Section 6 concludes the paper. 2.0 Hypotheses 2.1 Hypotheses about the stock market reaction to SCOME appointments Creating a new SCOME position that is part of the TMT or reports to the TMT sends a signal to investors about the importance and prominence a firm places on its SCOM activities. It elevates the role of SCOM in the firm and indicates to the market that the firm is making a commitment to improving the effectiveness of SCOM. Besides this external signal, internally the creation of a new SCOME position can help inform other members of the TMT about the challenges, opportunities, and potential of SCOM activities, which in turn can help build awareness and cooperation among various functions. Since corporate strategy is largely driven by the TMT, appointments of SCOMEs provide an opportunity to articulate the vision, role, and contribution of SCOM in the development and execution of corporate strategy. This link between SCOM activities and corporate strategy is particularly important as SCOM decisions can affect a significant portion of the firm resources. For example, SCOM has a direct impact on cost of goods sold, which can often account for 40% or more of revenues in many firms. The effectiveness of corporate strategy is also influenced by the integration and alignment of various functional strategies. The presence of SCOMEs in the TMT can ensure that the relationships of SCOM strategies with other functional strategies are discussed, and adjustments are made to ensure integration and alignment of various functional strategies. As part of the TMT, SCOMEs have the power to influence corporate decisions as it relates to SCOM, and can make the business case for investments and resources to build capabilities that increase the effectiveness of SCOM. 4

7 In short, given that the importance and prominence of SCOM has increased in the recent years and the above discussion on newly-created SCOME positions, our first hypothesis, stated in alternate form, is: H1. The stock market will react positively to announcements of newly created SCOME positions. Our second hypothesis examines whether the stock market reacts differently to hiring outsider or insider SCOMEs. To the best of our knowledge, there does not exist any literature that provides theory and evidence directly related to SCOMEs that we can use to develop this hypothesis. However, there is literature on executive appointments, successions and turnover that we can use to build our hypothesis. There are several benefits of hiring an outsider. The outside SCOME brings new experience, knowledge, and perspective of how SCOM is managed in other firms and industries, which could provide new insights and best practices about SCOM (Boeker (1997), Guthrie and Datta (1997)). He\she can more objectively assess the situation since he\she does not have a stake in the decisions made in the past. Moreover, he\she is not likely to be obligated or committed to follow the past policies and investments decisions, and may be more willing to challenge the status quo and change existing strategies and practices (Finkelstein and Hambrick (1996), Peteraf and Shanley (1997)). Being new to the firm, the outside SCOME can also start fresh in building relationships with other senior executives without the burden of past conflicts and disagreements, if any. Additionally, hiring an outsider may send a strong signal to investors and other internal and external stakeholders that the firm is serious about SCOM functions and activities (Friedman and Singh (1989)). Alternatively, there are also a number of benefits of hiring an insider. These benefits are associated with the SCOME having firm-specific knowledge about the markets, technology, systems, and internal networks (Harris and Helfat (1997) and Vancil (1987)). Internal promotions can also facilitate loyalty and boost employee morale (Howard 2001) as the existing employees may view internal promotions as rewards for good performance and opportunities for advancing in the firm. Firms have better information about the capabilities and performance of their internal candidates (Harris and Helfat (1997) and Zajac (1990)), which reduces the chances of hiring an individual not suited for the position. Internal promotions 5

