STRATEGIC ORIENTATION AND CORPORATE SOCIAL PERFORMANCE: AN EMPIRICAL EXAMINATION Roy L. Simerly East Carolina University Greenville, NC Anisya S. Thomas Florida International University Miami, FL Theoretical Background The strategic choice perspective (Child, 1972) suggests that the top managers of an organization determine its future direction by making important choices about products, markets and technologies. Over time, strategies, or streams or patterns of major and minor decisions (Miles and Snow,1978; Mintzberg,1978) contribute to the development of inter- nally consistent configurations of systems and processes that guide an organization's interaction with, and response to the external environment. Since corporate social perfor- mance (CSP), has been defined as the identification ofthe domains ofan organization's social responsibility, the development of processes to evaluate environmental and stakeholder demands and the implementation of programs to manage social issues (Carroll, 1979; Warlick and Cochran, 1985; Wood, 1991a; Wood, 1991b), it would seem logical to infer that the strategic orientation ofan organization influences its social policies and therefore its CSP. Typologies of Strategy: The Miles and Snow Framework Typologies provide a means of classifying organizations in a comprehensive yet parsimonious framework for theory development (pinder and Moore, 1979) because they reduce the myriad variables that characterize the organization to a manageable set of internally consistent ones. Thus the primary benefit of the typological approach is that the evidence of a few significant characteristics in an organization's profile facilitates the prediction of others. The strategic management literature has developed a variety of strategic typologies that describe an organization's competitive posture in an industry. These typologies ofstrategy suggest that over time organizations develop configurations ofsystems and processes which contribute to the development of a distinctive competence and consequently competitive advantage (Miles and Snow, 1978; Porter, 1980). Prominent among these is the Miles and Snow (1978) typology of strategy which describes bundles of organizational characteristics that have been found to lead to successful performance. This typology has been the subject of extensive testing in the strategic management literature and has been judged to be reasonably valid (Doty, Glick and Huber, 1993; Zahra and Pearce, 1990). Miles and Snow identify three viable strategies, Prospectors, Analyzers and Defenders. While Prospectors and Defenders are maximally different forms, Analyzers are 114 Journal ofBusiness Strategies Vol. 11, No.2 hybrid strategies that exhibit some of the key features of both. Prospectors are externally oriented organizations that compete by pioneering new products and developing innovative marketing techniques. Since they are constantly involved in monitoring the external environment and developing alternative responses to emerging trends, they are the creators of change in an industry. In contrast Defenders are internally oriented and focus on penetrating existing markets by improving operational efficiency. The main thrust of the Defender strategy involves "the creation of a narrow, stable domain ... through a limited mix ofproducts and customers and aggressive efforts to "protect" the domain from competitors" (McDaniel and Kolari, 1987:39). We suggest that the manner in which Prospectors, Analyzers and Defenders perceive and react to the needs of their manifold constituencies will be as distinct as their market approaches. Since Prospectors are externally oriented, they develop and value competencies that facilitate the scanning of the external environment and the early identifi- cation of trends. Thus it is expected that Prospectors are more adept at managing the needs of their stakeholders and are likely to achieve superior CSP. In contrast, Defenders operate in narrow and stable domains, are internally focussed and value efficiency above all else. The backgrounds and skills oftop managers reflect this orientation and tend to focus on activities that enhance productivity. Consequently, Defenders are less likely to assign much impor- tance to factors outside the organization which in tum translates to less proactive social programs and policies and lower CSP. Based on this logic we propose the following hypothesis: HI: Prospectors are more likely to achieve higher levels ofCSP than Defendersl . Method Setting and Sample The sample for this study was drawn from the 305 fIrms that comprise Fortune magazine's list of Most Admired Corporations. This list encompasses fIrms in the 32 industries included in the Fortune 500 and Fortune Service 5002 . The initial sample comprised only manufacturing fIrms. To maintain a basic level of homogeneity in the sample, service fums were deliberately excluded as it is commonly recognized that they differ in their strategic approaches and competitive maneuvers (Fombrun and Shanley, 1990). Industries with fewer than eight ftrms on the list were also eliminated. This step was necessary to ensure a minimum level of variance in the dependent measure, CSP, which was described as reflecting a fIrm's performance relative to the mean of its industry. To justify the use of the Miles and Snow (1978) typology which is recognized as being applicable at the business level (Hambrick, 1983), only fIrms that received more than 70 percent oftheir sales from a single