5. Estimation results
5.1. Correlations between errors
The estimation results show that the error terms in the assess equation and in the require equation are positively and significantly correlated; the estimated correlation is 0.554 with the standard error of 0.059. These findings suggest that similar unobservables affect both ASSESS and REQUIRE.
The results also indicate the endogeneity of ISO 14001 in the assess equation. The correlation between the error terms in the ISO equation and in the assess equation is estimated to be 0.576 with the standard error of 0.184. The negative correlation is in line with findings in previous studies. For example, Arimura et al. [2] find that ISO 14001 is endogenous in the environmental performance equation. They find a negative correlation between the error terms in the ISO 14001 adoption equation and those in environmental performance.
By contrast, ISO 14001 is found to be exogenous in the require equation; the error term in the ISO equation is not significantly correlated with that in the require equation (the estimated correlation is 0.259 but the standard error is 0.224). These findings suggest that a univariate model indeed suffices for consistent estimates of the require equation. After presenting the results of the multivariate model, we will compare them with the results of the univariate model.
. ISO equation
Estimated coefficients of the ISO equation and the corresponding average partial effects are presented in Table 3 (columns 5 and 6, respectively). We summarize these results only briefly here because our focus is on the assess and require equations. We find that facilities that are large, have a quality management system in place, or have a global market scope are more likely to adopt ISO 14001. These results are consistent with those in previous studies [2,22,31]. We also find that facility managers for whom corporate profile/image is an important motivation for environmental practices are more likely to adopt
ISO 14001.
5.3. Assess and require equations
Columns 1 and 3 of Table 3 present estimated coefficients of the assess equation and the require equation. Columns 2 and 4 of Table 3 show the corresponding average partial effects. The coefficients for the adoption of ISO 14001 in both equations are positive and highly significant, indicating that the adoption of ISO 14001 increases both the probability of assessing suppliers’ environmental performance and the probability of requiring suppliers to undertake environmental measures. The effects seem to be large; the average partial effects are estimated to be 42.2% for ASSESS and 50.6% for REQUIRE.
Our findings suggest the possibility of ISO 14001’s positive externality in that a facility’s adoption of ISO 14001 may positively affect the environmental performance of its suppliers. This is because the adoption of ISO 14001 has positive effects on the implementation of GSCM practices, which in turn may make it likely that suppliers reduce their environmental impacts.7
It should be stressed that all previous empirical studies assessing the environmental benefits of ISO 14001 have estimated its impact on the environmental performance of the adopters alone. These studies therefore do not capture the entire effects of ISO 14001 and may underestimate the total environmental benefits of EMS adoption and certification.
In addition to ISO 14001, other variables also influence the probability of assessing suppliers’ environmental performance. For example, the facilities with quality management systems are 8.5% more likely to assess the environmental performance of their suppliers. Facilities that report input bans are applicable in terms of their production activities are also 8.6% more likely to assess the environmental performance of their suppliers. These results are consistent with the expectation that facility managers who face input bans will be wary of using noncompliant components in their products. For example, Sony had to recall 1.3 million exported game machines when Dutch authorities determined that cadmium levels in the peripheral cables, provided by suppliers, did not meet environmental standards.8 Since then, numerous manufactures, including Sony, have taken actions to prevent such incidents.
The results also show that facility managers who consider new product/technology development as moderately or very important motivators with respect to the environmental practices of their facility are more likely to assess their suppliers’ environmental performance. These findings are consistent with evidence suggesting that R&D promotes more environmental practices [1], since GSCM can be considered one type of environmental practice.
Other results, however, seem counterintuitive. For example, the coefficient on the dummy variable for global market scope is negative and significant. That is, the facilities exporting to global markets seem to be less likely to assess their suppliers