4.3. Descriptive Statistics
Among the sample organizations, the four most disregarded dimensions of SP were found to be “purchase from minority and women-owned business enterprises” (1.94), “the presence of a formal program for making purchases from minority and women-owned business enterprises” (1.96), “participates in the design of products for recycling or reuse” (2.61), and “participates in the design of products for recycling or reuse” (2.69). This finding suggests that “donates to philanthropic organizations” and “volunteers at/for local charities” are the most commonly embedded SP dimensions instead of environmental practices.
In order to determine the SP variation between private and public organizations and between local and joint-venture organizations, an independent samples
t-test was conducted. The results in
Table 3 reveal that, in the engagement of SP activities, the private organizations were well ahead of the public organizations. The mean score of three dimensions of SP, “asks suppliers to commit to waste reduction goals”, “donates to philanthropic organizations”, and “purchases from small suppliers”, was significantly greater (
p ≤ 0.05) for the private organizations than the public organizations, while the remainder of SP practices were statistically insignificant.
The independent samples t-test results also identified that overall the joint-venture organizations were much better position than the local organizations in the engagement of SP activities. Joint venture organizations were significantly ahead (p ≤ 0.05) of local organizations in adopting only two dimensions of SP practices, “asks suppliers to commit to waste reduction goals” and “donates to philanthropic organizations”, while the remainder of SP practices were statistically insignificant.
In order to test Hypotheses 1–4, a correlation analysis was performed on the SP practices of organizational nonfinancial and financial performance scales. One-tailed significance tests were performed due to the limited sample size and expected positive correlations; the results are shown in
Table 4. The results in
Table 4 reveal that three out of six SP practices scales were significantly and positively associated with both financial and nonfinancial organizational performance, with the strongest relationship between environmental purchasing, as a practice of SP, and organizational quality and operational performance (
r = 0.372). The remaining three SP scales were positively but insignificantly related to performance scales. In addition, we carried out correlations among the summary measure (created by averaging scores for the scales) of SP practices and organizational performance scales, the results of which are shown in
Table 5. The correlation analysis showed that all scales of organizational performance were significantly and positively related to overall SP practices, with the strongest relationship between ”social and environmental performance” (
r = 0.395) as a facet of organizational performance, and SP practices. Overall, this result provided acceptable evidence that all facets of SP practices impacted on organizational nonfinancial performance as well as financial performance, with all correlations being in the predicted direction.
Table 6 and
Table 7 identify the variation in the benefits of the adoption of SP dimensions across the sectors, types, and ownerships of the organizations. The procurement directors/managers of the construction sector registered significantly higher levels of SP practices compared to other sectors. For example, the mean scores for the construction sector are greater than that of other sectors on all benefits. Particularly, this sector scored significantly greater on social and environmental performance (
p ≤ 0.05), and financial performance benefits (
p ≤ 0.05).
The results in
Table 7 indicate that joint ventures compared to local organizations and private compared to public organizations had both higher financial and nonfinancial benefits from the implementation of SP practices but the differences were statistically insignificant. Possible explanation of this result could be that the nature, culture and practices of purchasing of joint ventures differ from local organization’s practices and similarly public sector purchasing practices differ from private sector practices. In addition, a possible explanation could also be that private organizations and joint ventures are profit driven whilst public organizations are not necessarily for-profit. Moreover, public organizations are generally more politically constrained than private organizations and most private organizations have significantly different organizational objectives than their public counterparts. Organizational structure and level of commitment associated with public and private sectors could be another potential explanation.
For further analysis of H1, H2, H3, H4, and H5, multiple regression analyses were conducted by using a total nonfinancial performance score and financial performance score (created by averaging scores for the all scales). For regression analysis, the F-test and adjusted R2 value confirmed that regression is meaningful in the sense that the dependent variable is related to each specified explanatory variable. The correlation matrix of the independent variables was studied to identify the occurrence of multicollinearity. The correlation matrix also ensured that the Pearson’s r between each pair of independent variables did not exceed 0.80. It means that the independent variables are not too highly related to each other. Moreover, for collinearity diagnostics, the eigenvalues of the explanatory variables were studied by factoring the scaled (so that diagonal elements are 1’s), uncentred cross products matrix of the independent variables. Eigenvalues provide an indication of how many distinct dimensions are there among the independent variables. Since in the present model, several eigenvalues of the explanatory variables are not close to 0, thus the variables are proved to be not intercorrelated and the matrix is said to be efficient-conditioned.
