CHAPTER 6 METHODOLOGY
6.2 Quantitative Sample and Data Collection
6.2.4 Firm Profits
Firm profits were measured in a similar way to industry profits; an index for firm profits was calculated using the five year average of return on assets and equity for the firm in question. Similar to the determination of industry profits, this measure was compared to the average market return on equity and return on assets to produce a ratio. The market average was calculated as described in section 6.2.3. Once again, these two ratios were summed to produce an index of firm profits which was used in the analysis with the dependent variable. The data used were taken from the Compustat database.
6.2.5 Munificence
The determination of the level of munificence, and what should be included in the construct itself, has been calculated in a variety of ways in studies on illegal corporate activity. Staw and Szwajkowski (1975) used the mean return on equity and sales for the preceding five years of the industry’s, firm’s and market’s performance to measure munificence. While the authors rejected the measure of a percentage change in profits and sales as less valid indicators, Dess and Beard (1984) found that growth in sales, price to cost margin, total employment, and number of establishments loaded on munificence. Other researchers in this area included different factors in their munificence construct.
Keats and Hitt (1988) used growth in net income and growth in operating income, Baucus and Near (1991) employed growth in industry value added and finally Goll and Rasheed (1997) measured munificence as the regression slope of the value of shipments. I adopted Keats and Hitt’s measurement by measuring munificence as the regression slope of the difference between industry sales and cost of goods sold (a rough estimate of net income) over a five year period before the incidence of illegal cartel activity. I regressed the difference of sales from the cost of goods sold on time and use the regression slope from this OLS regression. This data were taken from the Compustat database.
6.2.6 Dynamism
Dynamism, as a construct, has been considered in a number of different ways, some more robust and inclusive than others. Dess and Beard (1984) found that the measures of instability of sales, price to cost margin, employment, and value added loaded on their measure of environmental instability whereas Keats and Hitt (1988) (using the name instability as opposed to dynamism) only looked at the instability of net sales and operating income over a five year period prior to the time frame in question. Baucus and Near (1991) measured complexity of a firm by the number of four digit businesses the firm was in. I followed the measurement technique used by Keats and Hitt (1988) by taking the standard error of the regression slope co-efficient and dividing it by the mean value of the net sales and net incomes over a five year period. These two measures were combined in an index for analysis with the dependent variable. This data were available from the Compustat database
6.2.7 Board Scanning
Board scanning is a construct that has very rarely been dealt with empirically in the literature. To inform this construct, I searched both the ProQuest ABI Global Research Database and the Web of Science Database on such topics as “board scanning”, “director scanning”, “board of directors scanning” and found only one article on this topic, a 1979 study by Ritvo, Salipante and Notz (1979) that looked at board problem recognition and used board scanning as the independent variable. Methodologically, they used a survey that was sent to hospital board members. I then expanded the search to include the concept of corporate scanning more broadly. This search was more fruitful, locating a number of articles which incorporated the CEO and the TMT rather than just scanning carried out by the board. This research stream relied upon a broader toolbox of methods to demonstrate the validity of the scanning construct including the survey method (Boyd & Fulk, 1996; Garg et al., 2003; Goll & Rasheed, 1997), interviews (Daft, Sormunen, & Parks, 1988), case studies (Ngamkroeckjoti & Johri, 2000), and simulations (Hough & White, 2004).
It has been suggested that the board scanning function is partially developed from the knowledge that a director obtains from their experience in general (Murphy & McIntyre, 2007) and from serving on other boards in particular (Carpenter & Westphal, 2001). The experience of individual directors and the fact they may serve on other boards indicates a superior scanning ability. A reasonable proxy for this ability is to determine whether the director is an outside or an inside director. An outside director will arguably have a better scanning potential than an inside director as a result of a more intense immersion in the external environment.
In scanning the external environment and its attendant information processing requirements, a larger board has greater ability than a smaller board (Sanders & Carpenter, 1998); in general this also applies to the general monitoring function of the board (Petra & Dorata, 2008). Furthermore, I assert that having more directors on the board will result in an enhanced scanning ability as each director brings with them the skill, time and effort to scan the environment and bring the information back to the organization. This ability accumulates as the number of directors increases. Finally, the relative size of the firm will also affect the scanning effectiveness of the board. The larger the firm the more overall scanning is required to achieve a given level of effectiveness. As the proportion of outside directors and the total number of directors are important when considering the board’s effectiveness in scanning the external environment, my measure for this construct was an index of the two measures. As noted previously, an index provides better construct validity than single indicators, single ratios or discrete indicators (Boyd et al., 2005). This index was compiled by adding the percentage of outside directors to the log of the number of directors, generating an index measure of between 0 and 2.5. This data were obtained from the companies’ annual reports and the Compustat database.
Given that board scanning is not a common variable, it was important to ensure that this conceptualization was appropriate. I developed face validity for board scanning by interviewing six current and former board members on their views of board external environmental scanning. These board members represented both for-profit (manufacturing, consulting and financial services) and not-for-profit (economic development commissions and religious institutions) organizations. The goal of these
interviews was to provide insight into the conceptualization of board scanning, ensuring that it was consistent with both theory and practice. The interviews followed a semi- structured format (Glaser & Strauss, 1967) and were conducted by telephone. I began by introducing the research topic and asked interviewees to describe their thoughts and experiences with board scanning. I followed three main areas in the interviews:
1. How do board members scan the external environment for issues that may have some impact on the organization;
2. Does the size of the board have any impact on the effectiveness of the board’s scanning activities; and,
3. Does the percentage of outsiders affect the board’s effectiveness in scanning? All interviews were approximately thirty minutes long and followed the semi structured questions outlined in Appendix A. The ethics approval is contained in Appendix B. The individuals interviewed confirmed that the proportion of outside directors and the total numbers of directors are important when considering the board’s effectiveness in scanning the external environment. Thus, the index I constructed was used as the indicator for board scanning.