8 may be also better if the value of continuity and stability of existing strategies is high and if the value of existing social networks is high (Ocasio 1999). The empirical evidence on the stock market reaction to outsider or insider executive successions is somewhat mixed. Kind and Schlapfer (2010) summarize the results of 10 studies that examine the stock market reaction to CEO successions. In seven of these 10 studies, the stock market reaction is more positive (or less negative) for outsiders than insiders. Mian (2001) in his study of CFOs and Chatterjee et al. (2001) in their study of CIOs also find that the stock market reacts more positively when the new appointee is an outsider than an insider. Considering the above discussion and the empirical evidence on the stock market reaction to top management changes, we take the position that H2. The stock market will react more positively when the new SCOME is an outsider than an insider. 2.2 Hypotheses on the likelihood of insider or outsider SCOME Next we develop our hypotheses on the effects of firm size, industry homogeneity, industry concentration, and prior performance of the firm on the choice to appoint an insider or outsider SCOME. Larger firms are likely to have a more global, dispersed and complex supply chains than smaller firms. To coordinate these supply chains, larger firms are also more likely to have mid-level and junior executives who have experience in managing various aspects of supply chains. For example, larger firms may have directors or vice presidents managing different parts of the supply chains such as manufacturing, distribution, sourcing, purchasing, and warehousing. These executives provide a pool of potential candidates for the top-level SCOME position, and some of them may have the qualification and experience to be appointed as the SCOME. Given this internal pool, the benefits of hiring an outsider must be balanced with the search costs involved in scouting for an outsider as well as the risk that the outsider may not mesh well with the culture of the firm. Any outside appointee must acquire firmspecific human capital. This will be more challenging and time consuming in larger firms given the greater scope, complexity, and idiosyncrasies of supply chains in larger firms. Furthermore, when a qualified pool of internal candidates is available within the firm (more likely in the case of larger firms 6

9 than smaller firms), hiring an insider not only ensures a smooth transition to the new role but it also sends a signal to mid-level and junior executives that opportunities for advancement and promotions exist within the firm. This signal can motivate and encourage junior executives to stay with the firm, thereby reducing the cost and risk associated with turnovers. Finally, larger firms are more likely to pay attention to succession planning (Shen and Cannella (2003)). They may have designated a group of candidates as potential heirs to be appointed as the top SCOME in the future or serve as a backup in case the incumbent SCOME needs to be replaced. Given this, larger firms are likely to give preference to an insider over an outsider. Research on appointments of CEOs provides evidence that larger firms select a greater proportion of insider CEOs than smaller firms (Warner et al. (1988) and Furtado and Rozeff (1987)). Given the above discussion we hypothesize that: H3. The likelihood of hiring outside SCOMEs will be negatively related to firm size. Even if a firm has pool of potential inside candidates for appointments as SCOMEs, it is likely to consider potential outside candidates before appointing a SCOME. However, outsiders are likely to have less firm-specific human capital than insiders. Compared to outsiders, it is easier for the firm to judge how well insiders have performed in their careers as firms have better and more reliable information to judge the performance of insiders. Given these issues, firms may be more inclined to hire insiders. However, Parrino (1997) argues that the importance of firm-specific human capital and the performance measurement issues in the decision to hire outsiders will vary significantly across industries and can depend on the homogeneity of the industry. Parrino s (1977) argument of industry homogeneity applies to hiring an outsider but within the same industry. Firms operating in a more homogenous industry are likely to compete in similar product markets, using similar inputs and technologies, and similar supply chain structures. Thus, hiring outsiders in more homogenous industries can mitigate the concern about the lack of firm-specific human capital associated with outsiders. Performance evaluation of outsiders in a more homogenous industry can be more reliable because changes in factors such as input prices, technology, and competitive environment are likely to have similar effect on all firms in a more homogenous industry. Thus, the errors in judging the relative performance of outsiders are likely to be 7

