**4. Results**

In our study, the values of the dependent variable were measured a year after those of the independent variables. Consequently, our dependent variable (environmental performance) was measured from 2014 to 2018, while the control and independent variables were measured from 2013 to 2017. This time-lapse of one year between when the dependent and independent variables were measured served to reduce the risks which derive from inverse causality. In fact, strategic change sometimes leads to better green performance. Pettigrew and Whipp suggest that such a time lapse between the disclosure of independent and dependent variables would seem appropriate because the positive consequences of changes in strategy may not become evident for some time [72]. The five years of measuring variables resulted in data for a panel with 265 different combinations of the values of our variables (*environmental performancet*, *board independencet-1*, *erp spending t-1*, *financial performancet-1*, *firm sizet-1*, *firm aget*-*1*, *Leveraget-1, Tobin's Qt-1*, *polluting industryt-1*, where *t* is the conclusion of a generic year from between 2014 and 2018), a combination for each of the firm-year observations that make up the sample (5 years × 53 firms). The standard deviations, the average values and the values of the Pearson correlation coefficient relative to the variables used in our analysis are presented in Table 1, which also shows some meaningful correlations between the variables when considered as pairs.

**Table 1.** Mean averages, standard deviations, and correlation matrix.


N = 265; 1-tailed: \* *p* < 0.05; \*\* *p* < 0.01.

A hierarchical regression model was run so that our hypothesis could be tested, and the results are presented in Table 2. Initially, the variance inflation factor (VIF) for every one of the independent variables within the regression models was calculated in order to guarantee that no potential multicollinearity problems existed with the variables. However, as the VIF values all lay within a range of between 1.1 and 1.7, it was clear that they had no influence on the validity of our three models [73].


**Table 2.** Hierarchical regression analysis of environmental performance (N = 265).

N = 265; 1-tailed: † *p* < 0.10; \* *p* < 0.05; \*\* *p* < 0.01.

We hypothesised that variables for board independence and ERP spending were, on the one hand, individually capable of producing positive effects on corporate environmental performances (H1 and H2) and, on the other, able to improve these positive effects further together (H3). Therefore, we started by placing only the Model A control variables in Table 2. We then carried out ordinary least squares (OLS) regression analysis and presented the results in the first two columns of Table 2. Next, we also added the independent variables from the testing of our hypotheses to the control variables so as to carry out another ordinary least squares (OLS) regression analysis in Model B, the results of which are reported in the third and fourth columns of Table 2. The results for the variable of board independence show that a firm's green performance is positively influenced by its having a large proportion of its board constituted by independent directors. This is a statistically significant result (β = 1.23, *p* = 0.008) and is in line with Hypothesis 1, which was formulated according to agency theory predictions. Hypothesis 2 focuses on how ERP technologies are put to use, and the prediction was that they would affect the way our sampled firms perform environmentally. However, this hypothesis is not supported by our results (β = 0.961, *p* = 0.239). Finally, in columns 5 and 6 of Table 2, the results of the addition of the term for the interaction between the use of ERP technologies and the share of independent directors on boards are reported. The interaction is statistically significant (β = 0.141, *p* = 0.009). The addition of this interaction term to Model 3 gives an explanatory contribution above and beyond effects-only Model 2. Explained variance increases by 2.1%, and this increase is statistically significant (Fchange = 4.211, *p* < 0.01). Therefore, this empirical analysis provides support for Hypothesis 3. The results found in the three steps (Model A, Model B, and Model C) are significant and robust. As is evident from Table 2, all models are significant (at *p* < 0.01), with R<sup>2</sup> ranging from 0.156 for the base model to 0.192 for the full model. In particular, the full model (Model C) is fit and explains about 19.2% of the variance, with Fsign = 6.723 and significance at the 0.01 level.

#### *Robustness Checks*

The Breusch and Pagan heteroscedasticity test was applied to the outcome of the multiple OLS regression analysis (Models A, B, and C, Table 2) so as to test whether our model is robust [74]. Table 3 presents the results from an auxiliary regression of the Breusch–Pagan test.


**Table 3.** Heteroskedasticity Test.

Note: N = 265.

#### **5. Discussion and Conclusions**

The aim of this paper was to establish whether the corporate governance structure and digital technologies a firm adopted could have an impact on its environmental performance. We began our analysis by looking at the composition of the board of directors. From an agency point of view, the more independent the members of a board of directors are, the more effectively the board will perform its internal control function of offsetting any agency problems the firm might have. Consequently, we formulated Hypothesis 1 because we expected greater involvement of independent directors in the board's activities to increase the firm's likelihood of achieving good environmental performance. ERP-type digital technologies provide functionalities that can be used relatively easily to implement practices and provide functionalities that can help manufacturers to establish their green supply chain. ERP systems might also be useful as support for environmental accounting instruments. Consequently, we formulated Hypothesis 2, according to which increased use of ERP systems can bring about better environmental performance. Finally, we formulated Hypothesis 3 to check the environmental effects of board independence and ERP technologies considered in combination and no longer individually.

