Measures

3.3.1 Independent Variables

EO. The independent variable in this study is EO. Three dimensions are used to measure EO, including innovation, proactiveness, and risk-taking. The three dimensions are measured by using nine questions developed by Miller (1983) and Covin and Slevin (1986, 1988, & 1989). Following these studies, a semantic differentials method is used in the questionnaire. This means that two opposing phrases are offered for each question, and respondents are asked to rank the indices on a seven-point Likert scale, ranging from 1 to 7. The higher the score, the stronger the EO is the firm. Reliability is estimated by using both coefficient alpha (Peter, 1979) and composite reliability (Fornell & Larcker, 1981). The Cronbach’s alpha values of the three dimensions are 0.800, 0.850, and 0.903 respectively, with an overall Cronbach’s alpha of 0.868. The test of reliability in our sample is consistent with past studies (Runyan, Droge, &

Swinney, 2008). The measure items of EO are shown in the Appendix A.

Resource-Capability Combination. These are the mediated variables of this study and are operationalized as two dimensions of resource attributes: value and rareness. The dimensions of value and rareness are measured by using the scales developed by Newbert (2008). Again, a seven-point Likert scale is used.

Value. The value of resource-capability combination is operationalized as an attribute in which the value of a resource (or a capability) can be enhanced when it is combined

- 80 -

with a capability (or a resource) to reduce costs and exploit market opportunities (Newbert, 2008). It is measured using four questions, each with five items developed by Newbert (2008), including financial, human, intellectual, organizational, and physical resources/capabilities. An averaged score of the questions is then calculated to indicate the overall value of a firm’s resource-capability combination. The higher the score of a firm, the higher is the value of its resource-capability combination.³ This construct has an overall Cronbach’s α of 0.887. The measure items of value are shown in the Appendix A.

Rareness. The rareness of resource-capability combination is operationalized as a firm’s exploitation of a common resource (or capability) with a unique capability (or resource) or a firm’s exploitation of unique resource-capability combinations, in order to reduce costs, utilize market opportunities, or withstand competitive threats. Following Newbert (2008), this construct is measured using three questions, each with five items—financial, human, intellectual, organizational and physical resources/capabilities. Similarly, the averaged score of the questions is then calculated to indicate the overall rareness of a firm’s resource-capability combination. The higher the score of a firm, the higher is the rareness of its resource-capability combination.⁴ This construct has an overall Cronbach’s α of 0.925. The measure items of rareness are shown in the Appendix A.

Environmental dynamism. Environmental dynamism is a moderating variable in this

3 The respondents are asked to rank the extent to which they agree on a seven-point Likert scale, ranging from extremely disagree (=1) to extremely agree (=7).

4 Ibid, footnote 2.

- 81 -

study. It is measured by using five questions, including extreme changes in marketing practices, a rapid rate of obsolescence in fashion goods/semi-conductors, the unpredictability of competitors, unpredictable demand and tastes of customers, and the modes of production/service change. The scales of environmental dynamism developed by Miller and Friesen (1982), and a semantic differential method is used in the questionnaire. Each question offers two opposite phrases. The overall Cronbach’s α is 0.81. The measure items of environmental dynamism are shown in the Appendix A.

3.3.2 Dependent Variables

Firm performance. Firm performance is the dependent variable in this study. Consistent with prior studies, this study uses two categories to measure firm performance:

subjective and objective measures, according to the recommendation of Weinzimmer, Nystrom, & Freeman (1998).

First, subjective measures are divided into competitive advantage and satisfaction, representing long-term performance. Following Newbert (2008), competitive advantage is based on the respondents’ answers to three questions, including cost reduction, opportunity exploration, and the defense of competitive threats. Each question includes five items to indicate different types of resource-capability combinations, i.e. financial, human, intellectual, organizational, and physical resources/capabilities.⁵ It has an overall Cronbach’s α of 0.903. The measure items of competitive advantage are shown in the Appendix A. In term of satisfaction, it is used in strategic management literature

5 Ibid, footnote 2.

- 82 -

and organization literature (Murphy et al., 1996; Venkatraman & Ramanujam, 1986).

Following the studies of Beal and Yasai-Ardekani (2000), Newbert (2008), and Venkatraman and Ramanujam (1986), satisfaction consists of five indicators: sales growth rate, return of assets, rate of profits, customer satisfaction, and brand image.⁶ The overall Cronbach’s α is 0.844. The measure items of satisfaction are shown in the Appendix A.

Second, objective measures for performance include return on assets (ROA) and Tobin’s q (TQ), representing short-term performance (Beal & Yasai-Ardekani, 2000;

Fitzsimmons, Douglas, Antoncic, & Hisrich, 2005; Luke, Verreynne, & Kearins, 2007;

Venkatraman, & Ramanujam, 1986). The averaged annual rate of profit after taxes but before interest on total assets (ROA) between 2007 and 2009 is appropriate to estimate the effectiveness of the business operations (Combs & Ketchen, 1999) due to the high debt-equity ratio and imperfect capital markets in developing economics (Chang & Choi, 1988). Moreover, Tobin’s q has been widely used to examine the source of unequal profitability (Lindenberg and Ross 1981). The stock market of firm performance is Tobin’s q, the ratio of the firm’s market value to the replacement costs of its assets between 2007 and 2009. Following Khanna and Palepu (2000), Miller and Breton-Miller (2011), and Villalonga and Amit (2006), a proxy variable for Tobin’s q is defined as: (market value of equity plus book value of preferred stock plus book value of debt)/(book value of assets). ROA and Tobin’s q were collected via a secondary

6 The respondents are asked to rank the extent to which they agree on a seven-point Likert scale, ranging from not satisfactory at all (=1) to extremely satisfatory (=7).

- 83 -

database maintained by the TEJ.

3.3.3 Control Variables

Control variables. Several variables that might influence the competitive advantage of firms are controlled in the regression models, including firm size, firm age, debt-to-market ratio (DEMKT), environmental dynamism, and industry affiliation.

Firm size reflects the economies and diseconomies of scale and may form barriers to entry (Bain, 1968) and is operationalized as the natural logarithm of the three-year average of total employees. Firm age is controlled because prior studies suggest that the established organizations are more bureaucratic, and this factor influences their competitive advantages (Hannan & Freeman, 1989). A firm’s age is measured as the company’s age since its establishment. DEMKT is controlled because a firm with a low debt-to-market ratio is more likely to create a competitive advantage (Chatterjee &

Wernerfelt, 1991; Palepu, 1986). Industrial environments are controlled by using industry affiliation (Khandwalla, 1976; Lumpkin & Dess, 2001, Miller & Friesen, 1982). Possible performance differences resulting from industrial affiliation are also controlled. Based on the industry classification of TSE, 17 dummy variables are used to classify the sample firms into 18 industries.⁷ Table 3.1 summarizes the number and percentage of firms based on their industrial categories.

7 According to the TSE database, these industries include cements, food and beverage, plastics, textile, electric machinery, electrical wire and cable, chemicals and biotechnology, glass and ceramic, paper, iron and steel, rubbers, information and electronics, building and construction, shipping and transportation, tourism, wholesale and retail trading, electricity, and other miscellaneous industries.

- 84 -

TABLE 3.1 Number and Percent of Public Firms by TSE Industry Code

TEJ

23 Information and electronics 729 88 0.12 43.78%

25 Building and construction 53 20 0.38 9.96%

- 85 -