(1) Innovation input in technology research and development stage

In order to explore the relationship between the two in depth, model (6) is used to test Hypothesis H1a to verify the relationship between environmental uncertainty and technological innovation investment. Column (1) of Table 3 shows environmental uncertainty (EU) and technological innovation regression results in the technological development stage, and the regression coefficient between EU and technological innovation input (R&D) is −0.521, which is significantly negatively correlated at the 1% confidence level, indicating that EU has an inhibitory effect on innovation input (R&D). It can be seen that the greater the environmental uncertainty, the more significant the inhibitory effect on enterprises’ investment in technological innovation, which verifies Hypothesis 1a.

 

In the achievement transformation stage, using models (6)–(9) to test Hypothesis 2 to verify the effect of environmental uncertainty on firms’ innovation output and the mediating effect of the second type of agency problem between environmental uncertainty and technological innovation output. The results of the multiple regressions are shown in Table 5. Column (1) of Table 5 shows the regression results of environmental uncertainty (EU) and technological innovation in the achievement transformation stage. Innovation output (LPatents), which is the natural logarithm of the number of patent applications, is a continuous variable, and the OLS model was used to test Hypothesis 1. In column (1) of Table 5, the regression coefficient of EU and innovation output (LPatents) is −0.612, which is significantly negative at the 10% confidence level, indicating that EU has an inhibitory effect on innovation output (LPatents). Further testing the mediating effect, in column (2) of Table 5, the correlation coefficient between agency and innovation output (LPatents) is −3.934, which is significantly negative at the 5% confidence level, indicating that the second type of agency problem is prevalent among firms, which affects their technological innovation output and plays an inhibitory role on innovation. As can be seen in column (3) of Table 5, the regression coefficient between Agency and EU is 0.006, which is significantly positive at the 1% level, that is, the higher the environmental uncertainty, the more serious the second type of agency problem of firms. In column (4) of Table 5, it can be seen that in model (9), the coefficient of agency with firms’ technological innovation output (LPatents) is −3.804, which is significantly negative, and the coefficient of environmental uncertainty (EU) with firms’ technological innovation output (LPatents) is -0.586, which is significantly negative at the 10% level. Combining the regression results of the previous models, it can be concluded that the greater the environmental uncertainty, the more unfavorable the technological innovation output of the firm, and also, it indirectly has a negative impact on the technological innovation output of the firm by exacerbating the second type of agency problem. This verifies Hypothesis 2.

Among the control variables, it can be seen from column (1) of Table 3 that the regression coefficients of enterprise size (SIZE) and profitability index (ROA) are 0.789 and 2.791, respectively, which are both significantly positive at the 1% level, indicating that the larger the scale, the better the profitability. The better innovation foundation and innovation resources companies have, the more they invest in innovation activities. The regression coefficient of investment opportunity (TOB) is positive, indicating that the more investment opportunities a company has, the more likely it is to choose an innovative project to invest in, thereby promoting the company’s input in technological innovation. The regression coefficient of the net business cycle (JYYZ) is significantly negative, which shows that the longer the net business cycle, the more unfavorable the company’s innovation investment.

(2) Innovative output at the stage of achievement transformation

In order to explore the relationship between the two in depth, model (6) is used to test Hypothesis 1b to verify the impact of environmental uncertainty on the innovation output of the enterprise in the transformation stage of the results. The results of multiple regression are shown in Table 3. Column (2) of Table 3 shows the regression results of environmental uncertainty (EU) and technological innovation in the achievement transformation stage, and innovation output (LPatents) is the natural logarithm of the number of patent applications, which is a continuous variable, using the OLS model test Hypothesis 1b. In column (2) of Table 3, the regression coefficient between EU and innovation output (LPatents) is −0.612, which is significantly negative at the 10% confidence level, indicating that EU has an inhibitory effect on innovation output (LPatents). It can be seen that the greater the environmental uncertainty, the more unfavorable the technological innovation output of the enterprise, which verifies Hypothesis 1b.

Taking into account the data characteristics of the number of patent applications in China, the explained variables of the model are converted from continuous variables to the number of patent applications to explain. As patent data have many values of 0, the Poisson model is more consistent with the number features, and multiple regression models make the regression results more robust. From column (3) of Table 3, it can be seen that the regression coefficient between environmental uncertainty (EU) and technological innovation output (Patents) is −0.344, which is a significant negative correlation, indicating that EU’s influence on innovation output (Patents) has an inhibitory effect. It can be concluded that the greater the environmental uncertainty, the more unfavorable the technological innovation output of the enterprise, which verifies Hypothesis 1.

In the technology research and development stage, models (6)–(9) were used to test Hypothesis 2 to verify the relationship between environmental uncertainty and technological innovation input and the mediating effect of the second type of agency problem on environmental uncertainty and technological innovation input. From column (1) of Table 4, it can be seen that the regression coefficient between environmental uncertainty EU and technological innovation input (R&D) is −0.521, which is significantly negatively correlated at the 1% confidence level, indicating that EU’s contribution to innovation input (R&D) has an inhibitory effect. Further testing the mediation effect, from column (2) of Table 4, it can be seen that the correlation coefficient between agency and innovation input (R&D) is −2.693, which is significantly negative at the 1% confidence level, indicating that the second type of agency problem is common among enterprises and affects enterprises’ input in technological innovation and inhibits innovation. From column (3) of Table 4, it can be seen that the regression coefficient between agency and the environmental uncertainty EU is 0.006, which is significantly positive at the 1% level, that is, the higher the environmental uncertainty, the more serious the enterprise’s second type of agency problem. From column (4) of Table 4, we can see that in model (9), the coefficients of agency, Environmental Uncertainty (EU) and the company’s technological innovation input (R&D) are −2.606 and −0.505, respectively, which are both significantly positive at the 1% level. Combining the regression results of the previous models, we show that the greater the environmental uncertainty, the more significant the inhibitory effect on the enterprise’s technological innovation investment. In addition, it exacerbates the second type of agency problem, which has a negative impact on enterprises’ investment in technological innovation indirectly. This verifies Hypothesis 2.

Table 6 shows the regression results of the Poisson model. It can be seen from column (1) of Table 6 that the regression coefficient between environmental uncertainty (EU) and technological innovation output (Patents) is −0.344, which is significantly negative, indicating that EU’s impact on innovation output (Patents) has an inhibitory effect. To further test the mediation effect, in the column (2) of Table 6, the correlation coefficient between agency and innovation output (Patents) is −4.682, which is significantly negative, indicating that the second type of agency problem generally exists between different companies and affects the company’s output of technological innovation, inhibiting innovation. In column (3) of Table 6, the regression coefficient between agency and EU is 0.006, which is significantly positive, that is, the higher the environmental uncertainty, the more serious the second type of agency problem of the enterprise. In column (4) of Table 6, using model (9) to test, the coefficient of agency and firm’s technological innovation output (Patents) is −4.603, which is significantly negative. At the 1% confidence level, the environmental uncertainty EU and the technological innovation output (Patents) of the enterprise are significantly negative with a regression coefficient of −0.321. Combining the regression results of the previous models and using the principle of mediating the effect test, it is found that the greater the environmental uncertainty, the more unfavorable the technological innovation output of the enterprise. In addition, it also indirectly negatively affects the technological innovation output of enterprises by exacerbating the second type of agency problem. This verifies Hypothesis 2.