Our predictions were tested by taking a sample of 53 firms that were quoted on the Italian stock exchange in Milan and looking at their end-of-year reports (Sustainability Reports/non-financial declaration) for a period of 5 years, for a total of 265 firm-year observations. Sampled firms were also asked for data on the costs they faced in implementing ERP digital technologies. The results of our analysis:


environmental performance is in addition to the impact brought about by the board's greater independence (considered alone as in for the verification of Hypothesis 1).

The verification of Hypothesis 1 and the non-verification of Hypothesis 2 may be considered due to the differing behaviour that independent directors and managers have toward the pursuit of green strategies. Hill and Jones point out that managerial behaviour is influenced by particular utility functions, which means that the managers might not consider it convenient to invest a lot of their high-quality time in redesigning the company's internal procedures simply to reduce the causes of the firm's polluting in the future [42]. This circumstance may negatively influence results which, from an environmental point of view, might be achieved by a firm's implementation of an ERP system. For such a system to be successful, the aims and expected results of its implementation should first be clearly defined and shared within the organisation, and then the efforts necessary to change management and reproject the firm's processes and operations should be undertaken [56,57]. Managers' utility functions, which are in contrast with the attribution of pollution control aims to ERP systems, could provide an explanation for the failure to test Hypothesis 2. Nevertheless, from an Agency Theory point of view, the powers that managers have to control firm activity are counterbalanced by board independence. Through the board, independent directors compensate for the behaviour of managers who are only interested in short-term interests by forcing them to consider the long-term interests of shareholders and stakeholders [34,35]. Literature on agency theory indicates how an independent board of directors is more likely to recognise the potential that green investments may have in the long term and, therefore, resist managerial pressure to adopt a different investment strategy. Consequently, there is a greater inclination on the part of independent boards to adopt environmentally-friendly policies, even where they are expensive [15]. These aspects might provide an explanatory verification of Hypothesis 1. Today, many stakeholders require disclosures of corporate responsibility and environmental performance in particular [14,75]. Society expects firms to take greater environmental responsibility for their activities. These expectations correspond to, on the one hand, new opportunities that are emerging for those firms which are most attentive to environmental performance and, on the other, more severe sanctions and fines for those firms that commit environmental crimes [76]. The Exxon Valdez and BP oil spills (of 1989 and 2010, respectively) are two examples of how environmental disasters may lead to harsh financial consequences for a firm that does not pay attention to environmental problems. Adopting environmentally responsible business practices ought to be a primary consideration for a firm's board of directors and owners. One explanation for the verification of Hypothesis 3 is that reinforcing a board's independence releases the potential that ERP digital technology has to improve environmental performance. A more independent board of directors will be more effective in blocking and overcoming management resistance to making the efforts necessary to redesign company processes and adequately configure ERP systems so as to obtain maximum advantage for the environment, too. Some boards know how to shape behaviour that helps an organisation use data and digital technologies in ways that are perceived as socially, economically, and environmentally responsible. In this sense, a board may constitute an antecedent with respect to CDR.

Our study is not without its limitations. We opted for a sample and a hierarchical regression model that were capable of explaining a part of the complexity of the entire phenomenon. However, a firm's green performance and control of pollution are complex phenomena, while governance mechanisms and ERP digital technology only represent a limited part of the variables affecting a firm's environmental achievements. Finally, the data for this study were gathered in Italy. Therefore, special attention should be given when generalising with regard to other national contexts on the basis of these discoveries. In Italy, firm ownership is generally rather concentrated, and family firms make up the most sizeable set of blockholders on the stock market, while the next largest set consists of the state or other public bodies [76].

Our conceptualisation paves the way for further future research, especially regarding antecedents to CDR and their relationship to corporate governance. It would also be advantageous to extend the empirical research to include other countries, and analysis should be made of further possible relationships between CDR and other characteristics of board composition, such as CEO duality or interlocking directorates. Further food for thought might emerge from the drawing up of environment, social, and governance (ESG) parameters that take account of CDR problems. With regard to these, particular attention ought to be given to the malleability of the technology and data, which, as this paper has shown, may be positively influenced by effective governance mechanisms.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of e-Campus Telematics University (protocol code 03/2018).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** Author declares no conflict of interest.