The R2 statistic for the first regression showed that more than 41 percent of the variance in nonfinancial performance was predictable from all of the SP practices scales, while R2 for the second regression showed that more than 35 percent of the variance in financial performance was predictable from all of the scales of SP practices. These results show a good fit for these data. The F-test showed a highly significant result (p ≤ 0.01), which implies that the variables in the models, taken together, are significant predictors of organization’s nonfinancial as well as financial performance.
Table 8 provides results of the regression analyses, which reveal that, controlling for types of sector, types of ownership, types of organization, education, and duration with the organization in the model, three of the six SP practices dimensions was important predictors of organizational nonfinancial performance as well as financial performance. The SP practices dimension “purchases from small and local suppliers” was the most important and significant SP practice that positively effects of organizational nonfinancial performance (β = 0.246,
p ≤ 0.01) as well as financial performance (β = 0.207,
p ≤ 0.01), followed by the SP practices dimensions “environmental purchasing” and “philanthropy” with nonfinancial performance (β = 0.237,
p ≤ 0.01; β = 0.204,
p ≤ 0.01) and “philanthropy” and “safety” with financial performance (β = 0.200,
p ≤ 0.01; β = 0.164,
p ≤ 0.05). The rest of the SP practices dimensions did not have strong influence on organizational performance. These findings confirm that the impact of SP practices is not restricted to the nonfinancial and financial performance of an organization.
Table 9 shows that all dimensions of organizational nonfinancial performance were significant predictors of organizational financial performance. The nonfinancial performance dimension “quality and operation performance” was the most strongly related to organizational financial performance (β = 0.277,
p ≤ 0.01), followed by “social and environmental performance” (β = 0.235,
p ≤ 0.01) and “local and international business performance” (β = 0.192,
p ≤ 0.05). Indeed, the path from these scales to organizational financial performance was not significantly different from zero. The organizational nonfinancial performance dimensions accounted for 33 percent of the variance in organizational financial performance. Therefore, taken as a whole, the present study confirms that the impact of nonfinancial performance dimensions on financial performance of an organization is positive and significant.
In order to test H6—“the relationship between SP practices and financial performance is mediated by non-financial performance in an organization”—we carried out the regression-based path analysis described by Hair et al. [
41] as appropriate for such cases. Basically, regression-based path analysis involves the estimation of the partial coefficients in a path model. In other words, it determined the extent to which predictor variables have direct and indirect influences on the dependent variable. This estimation provided a test of whether the relationship between SP practices and financial performance is mediated by the organizational nonfinancial performance.
A regression analysis was performed using the summary measures of SP practices, organizational nonfinancial performance, and organizational financial performance to examine whether SP practices predicted organizational nonfinancial and financial performances. First, the scale of SP practices was regressed onto organizations’ nonfinancial performance; this model was significant (
F = 4.12,
p ≤ 0.01) (
Table 10). Multiple regression analysis found that when controlling for types of sector, types of ownership, types of organization, education, and duration with the organization in the model, the index of overall SP practices was significantly and positively related to nonfinancial performance (β = 0.417,
p ≤ 0.01; see
Table 10), explaining 25.3 percent of the variance in the nonfinancial performance construct. Second, the scale of SP practices was regressed onto organizational financial performance; this model was also significant (
F = 3.71,
p ≤ 0.01; see
Table 11). The result of this regression analysis found that, controlling for types of sector, types of ownership, types of organization, respondents’ educational level, and respondents’ duration with the organization in the model, SP practices were significantly and positively related to organizational financial performance (β = 0.258,
p ≤ 0.01), explaining 23.2 percent of the variance in the organizational financial performance construct.
Finally, both SP practices and organizational nonfinancial performance were simultaneously regressed onto organizational financial performance. The results in
Table 12 show that nonfinancial performance was a significant predictor (
p ≤ 0.01) of financial performance, while combined practices of SP was not. Only three of six SP practices dimensions showed evidence of mediation, using tests of indirect effects. The indirect effects from “environmental purchasing,” “philanthropy,” and “purchasing from local and small firms” factors were statistically significant. The partial effect (β = 0.463) of nonfinancial performance on financial performance was significant (
p ≤ 0.01), while the coefficient (β = 0.417) for the direct effect of SP practices on organizational nonfinancial performance was significant (
p ≤ 0.01). The direct effect of the summary measures of SP practices (taken together six scales of SP practices) on financial performance was not significant (β = 0.065;
p = 0.342), while the indirect effect of SP practices via organizational nonfinancial performance on financial performance was statistically significant (β = 0.193). Significant indirect effects indicate that organizational nonfinancial performance did significantly mediate the relationship between SP practices and organizational financial performance.