Does negative feedback affect a firm’s R&D efficiency? An empirical study of Chinese listed firms

ABSTRACT

Based on the behavioral theory of the firm, we investigate the impact of negative feedback on R&D efficiency. The results show that when performance falls below aspirations, firms pay more attention to R&D projects with high efficiency to get rid of the operating pressure and reputation pressure caused by the decline in performance. This effect may be influenced by resource differences. In the case of negative feedback, firms with fewer idle resources have more incentive to improve R&D efficiency. Compared with state-owned firms and high-tech firms with higher resource acquisition ability, the impact of negative feedback on R&D efficiency is statistically significant only in the sample of private firms and non-high-tech firms. A firm’s strategy to increase its R&D efficiency may be correct, as the improvement in R&D efficiency effectively reverses investors’ expectations of the firm’s future market value.

KEYWORDS:

• Performance feedback
• aspirations
• negative feedback
• R&D efficiency
• resources

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. The specific calculation process of R&D efficiency: the input (total amount and intensity of R&D input) and output variable (number of patent applications) of all firms are sorted out. The data is then imported into the DEA software, which calculates the R&D efficiency directly without any command. During the calculation process, the weights of the input and output are automatically obtained.

2. For example, if the actual performance is lower than aspirations by 0.009 (P_i,t-1-AP_i,t-=-0.009), then I₁  = 1, BHA=I₁×(P_i,t-1-HAP_i,t-1) = 1×(-0.009)=-0.009. If the actual performance is higher than aspirations by 0.009 (P_i,t-1-AP_i,t- = 0.009), I₁   = 0, then BHA=I₁×(P_i,t-1-HAP_i,t-1) = 0×(-0.009) = 0 is excluded. By calculating the equation we can get continuous negative values.

3. The normalization formula for the variable is: (x-m)/std, where m is the mean of x and std is the standard deviation of x.

4. Slack index = Normalized absorbed redundancy + Normalized unabsorbed redundancy + Normalized potential redundancy.

Additional information

Funding

This work was supported by Chongqing Social Science Planning Fund [2021BS084]; Social Science Research Fund of Chongqing Education Commission [23SKGH181]; Chongqing Technology and Business University Research Start-up Fund [2255006]