This study uses multiple linear regression models to analyze the influencing factors of companies’ carbon emission reduction behaviors which are taken as dependent variables.
2.2.1. Independent Variable Selection and Research Hypothesis
We divide the possible influencing factors into external pressure and internal driving force [28
]. Institutional theory is widely used to analyze the influence of policies and social external pressures [11
]. It states that institutional factors such as governmental regulation, market requirement and social expectations play an important role in companies’ behavior. In the socio-political sphere, Liu et al. believe that policies, social pressure and expectations from special stakeholders are often a decisive factor in corporate behavior, while competitors, suppliers and consumers are often the influential factors [28
]. Under environmental protection pressure, companies’ environmental strategies and management performance are affected by factors such as energy price [30
], market competition [31
], the carrying out of order and control policies or those with market incentives [32
], regional differences [33
This study takes three variables that reflect external pressure: energy price, competitive pressure, and whether the company is in the carbon pilot region. If the company believes that the current energy price is too high, it may be reluctant to put more efforts on the use of clean energy to reduce carbon emissions. If the competitive pressure is too large, companies may not be willing to reduce carbon emissions with excessive resources to avoid the increase in costs and guarantee the larger market share. Companies in carbon pilot regions will take more actions, because they may have a better understanding of carbon market policies, show more awareness of the importance of energy-saving and carbon emission reduction and have larger mitigation pressure, compared with other enterprises.
Regarding the external pressure that companies face, the following hypotheses are presented:
Hypothesis 1a (H1a).
The larger the pressure on energy prices, the less companies will engage in carbon reduction behavior.
Hypothesis 1b (H1b).
The larger the pressure on market competition, the less companies will engage in carbon reduction behavior.
Hypothesis 1c (H1c).
Companies in pilot region will take more actions on carbon emission reduction compared with others.
However, institutional theory cannot explain differences in behavior between companies facing the same external pressure [34
]. The theory of rational action [12
] and the theory of planned behavior [13
] state that attitude, subjective norms, and perceived behavioral controls can affect individual behavior. As a result, for companies’ decision-making on behavior, instead of external pressure, a company’s own resources and capacities [35
] as well as its perceived situation for external pressure [14
] need to be considered. The greater potential companies have in enhancing energy efficiency and reducing carbon emission, the more room they have to reduce costs brought by climate policies [36
]. Companies will act more with clear and accurate expectation on carbon market policies [26
This study incorporates five variables representing the internal driving forces of companies: the investment in technological transformation projects of energy-saving and carbon emission reduction, the familiarity of mitigation technology, the familiarity of carbon emission trading market policies, the potential for carbon reduction, and the expectation of national carbon market policies. When larger investments are made, it indicates more attention has been paid on environmental management, and more actions will be taken to reduce carbon emission by the company. Companies will do more with a further understanding of technology, policies, and larger mitigation potential. Corporate expectations on the time and scale of the national carbon emission trading scheme reflect the pressures of carbon emission reduction. If they believe the establishment of the scheme is urgent and with high reduction requirements, they may take a variety of measures in advance to ensure a better response.
In terms of companies’ internal driving force, the following hypotheses are presented:
Hypothesis 2a (H2a).
As investment are made in improving energy saving and carbon reducing technology, the more inclined companies are to engage carbon reduction behavior.
Hypothesis 2b (H2b).
As companies become familiar with carbon reducing technology, the more inclined they are to engage carbon reduction behavior.
Hypothesis 2c (H2c).
As they become familiar with the policies of carbon emission trading scheme, the more inclined they are to engage carbon reduction behavior.
Hypothesis 2d (H2d).
The larger carbon reduction potential companies have, the more inclined they are to take carbon reduction behavior.
Hypothesis 2e (H2e).
The sooner companies expect carbon market policies to be put into effect, the more inclined they are to act.
Hypothesis 2f (H2f).
The larger the expected requirement of carbon reduction, the more inclined companies are to acts.
Enterprise scale also has great influence on company environmental strategies. The larger the scale of an enterprise, the more equipped it is for pollution control and better environmental performance [33
]. The scale of companies can be characterized by the number of employees. The industry-effect will also have an impact on the behavior of companies [33
]. As most respondents are private enterprises, this study does not take the impact of ownership type into account. A company’s scale and economic sector are the two control variables, which reflect the characteristics of the firm.
In terms of control variables, the following hypothesis is made:
Hypothesis 3a (H3a).
The larger a company’s scale is, the more inclined they are to act on carbon emission reduction.
2.2.2. Variable Processing and Assignment
As mentioned above, in the regression analysis, there are four types of companies’ carbon emission reduction behaviors: mitigation from combustion processes (), mitigation from production processes (), mitigation by reducing output () and strengthening environmental management (). The value of dependent variable is the average score of each type of carbon emission reduction behaviors in the five-level Likert scale.
The name, symbolic representation, description and value of independent variables are shown in Table 3
. For example, “
” demonstrates the pressure on energy price which is investigated by asking companies to score their evaluation of the economic pressure of energy price from 1 to 4, while 1 means the pressure is low and 4 means high. Continuous variables can be directly put into the regression model. Some categorical variables, including
can be assigned according to the actual meaning and added into the regression model as continuous variables; this simplifies the model. Variables that cannot be assigned based on the meaning, for example
, are transformed into dummy variables.