10 less of a concern in a more homogenous industry. The above discussion leads to the following hypothesis: H4. The likelihood of hiring outside SCOMEs will be positively related to industry homogeneity. A firm s choice to hire an outsider SCOME, particularly outside its own industry, is likely to be influenced by the level of industry concentration. A highly concentrated industry is one where a single firm or few firms dominate and account for a significant portion of the industry sales. To augment the internal pool of candidates, a dominant firm in a highly concentrated industry is likely to consider candidates from other dominant firms in the same industry as such firms are often similar in terms of structure and operating environment. Since the number of such firms may be few in a highly concentrated industry, the pool of potential applicants from the same industry will be quite limited. Given this, a dominant firm may broaden the pool by considering candidates outside their industry. In a highly concentrated industry the SCOM expertise of the industry is likely to reside among the few firms that dominate the industry. This may pose a hiring challenge for the non-dominant firms if they plan to appoint a SCOME from their own industry. This challenge for a non-dominant firm in the industry to hire from a dominant firm in the same industry arises because of the prestige, visibility, and perks associated with being part of a dominant firm in the industry. Given this, a non-dominant firm may broaden the pool by considering candidates outside the industry. Based on the above discussion our hypothesis is: H5. The likelihood of hiring outside SCOMEs will be positively related to industry concentration. The choice of hiring an insider or outsider SCOME can be affected by the performance of the firm in the period prior to hiring the new SCOME. Poor prior performance could suggest that the existing strategies and policies in SCOM are not delivering the expected performance, and continuation with the status quo will not help. Reversing the trend in poor performance may require a change in SCOME leadership. Even if a firm has pool of internal candidates who may be capable of assuming this leadership 8

11 role, a firm may not want to appoint an insider as he/she helped develop and implement the strategies and policies that have contributed to the poor performance. In such cases, hiring an outsider might be more attractive. An outsider can bring new and fresh perspectives on running the SCOM function, and will have no vested interest in continuing the past strategies and policies; he/she can be more objective in evaluating the value of past strategies and policies. On the other hand, good prior performance indicates that existing strategies and policies are working well and the firm should not change its course. Hiring an insider will ensure continuation of existing strategies and policies. It may also be viewed by mid-level and junior executives as a reward for good performance, which can help the morale and retention of these executives. Our next hypothesis is: H6. The likelihood of hiring outside SCOMEs will be negatively related to performance in the period prior to hiring. 3.0 Data and sample characteristics The sample consists of firms that appoint SCOMEs to existing or new positions. Sample firms are collected from announcements made in business publications and newswires. To generate our sample, we use a preliminary set of key words to collect a small sample of announcements about SCOME appointments. We read these announcements to identify additional phrases and words that are commonly used to announce SCOME appointments and the proximity of the key words to each other. The final set of key words includes chief or president or director or head within ten words of supply or procurement or sourcing or manufacturing or logistics or distribution or purchasing. We search the headlines and lead paragraphs of all announcements in the The Wall Street Journal, Dow Jones News Service, PR Newswire, and Business Wires during and download announcements that meet the search criteria. For an announcement to be included in the sample, we use the following criteria. (1) The appointment in the announcement is related to some aspects of the SCOM function such as manufacturing, operations, supply chain, procurement, purchasing, logistics, or distribution. 9

12 (2) If an announcement includes two or more simultaneous personnel changes, then the SCOME appointment in the announcement was not included in the sample. For example, if the announcement mentions that in addition to the SCOME appointment the firm is also appointing a new CIO, this SCOME appointment will not be included in our sample. The reason for this exclusion is to reduce the possibility of conflicting events driving the stock market reaction. (3) Only those appointments where the SCOME is part of the TMT or where the SCOME reports to a member of the TMT are included in the sample. This criterion attempts to focus the sample to critical high-level appointments that are also likely to be of interest to investors. Some announcements clearly indicate that the appointee is part of the TMT or reports to a member of the TMT. In cases where such information is not available from the announcement, we examine the 10-K reports, annual reports, and other filings with the Securities and Exchange Commission to determine whether the appointee is part of the TMT or reports to a member of the TMT. (4) To be included in the sample, the firm must have stock returns information available on the University of Chicago s Center for Research in Security Prices (CRSP). Based on the above criteria, the sample consists of 496 SCOME appointments. Examples of some appointments are: The Talbots, Inc. today announced that it has named Gregory Poole Executive Vice President and Chief Supply Chain Officer. In this newly created position, Mr. Poole will oversee the global manufacturing, sourcing, transportation and distribution centers for the Talbots and J. Jill brands, reporting directly to Talbots President and Chief Executive Officer, Trudy F. Sullivan. Business Wire, 17 June Under Armour, Inc. today announced the appointment of James Calo as the company's first Chief Supply Chain Officer. The new position is designed to further strengthen Under Armour's executive management team as the company continues its expansion in international and domestic markets. Mr. Calo will report to Kevin Plank, Chairman and Chief Executive Officer of Under Armour, Inc. In his role as Chief Supply Chain Officer for Under Armour, Mr. Calo will 10

