Internal Control, Environmental Uncertainty and Total Factor Productivity of Firms—Evidence from Chinese Capital Market

Abstract

Based on the data of China’s A-share listed companies from 2009 to 2019, this paper empirically examines the relationship between internal control and total factor productivity of enterprises in the presence of environmental uncertainty. The research shows that high quality internal control can effectively improve the total factor productivity of enterprises. Environmental uncertainty negatively moderates the relationship between internal control quality and total factor productivity. Further research in this paper shows that there are heterogeneous effects on the adjusting effects of different life cycles, property rights and regional distribution of enterprises, that is, when enterprises are in the growth stage, the quality of internal control has a significant effect on the improvement of total factor productivity of enterprises. For enterprises and non-state-owned enterprises located in the eastern region, the inhibition and adjustment effect of environmental uncertainty is more significant. At the same time, the supplementary research finds that internal control directly affects the total factor productivity of enterprises through the intermediary role of promoting enterprise development and innovation and easing financing constraints. The research conclusions enrich the literature on the mechanism of internal control affecting enterprises’ total factor productivity and provide a new reference and basis for enterprises to effectively manage the internal environment, strengthen the risk control management mechanism and improve the enterprise value.

1. Introduction

In April 2010, the Ministry of Finance of China jointly issued internal control guideline documents such as the Guidelines for the Application of Internal Control in Enterprises, which require that listed companies in China conduct self-evaluation of the effectiveness of internal control and disclose annual self-evaluation reports, as well as engaging accounting firms to audit the effectiveness of internal control over financial reporting and issue audit reports. The report focuses on the effectiveness of internal control, and promoting the internal control system has been improved in five aspects: internal management, risk assessment, preventive control, information transmission and internal supervision, and the internal control index reflects the situation of enterprises more accurately. This means that Chinese listed companies have officially entered a new period of standardized internal control management.

At present, the global epidemic is creating a long-term trend in development. In the context of the transformation of the economic growth mode and the fluctuation of the external environment, China is facing enormous pressure for economic development. Many uncertain factors in the ecological, international and economic environment have contributed to the uncertainty of the macro environment. For the Chinese economy, which is in the critical period of development transformation, how to ensure the sustainable and high-quality development of the economy is particularly urgent. However, the realization of macroeconomic goals must be based on development at the micro level. How to further optimize the structure of enterprises at the micro level and improve the management level to improve the quality and efficiency of production is an important issue to be solved urgently. Therefore, it is clearly pointed out in the report that we should accelerate the construction of a modern economic system and focus on improving total factor productivity (TFP). Enterprise total factor productivity refers to the additional production efficiency under the given factor level. It can comprehensively consider the contribution of various input factors to enterprise output [1], so it can be used as an important indicator representing the high-quality operation efficiency of enterprises. Shen Danhong (2022) pointed out that the average growth of total factor productivity of the manufacturing industries in China’s six regions during 2003–2019 was only 3.46%, but the contribution rate to the added value of the manufacturing industry was 33.18%, and after 2018, the growth rate of total factor productivity of enterprises reached 11% [2], indicating that in the process of China’s future development, the total factor productivity of enterprises will exert greater economic contribution. Therefore, how enterprises optimize the internal environment, improve the total factor productivity of enterprises and maintain high quality development has become a problem studied and discussed within enterprises and by many scholars.

Current studies mainly discuss the influence of total factor productivity on the external environment and internal enterprise. With regard to the external environment, Duan and Li (2019) showed that economic policy uncertainty significantly inhibits the total factor productivity of firms [3]. Harrish and Trainor (2005) pointed out that compared with other methods, government subsidies can effectively improve the total factor productivity of subsidized enterprises [4]. Based on the internal environment, Zhao (2015) shows that financing constraints reduce the efficiency of enterprise resource allocation and have a negative impact on total factor productivity [5]; Sheng and Jiang (2019) find an inverted “U-shaped” association between executive monetary compensation incentives and enterprise total factor productivity [6]. Qing Li et al. (2021) found that information management ability can improve risk prediction, performance incentive, integrated innovation and, finally, effectively improve the total factor productivity of enterprises [7]. According to the study of Guo Mengnan and Li Xiaohong (2020), internal control can effectively alleviate resource mismatch, improve organizational efficiency and promote the growth of enterprise value [8].

To sum up, at present, domestic researchers rarely explore the direct relationship between the concept of internal control within enterprises and the total factor productivity of enterprises. Meanwhile, there are few studies that take environmental uncertainty as the research angle of total factor productivity of enterprises. Under the current situation, the operating efficiency of enterprises cannot ignore the impact of changes in the external environment. Environmental uncertainty may lead to sharp changes in market demand, intensify competition and information asymmetry, and seriously affect the development of all aspects. Therefore, from the perspective of internal control quality, this paper explores the optimization mechanism of total factor productivity and the influence mechanism of environmental uncertainties on the relationship between the two factors, which has strong practical significance and application value.