industry were included in the sample (Rumelt, 1974). This step is useful in selecting a sample of comparable fIrms and is consistent with methods employed in prior literature operationalizing the Miles and Snow typology (fhomas, Litschert and Ramaswamy, 1991). Firms that were classifted as conglomerates, holding Fall 1994 Simerly & Thomas: Strategic Orientation 115 companies, subsidiaries or those that had the majority of their sales or manufacturing outside the U.S. were excluded. These procedures resulted in the identification of 134 fIrms. Data and Measures Strategy. Ginn and McDaniel (1987) note that an organization's strategy is reflective of all its systems, allowing a limited selection of theoretically relevant variables to capture strategic type. In this study, strategy was viewed as being implied :from behavior rather than stated by management. Thus, the variables selected to operationalize the strategies represent a series of resource allocations and competencies necessary for their suCC¥ssful implementation. The underlying assumption was that the value of these variables stems mainly from managers exercising choice (Child, 1972). Five different measures were used to operationalize the strategy types. Since strategy is recognized as a stable phenomena which becomes established over time and is relatively immutable, four year averages (1985-1988) were used to describe each of the strategy measures. This step also ensures the stability of the data, smoothing the effects of temporary fluctuations in resource allocations due to unusual circumstances in an organiza- tion or its external environment. Wherever possible, ratios were used to control for the possibly contaminating effect of fIrm size. The data for these measures were obtained primarily from COMPUSTAT tapes, Company Annual reports, and 10-K reports. Each of these measures is briefly discussed below. Research and Development (.R&D) Expenditures: Snow and Hrebiniak. (1980) found that of all the distinctive competencies examined, emphasis on research and develop- ment most clearly differentiated between the strategic types. Hence a ratio of R&D expenditures to total sales was used as a standardized indicator of a fmn's propensity to search for new products. This measure was derived from previous studies investigating the Miles and Snow typology (e.g. Hambrick:, 1983; McDaniel and Kolari, 1987). Since Prospectors engage in greater amounts ofinnovative activity, they are expected to rank highly on this indicator. Marketing Ewnditures: A ratio of marketing expenditures to total sales were used as an indicator of the fIrm's market focus. Typically marketing expenditures include direct promotion, advertising and other general selling expenses. In keeping with their external orientation, Prospectors were expected to rank highly on this measure. Net Sales per Employee: This ratio describes an organization's ability to produce and distribute goods efficiently. Since the Defender strategy is geared toward the maximi- zation of efficiency, it is expected that these organizations would have higher scores. Property Plant and Equipment Expenditures: A ratio of property plant and equip- ment to total sales was used to describe the capital intensity of the business. Derived from previous literature on the Miles and Snow typology (Hambrick, 1983), this measure attests to an organization's investment in efficient production. Thus, Defenders were expected to rank highly on this measure. CEO Experience in External and Internal Functions: The functional background of the CEO was coded as a categorical variable to reflect external or internal experience. 116 Journal ofBusiness Strategies Vol. 11, No.2 External experience included backgrounds such as marketing and product research and development while internal experience included backgrounds in finance, engineering and manufacturing. The area where the CEO spent the longest amount of time, prior to obtaining the position at the helm of the corporation, was used to determine his or her functional background. Previous research (e.g. Chaganti and Sambharya, 1987; Thomas, Litschert and Ramaswamy, 1991) suggests that Prospectors will have a greater proportion of externally oriented executives and that Defenders will have a greater proportion of internally oriented executives. The data for this measure was obtained from Dun and Bradstreet's Reference Book ofCorporate Management (1988). Col1'Orate Social Performance. Each year Fortune polls over 8000 senior execu- tives, outside directors and fmancial analysts to evaluate the reputations of firms which comprise its list oflargest industrial and service corporations. These experts are asked to rank the corporations in their industry, using a scale from 0 (poor) to 10 (excellent) on eight qualitative perceptual dimensions3 . These scores are then combined to derive a composite number which determines a fums rank on the list of "most admired corporations". The raw data reflecting the evaluations of the industry specialists on each of the eight dimensions was obtained from Fortune and used in this analysis. This data has similarly been used in a number of previous studies that attest to its validity (e.g. McGuire, Sundgren and Schneeweis, 1988; Wokutch and Spencer, 1987). The attribute community and environmental responsibility was used as a surrogate for CSP as there are no widely recognized, objective measures to denote this construct. This measure has been used in a variety of previous