13 be responsible for managing the global functions of production planning, sourcing, development, logistics and the distribution house. Business Wire, 24 October Dean Foods Company announced today the appointment of Gregg A. Tanner, Executive Vice President and Chief Supply Chain Officer, effective November 5. Reporting directly to Gregg Engles, Chairman and Chief Executive Officer, Tanner will lead the effort to optimize Dean Foods manufacturing and other supply chain systems. PR Newswire, 26 October The title of Chief Supply Chain Officer (CSCO) is quite extensively used in the practitioner literature to generically refer to the executive responsible for all or nearly all of the SCOM functions. While some SCOMEs in our sample have the title of CSCO, many have other titles such as Senior Vice President, Executive Vice President, Vice president, and President. Some also hold the additional title of Chief Procurement Officer or Chief Purchasing Officer. SCOMEs who are part of the TMT generally report to the CEOs or COOs. SCOMEs who report to a member of the TMT generally report to COOs, CFOs, Senior Vice Presidents, or Executive Vice Presidents. Panel A of Table 1 presents statistics on the sample based on the most recent fiscal year completed before the date of the SCOME appointment. The median observation represents a firm with market value of equity of $990.2 million, total assets of $ million, sales of $ million, and net income of $19.8 million. Panel B of Table 1 gives the number of appointments by year. Nearly 18% of the appointments are announced in 2001 and 4% in 2009, with each of the remaining years accounting for about 10% of the appointments. The sample is diverse in terms of industry representations with firms from 51 (167) distinct two-digit (three-digit) standard industrial classification (SIC) codes. In approximately 81% of the sample, the SCOME is part of the TMT and in the remaining 19% of the sample the SCOME reports to a member of the TMT. Announcements often provide background information on the new SCOME and the nature of the appointment. In 93% of the appointments (463 of the 496 appointments), the new SCOME is a male. 71% of the sample (352 of the 496 appointments) contains information about the educational background of 11

14 the new SCOME. The highest education degree for 38% of the new SCOMEs is a Bachelor s degree, for 54% it is a Master s degree, and for 8% it is a Ph.D. 47% have an MBA and 65% have a science, math, and/or engineering degree. The mean (median) work experience of the new SCOME at the time of appointment is 23 years (22 years). The evidence indicates that firms have a strong preference to appoint outsiders as SCOMEs. In 70% of our sample (346 appointments) the SCOME is an outsider. The preference for outside SCOMEs is much greater than that for outside CEOs. Based on CEO turnover studies, the percent of new CEOs that are outsiders ranges from 11% to 36% with an average of 22% (see the studies mentioned in Table 1 of Kind and Schlapher (2010)). Preference for outside SCOMEs is also greater than that for outside Chief Financial Officers (CFOs) but is similar when compared to the preference for outside Chief Marketing officers (CMOs). In his study of 2,227 CFO successions during , Mian (2001) finds that 50% of the new CFOs are outsiders. Based on a study of 88 CMO successions during , Boyd et al. (2010) find that 73% of the new CMOs are outsiders. In nearly 35% of the sample (or 171 cases), the SCOME is appointed to a newly created position. Mian (2001) reports that 12% of the CFO positions are newly created in the sample of 2,227 CFO successions during , whereas Boyd et al. (2010) report that 66% of the CMO positions are newly created in the sample of 88 CMO successions during Stock market reaction to SCOME appointments This section describes the methods used to estimate the stock market reaction to SCOME appointments, presents the results of the market reaction, and compares the results to appointments to other non-ceo top management appointments. 4.1 Methods for estimating the stock market reaction abnormal returns We use event study methodology to estimate the stock market reaction to announcements of SCOME appointments. The stock market reaction is commonly referred to as abnormal returns. Event study methodology offers a rigorous approach to estimate abnormal returns associated with specific events by 12

