*2.3. Hypotheses*

Since Schumpeter [41] proposed the innovation theory, innovation performance has been a research hotspot. Innovation investment not only improves the core competitiveness of the enterprise, but also significantly improves the productivity of the enterprise [42]. Meanwhile, it brings new products or new technology to the enterprise, which makes the enterprise operate di fferently and beneficially to improve the market share of the enterprise. On the one hand, due to the long cycle and uncertainty of innovation and R&D, especially in the field of advanced technology, the economic returns brought by innovation investment may lag many operation years [43], but in the long run, its cumulative effect will significantly promote the improvement in enterprise performance. On the other hand, based on the principal-agent theory and the managemen<sup>t</sup> defense hypothesis, the decision-making behavior of managemen<sup>t</sup> is subject to annual performance appraisal, while high-risk innovative projects may lead to higher R&D expenditure. To pursue the improvement of financial performance during the term of o ffice, managemen<sup>t</sup> may limit the scale of R&D expenditure, resulting in short-sighted behavior of self-interest, which is not conducive to the long-term development of enterprises. Therefore, this paper argues that there is an endogenous relationship between corporate innovation investment and sustainable financial performance.

As di fferent industries have di fferent forms of R&D, the extent of performance improvement brought by innovation investment will be di fferent. For technology-intensive energy enterprises, their development is based on technological innovation, which is the source of enterprise life, thus the input-output benefit of innovation is obvious; for capital-intensive energy enterprises, scale e ffect should be an important strategy for their development; and for labor-intensive energy enterprises, service mode change and managemen<sup>t</sup> process innovation could bring more profits for enterprises. This leads to the phenomenon of some enterprises investing a lot in innovation while others do not have R&D activities. Therefore, it is necessary to analyze the relationship between innovation investment and financial performance from the industry level. Based on the above analysis, this paper proposes the following hypotheses:

**H1:** *Innovation investment has a positive e*ff*ect on financially sustainable performance, and this e*ff*ect is the most significant in technology-intensive energy enterprises*.

**H2:** *There is an interactive endogenous relationship between innovation investment and financially sustainable performance, and financially sustainable performance has a reverse e*ff*ect on innovation investment, and the reverse e*ff*ect is the most significant in technology-intensive energy enterprises*.

From the perspective of the principal-agent theory, the managemen<sup>t</sup> will adopt the egoism behavior that pays attention to the short-term economic benefit based on their own interest, and ignore or avoid the innovative RD investment that has risk uncertainty but is helpful to improve the sustainable operation ability of enterprises. Therefore, enterprises must implement some incentive policies to enhance the motivation of managemen<sup>t</sup> to make innovative RD decisions, such as increasing short-term compensation returns or implementing equity incentive plans, so that their personal interests and the company's interests converge. Based on the above analysis, this paper proposes the following hypotheses:

**H3:** *Salary incentives have a significantly positive moderating e*ff*ect on the relationship between RD investment and financially sustainable performance*.

**H4:** *Equity incentives have a significantly positive moderating e*ff*ect on the relationship between RD investment and financially sustainable performance*.

#### **3. Proposed Methodologies**

#### *3.1. Energy Enterprises Background*

With the continuous development of the economy, the production factor density is widely used in the research field of enterprise classification. It was first proposed by Heckscher and Ohiln in H-O theory [44], which mainly refers to the degree of influence on production and managemen<sup>t</sup> activities and the degree of dependence on various production input factors in the production process. The classification method of production factor intensity categorizes enterprises into three types, including technology-intensive, labor-intensive and capital-intensive enterprises. This classification could not only reveal the productivity and resource advantages of enterprises, but also reflect the changes in the proportion of production input factors brought about by technological progress.

Technology-intensive enterprises refer to enterprises with large investment in technical knowledge, high costs for RD, high cultural and technical level of workers, and high added-value of products, including, for example, renewable energy enterprises; capital-intensive enterprises refer to enterprises

with large investment capital and grea<sup>t</sup> influence on production and operation activities, including, for example, electric power enterprises; labor-intensive industries refer to enterprises with large labor input and grea<sup>t</sup> influence on production and operation activities, including, for example, coal enterprises.

#### *3.2. Data Source*

Based on the CSMAR database, this paper collected data from the energy sector in the stock market. The executive incentive data and enterprise financial performance data were all from the CSMAR database; the RD investment data were obtained from the annual reports of each enterprise; some missing data of financial indicators were obtained through the sorting of the Tonghuashun database; and other relevant missing data were obtained from the WIND database and www.cninfo.com.cn. The statistical analysis was performed using Stata and Excel statistical software.

We used Ward linkage method in the cluster analysis to classify samples. According to the proportion of fixed assets, enterprises with a larger proportion of fixed assets belong to the capital-intensive companies, indicating that the capital is of higher importance. In terms of the RD expenditure-salary ratio, enterprises with higher ratios belong to the technology-intensive companies, indicating that the technological is of higher importance and others belong to the labor-intensive companies.