studies over the past five years (Boyd and Carroll, 1993; Fombrun and Shanley, 1990; Nayyar, 1992; Wartick, 1992). Wokutch and Spencer (1987) suggest that it is particularly appropriate as it reflects the opinion of a wide range of industry experts. Since the Fortune survey has been conducted annually for a period of fourteen years, it can be assumed that the raters account for shifts in corporate social performance over time, and evaluate firms in a dynamic rather than a static fashion. To assess the reliability of the data, correlation analysis was performed on a sub-sample of firms in several industries over a three year period (1987, 1988, 1989). Notwithstanding changes in economic and environmental conditions, the correlations were consistently positive and significant (rang- ing between 0.7 and 0.9). There was considerable variance in the CSP measure across industries, with the firms in some industries ranking consistently high while those in others clustered at the bottom ofthe overall list. Thus CSP was independently defined within each industry context by comparing each frrm in an industry to the mean CSP score for that industry. Companies that ranked above or below the mean CSP in their industry were classified as high CSP frrms and low CSP fums respectively. By classifying each fmn within its own industry, the problem of perceptual bias by raters of a specific industry was substantially reduced. Since CSP is understood to be an outcome of a variety of internal variables and resource allocations (Fombrun and Shanley, 1990), it was measured in 1989 following the classification of the strategies of the fums in the sample. Fall 1994 Simerly & Thomas: Strategic Orientation 117 Data Analysis Identifying Strategic Types: Miles and Snow (1978) contended that the three viable strategies, Prospectors, Analyzers and Defenders should be found in every industry. This assertion has subsequently been validated by empirical research in multiple industries (Snow and Hrebiniak, 1980). However, there is no defInitive evidence that details the numerical distribution for the three strategies in a particular context (Smith and Grimm, 1987; zahra and Pearce, 1990). Therefore a two-step process was used to identify Prospectors and Defenders. In keeping with the contentions of Miles and Snow, strategic orientation was operationalized as a continuumwith Prospectors and Defenders occupying the two extremes. To construct the continuum, a two variable criterion set (marketing expenditures/sales and research and development expenditures/sales) was used. These variables were specifIcally chosen for several reasons. According to Miles and Snow (1978), research and marketing define the external thrust of the Prospector strategy. Further previous research has identified these variables as having high predictive validity (zahra and Pearce. 1990). Finally, a correlation analysis empirically confIrmed high convergence between these two measures (r=O.98). Therefore the two measures were summed and the frrms arrayed in ascending order. A median split method was adopted to identify strategic types (see Drazin and Van de Ven.1985; Romanelli, 1989 for similar approaches). The top 25 percent (frrms having the highest summed score) and the bottom 25 percent (frrms having the lowest summed score) of the firms were respectively denoted as Prospectors and Defenders. This process resulted in two sets of 34 frrms each that were selected for further testing. Nevertheless, it is recognized that such a procedure can result in classifIcation errors at the margin. In other words, fIrms having the highest summed score in the Defender sample and firms having the lowest summed score in the Prospector sample could in fact be pursuing the hybrid Analyzer strategy. To establish the reliability of the above strategy typing approach, the use ofmultiple methods testing for convergence was necessary. Thus the next step in the analysis was one of validation. The remaining three strategy measures comprising the hold out set (sales per employee, property plant and expenses to revenues and functional background of the chief executive) were used for this purpose. As indicated in Tables 1 and 2, directional t-tests and chi-square tests on these measures (as well as the ones used in the criterion set) revealed statistically significant differences between the two strategy groups. Prospector fIrms were found to have a signifIcantly higher proportion of R&D expenditures and marketing expenditures than Defenders. They also had a signifIcantly greater proportion of executives with externally oriented rather than internally oriented functional backgrounds. On the other hand, Defenders exhibited higher levels ofemployee productivity and capital intensity. Their CEO's also tended to have a greater proportion of experience in internal functions as compared to external functions. These findings parallel the conclusions of Hambrick (1983) and Thomas et. al. (1991) who reported similar patterns in their investigations of the Miles and Snow typology. 118 Journal ofBusiness Strategies Vol. 11, No.2 Table 1: Summary of t-test results for strategy variablesa Criterion Set R&D Expenses/Sales Marketing Expenses/Sales Hold-Out Set Sales/Employee Prop. Plant and Eqip. Prospector (n=34) .0606*** (.0387) .3796*** (.0814) 119.04 (56.33) .3087 (.0858) Defender (n=34) .0094 (.0120) .0689 (.0814) 303.93** (380.99) .4931 *** (.1891) a Natural means reported. Standard deviations are in parentheses. • •• •••p