15 controlling for market-wide and other factors that influence stock returns (see Brown and Warner 1985, and MacKinlay 1997 for a review of this methodology). The abnormal returns are an estimate of the percent change in stock price associated with an event. All announcements that we use in creating our sample first appear in Dow Jones News Service, or PR Newswire, or Business Wires. These announcements indicate the time when the information is publicly released. We use the time of release of information to determine a one-day event period. If the SCOME announcement is released after 4:00 PM Eastern Standard Time (EST), then the announcement date is set to the next trading day to account for the fact that investors cannot act until the next trading day on the information contained in the announcement. If the announcement is made before 4:00 PM EST, then no adjustment is necessary to the announcement date. The announcement day is the one-day event period that we use for measuring abnormal returns. Calendar days are translated into event days where the announcement date is Day 0 in event time, the next trading day is Day 1, the trading day before the announcement is Day -1, etc. We estimate abnormal returns using two different models: the Four-Factor Model and the Market Model. The Four-Factor Model is based on the three factors identified by Fama and French (1993) and the momentum factor identified by Carhart (1997). The Four-Factor Model posits a linear relationship between the stock return and the four factors over a given time period as: (1) where R it is the return of stock i on Day t, is the intercept of the relationship for stock i, R ft is the riskfree return on day t, R mt is the market return on Day t (i.e., the return on the market portfolio), SMB t is the small-minus-big size portfolio return on day t, HML t is the high-minus-low book-to-market portfolio return on day t, UMD t is the past one-year winners-minus-losers stock portfolio return on day t. and is the error term. The Market Model is a single factor model and posits a linear relationship between the stock return and R mt, the market return, over a given time period as: 13

16 (2) where is the slope of the relationship for stock i with respect to the market return. We estimate the expected return for each sample firm using data from a 200 day estimation period that begins on Day 210 and ends on Day 11. We end the estimation period two weeks (10 trading days) prior to the event day to shield the estimates from the effects of the announcement and to ensure that any non-stationarity in the estimates is not an issue. In estimating the parameters we require that a firm must have a minimum of 40 stock returns during the estimation period of 200 trading days. Similar estimation periods are common in the literature (e.g., Hendricks and Singhal 1997, Sharma and Lacey 2004). Using ordinary least squares regression over the estimation period of 200 trading days we estimate the parameters,,,,, and (the variance of the error term ) for the Four-Factor Model (equation (1)) and parameters,, and for the Market Model (equation (2)). The abnormal return for firm i on Day t is the difference between the actual and the expected return. For the Four-Factor Model: (3) For the Market Model: (4) The mean abnormal return for Day t is given by: (5) where N is the number of announcements in the sample. To test the statistical significance of the mean abnormal return in equation (5), each abnormal return A it is divided by its estimated standard deviation to yield a standardized abnormal return. Since the abnormal returns are assumed to be independent across events, we have from the central limit theorem that the sum of the N standardized abnormal returns is approximately normal with mean 0 and variance N. Thus, the test statistic TS t for Day t is calculated as: 14

17 (6) The cumulative abnormal return (CAR) for a given time period [t 1, t 2 ] is: (6) The test statistic TS e for a multiple day period is derived in a manner similar to that for a single day. (8) We use both parametric t-tests to determine the statistical significance of the mean abnormal return as well as non-parametric Wilcoxon signed-rank tests on the median abnormal return. Unless otherwise noted, we report one-tailed p-values for all tests. 4.2 Event period abnormal returns Table 2 presents the abnormal returns using the Four-Factor Model and the Market Model for the announcement day (day 0 or the event period). The results are reported for the full sample and various subsamples. We are unable to compute the abnormal returns for 40 out of the 496 firms because some of the firms did not have the minimum 40 days of stock returns data during the estimation period and/or did not have stock returns on the day of the announcement. Since the results from the two models are very similar, we focus our discussion on the results from the Four-Factor Model. Where possible, we also compare the stock market reaction to appointment of SCOMES to non-ceo top management appointments such as CFOs, CMOs, and CIOs. To make it easier to compare the stock market reaction to announcement of non-ceo top management appointments across different functions, Table 3 summarizes the results for SCOMEs, CFOs, CMOs, and CIOs, which we use in our discussion below. Panel A of Table 2 presents the abnormal returns for the full sample of new SCOME appointments. The mean abnormal return for the event period is 0.25%, significantly different from zero at the 2.5% level. The median abnormal return is 0.05%, insignificantly different from zero. The stock market seems to react positively to appointment of new SCOMEs. The mean abnormal return of 0.25% is higher than 15

18 the mean abnormal return for non-ceo top management appointments in other functions. Mian (2001) reports that the mean abnormal return for new CFO appointments is -0.05%, insignificantly different from zero, and Boyd et al. (2010) reports that the mean abnormal returns for new CMO appointments is 0.003%, insignificantly different from zero. Panel B of Table 2 presents the abnormal returns for the newly created and existing positions. The mean (median) abnormal return for the newly created position is 0.68% (0.22%), significantly greater than zero at the 2.5% (10%) level. This is consistent with our hypothesis that the stock market will react positively to newly created positions. In contrast, when the new SCOME is appointed to an existing position, the mean (median) market reaction is 0.01% (0.01%), both insignificantly different from zero. The difference in the stock market reaction to newly created and existing positions is significantly different from zero at the 10% level. It is useful to compare our results for appointments of new and existing SCOMES with other non- CEO top appointments (see Table 3). In the case of CFOs, the mean stock market reaction to newly created (existing) positions is 0.41% (-0.11%). Chatterjee et al. (2001) study the stock market reaction for only newly created Chief Information Officers (CIOs). They find that the mean stock market reaction based on a sample of 96 positions is 1.16%. Boyd et al. (2010) do not report results for newly created or existing CMO positions separately. Panel C of Table 2 presents the abnormal returns for outsider and insider SCOMEs. As mentioned earlier, about 70% of SCOMEs are outsiders. The mean (median) abnormal return for outsider SCOMEs is 0.41% (0.05%), significantly different from zero at the 2.5% (10%) level. In the case of insider SCOMEs, the mean (median) abnormal return is -0.11% (0.06%), insignificantly different from zero. The difference in the stock market reaction to outsider and insider SCOMEs is significantly different from zero at the 10% level. This provides support for our hypothesis that the stock market will react more positively to appointment of outsider SCOMEs when compared to insider SCOMEs. This is consistent with the results for CFO appointments. Mian (2001) reports that the mean stock market reaction for outsider (insider) new CFO is 0.40% (-0.23%). In their study of newly-created CIO positions, Chatterjee 16

19 et al. (2001) find the mean stock market reaction for appointment of outsider (insider) CIO is 1.02% (1.37%). Panel D provides results for the breakdown of new and old positions into outsiders and insiders. In terms of magnitude, newly created SCOME positions staffed by outsiders have the most positive reaction with a mean abnormal return of 0.86%. This result in terms of magnitude is followed by new positions and insider with a mean abnormal return of 0.28%, existing positions and outsider with a mean abnormal return of 0.17%, and existing positions and insider with a mean abnormal return of -0.34%. To summarize, overall the stock market reacts positively to appointments of SCOMEs. Consistent with our hypothesis, the market reaction is positive for newly created SCOME positions. Although hiring outsiders versus insiders each have their advantages and disadvantages, the stock market reacts more positively to outsider appointments rather than insider appointments. This finding is not only consistent with our hypothesis but also with the evidence provided on other non-ceo top management appointments such as CFOs and CIOs. 4.3 Abnormal returns in the period prior to appointment of new SCOMEs To develop additional insights into the relation between stock price performance and SCOME appointments, this section analyzes the abnormal stock price performance in the period prior to SCOME appointments. There are two reasons for presenting this analysis. First, nearly 65% of the SCOMEs in our sample are appointed to existing positions and 35% to newly created positions. The need to replace an existing SCOME could be because the existing SCOME is retiring, promoted, fired for poor performance, leaving for personal and family reasons, or joining another firm, among others. The creation of a new SCOME position could be because of reorganization, promotion, the need to improve performance, or the need to attract a SCOME from another firm, among others. 88% of the announcements did not give information about the reasons for replacing existing SCOMEs or creating new positions. Given this lack of information, we explore the more general issue of whether prior stock price performance is a driver for replacing existing SCOMEs or creating new SCOME positions. Second, we also analyze whether the stock market reaction to announcements of appointment of new SCOMEs 17

20 (see Section 4.2) is related to prior stock price performance. One could conjecture that if the prior performance is good (poor) then the announcement to appoint a new SCOME could be bad (good) news as a good performer (poor performer) is being replaced. Our measure of prior stock price performance is the cumulative abnormal return (CAR) over a 250 day preannouncement period that starts on day -270 and ends on day -21. Since a month typically has 21 trading days, days -270 to -21 spans a year that starts 13 months before the announcement and ends one month before the announcement. The reason we end one month before the announcement is to reflect the possibility that firms are likely to have made the decision to find a new SCOME sometime before the announcement. The estimation period used to compute the CAR is from Day -480 to day We estimate the CAR using the Four-Factor Model and the Market Model. Since 64 firms did not have sufficient data during the estimation period and/or the pre-announcement period to compute CARs, the results for CARs are based on 432 firms. Table 4 presents these results. Although the magnitudes of the abnormal returns are different across the two models, there are no substantive differences in the inferences. Given this, we focus our discussion on the results from the Four-Factor Model. The results suggest that new SCOME appointments are preceded by poor stock price performance. Panel A of Tables 4 indicates that for the full sample the abnormal returns are negative during the 250 day pre-announcement period. The mean prior-period abnormal return is -8.61%, significantly different from zero at the 1% level. The median abnormal return is -3.74%, significantly different from zero at the 2.5% level. Panel B suggest that the much of this negative prior-period stock price performance is driven by the subsample of new SCOMEs in existing positions. The mean (median) prior-period abnormal return is % (-6.45%), significantly different from zero at the 1% (1%) level. For newly-created SCOME positions, the prior-period abnormal returns are insignificantly different from zero. This evidence suggests that many firms may be replacing existing SCOMEs after a period of poor performance. Panel C indicates that when prior performance is poor, firms are more likely to appoint outsider SCOMEs. The mean (median) prior-period abnormal return for outsider SCOMEs is % (-9.51%), significantly different from zero at the 1% (1%) level. On the other hand, the prior-period abnormal 18

21 returns for insider SCOMEs are insignificantly different from zero. This result provides some preliminary evidence to support our hypothesis that the likelihood of hiring outsider SCOMEs is negatively related to prior performance. Panel D provides results for the breakdown of new and old positions into outsiders and insiders. The results indicate that the worst prior performance is for the subsample where outsiders are appointed to existing positions. The mean (median) prior-period abnormal return is % (-14.40%), significantly different from zero at the 1% (1%) level. Overall, the evidence presented in Table 4 suggests that SCOMEs are replaced when the prior performance is weak and firms tend to appoint outsiders when the prior performance is weak. This suggests that performance related disciplinary mechanisms are at work in appointments of SCOMEs. To shed light on the issue of whether the stock market reaction to announcement of appointment of new SCOMEs is related to prior stock price performance, we estimate the correlation between the abnormal return on the announcement day and the CAR over a 250 day preannouncement period. The Pearson correlation coefficient is and the Spearman rank correlation coefficient is Both these correlation coefficients are insignificantly different from zero suggesting that the stock market reaction on announcement day is not influenced by the stock price performance during the preannouncement period. 5.0 Results on factors that affect the choice of an insider or outsider as the new SCOME In this section, we test our hypotheses on whether factors such as firm size, industry homogeneity, industry concentration, and prior performance affect the likelihood of the SCOME being an outsider or insider. We first describe how the various independent variables are measured and then present the results from the probit regression analysis. The independent variables are measured as follows: Firm Size: measured as the natural logarithm of sales in the most recent fiscal year completed before the date of the announcement of the appointment of the new SCOME. Industry Homogeneity: Recall that we posited that firms operating in a more homogenous industry are likely to compete in similar product markets, using similar inputs and technologies, and similar supply chain structures. Therefore, changes in factors such as input prices, technology, and competitive 19

22 environment are likely to have similar effects on firms in a more homogenous industry. This observation also suggests that the cash flows and therefore the stock prices of firms in a more homogenous industry are likely to be more positively correlated. Following Parrino (1997), we use a stock price based proxy for measuring industry homogeneity. We use the industry homogeneity in the month before the announcement of the new SCOME in our probit analysis. For a given month X, the proxy for industry homogeneity is calculated in three steps described as follows: 1. We identify all firms that belong to a particular three-digit SIC code in month X. For each firm in a three-digit SIC code, we estimate the monthly abnormal returns using the Market Model for 60 months. This 60 month period includes the month X and 59 months before month X. For a firm to be included in this estimation, it must have at least 30 months of data over the 60 month period. The monthly abnormal returns are the residuals from the following regression equation: R im = α i + β i R mm + ε im (9) where R im is the return of stock i in month M, is the intercept of the relationship for stock i, is the slope of the relationship for stock i with respect to the market return, and R mm is the market return in month M, and, the error term, is the portion of the return that cannot be explained by the market and is the abnormal return of stock i in month M. 2. For each firm Y in a three-digit SIC code, we correlate Y s monthly abnormal returns that is estimated in step 1 with an equally-weighted portfolio of monthly "industry" abnormal returns of all other firms that are in the same three-digit SIC code as Y. This gives one value of the correlation coefficient. We exclude Y s abnormal returns in computing the equally weighted portfolio of monthly "industry" abnormal to prevent any bias if the number of firms in the industry is small. If there are N firms in the industry, we repeat this for each firm in the industry, and estimate N correlation coefficients, one for each firm. 3. The average of the N correlation coefficients estimated in step 2 is our proxy of industry homogeneity. This average can be interpreted as the average correlation between the firm and its 20

23 industry after the effect of the overall stock market movements is removed. The higher the average correlation, the more homogenous is the industry. Industry Concentration: We use the Herfindahl index to measure industry concentration. The Herfindahl index for an industry is defined as the sum of the squared fraction of industry sales of each firm that is in the industry. For each sample firm, we compute the Herfindahl index of its industry using sales of all firms with the same primary three-digit SIC code as that of the firm sample. We use the sales in the most recent fiscal year completed before the date of the announcement of the appointment of the new SCOME. A higher (lower) value of Herfindahl index means a more concentrated (fragmented) industry. Prior Performance: We measure prior performance in a number of ways including stock price, profitability, and inventory performance. The specific measures used are the following: 1. Cumulative Abnormal Returns from Day-270 to Day -21: We use the Four-Factor Model described in Section 4 to compute the abnormal returns. 2. Return on assets (ROA) of the Firm: ROA is the ratio of operating income (sales less the cost of goods sold, and selling, general, and administration expenses), divided by the average of beginning and ending period book value of assets to get return on assets (ROA). We estimate ROA for the most recent fiscal year completed before the date of the announcement of the appointment of the new SCOME. 3. Median Industry ROA: Industry ROA includes all firms that have the same three-digit SIC code and is based on the most recent fiscal year completed before the date of the announcement of the appointment of the new SCOME. 4. Industry-adjusted ROA of the Firm: This measure is the difference between the ROA of the firm and Median Industry ROA of the firm, where industry is defined at the three-digit SIC code level. 5. Industry-adjusted Days of inventory of the Firm: To ensure comparability, we convert all inventory and cost of goods sold (COGS) data to First-In-First-Out (FIFO) basis by using the Last-In-First- Out (LIFO) reserves. Days of inventory is calculated as 360*(FIFO Inventory/FIFO COGS). We 21