In view of this, this paper takes China’s A-share listed companies from 2009 to 2019 as samples to explore the mechanism of internal control quality on the total factor productivity of enterprises in the presence of environmental uncertainty. The main innovations and contributions of this paper are as follows: (1) Enrich the objective research on the factors that affect enterprise TFP. Most existing literature uses internal control as evidence from the perspective of intermediary, and few studies demonstrate the direct relationship between internal control quality and enterprise TFP. This paper expands the research horizon and explores its direct effect on total factor productivity from the perspective of internal control quality. (2) Enrich the discussion about the economic consequences of internal control and make reasonable supplement to the existing literature. OLS analysis was adopted in this paper to test the influence and mechanism of internal control on corporate development. (3) Explore the environmental regulation effect between internal control and total factor productivity of enterprises in depth and select enterprise innovation and financing constraints as intermediate variables to explore the mediating effect, which makes up for the deficiency of relevant literature research. (4) Certain practical significance. Combined with the background of China’s economic focus shifting to high-quality development, this paper provides a channel for how to improve the internal environment of enterprises at the micro level so as to improve the total factor productivity of enterprises and also provides a new idea for how to form an internal control system at the macro level.

The rest of the basic structure of the paper is arranged as follows: The second part is the literature review and research hypotheses. The third part is the study design. The fourth part is the analysis of the empirical results. The fifth part is further research. The sixth part is research conclusions and implications.

2. Literature Review and Research Hypotheses

2.1. Internal Control and Total Factor Productivity of Enterprises

The total factor productivity of enterprises is mainly affected by internal factors such as technological innovation, resource allocation and information quality. As an effective operating mechanism of enterprise environment, internal control, with value creation as its main guidance, plays an important role in the implementation of internal management decisions, resource allocation, process supervision and other major links [9], thus affecting the total factor productivity of enterprises. In recent years, many studies have explored the influence mechanism of internal control on the economic consequences of enterprises from multiple perspectives, among which the influence of enterprise innovation and financing constraints is very significant. In terms of enterprise innovation, high-level internal control means that enterprises can reasonably control risks at the innovation level, and avoid unnecessary risks in a timely manner by establishing a sound internal control risk assessment mechanism, supervision and evaluation mechanism [10,11], so as to improve the possibility of innovation success. Meanwhile, the effective implementation of internal control can alleviate the degree of information asymmetry among the members of the company, restrain the agency conflicts between the board of directors and managers, shareholders and minority shareholders, enhance the internal innovation willingness of the company, help improve the driving force of innovation and promote the increase in innovation investment of the company. Under the dual effects of increasing innovation input and improving the possibility of innovation success, companies can fully transform R&D resources into new technologies and products [12], enhance market competitiveness, improve enterprise technical level [13] and thus improve total factor productivity of enterprises.

In terms of financing constraints, it refers to the inability of enterprises to obtain external financing due to capital market friction or the high cost of external financing, forcing enterprises to give up favorable investment opportunities [14]. High-quality internal management enterprises have lower capital cost and systematic risk [15] and better management of accounting information quality and liquidity, which can well reflect the profitability and operation of enterprises. Therefore, it can effectively improve the possibility of external investors to invest in the company and further improve the operation efficiency and total factor productivity of enterprises.

In summary, internal controls can avoid innovation risks, alleviate the degree of internal information asymmetry and improve the driving force of innovation and the possibility of innovation success so as to achieve the goal of improving the overall business performance of enterprises. Meanwhile, through reducing financing costs and improving the quality of accounting information and working capital management, enterprises can enjoy a good investment environment to improve the total factor productivity of enterprises.

Based on the above analysis, the following hypothesis is proposed.

H1. 

Internal controls can effectively improve the total factor productivity of an enterprise.

H2. 

Internal control can improve the total factor productivity of enterprises by promoting enterprise innovation and easing financing constraints.

2.2. Moderating Effect of Environmental Uncertainty

Enterprises always operate in a specific external environment, and the changes in the economic, international and political environment constitute environmental uncertainty. Environmental uncertainty is defined as the inability of management to accurately appraise the environment in which the firm operates, as well as its risk factors [16]. Specifically, it refers to the complexity of the environmental dynamics faced by the firm due to a lack of information. Environmental uncertainty leads to unpredictability in the future business risks and profitability of the firm [17], which will have an impact on the operational efficiency of enterprises.

In the case of high environmental uncertainty, external investors are less able to predict and monitor firm performance [18], and there are more information asymmetries between management and shareholders, which is detrimental to the functioning of the firm. As mentioned above, the internal information asymmetry of enterprises will reduce the willingness of enterprises to innovate, so the intensification of environmental uncertainty will inhibit the innovation of enterprises [19] and then negatively affect the development of enterprises’ total factor production. Meanwhile, environmental uncertainty can significantly increase the difficulty and cost of assessing risks, reduce the effectiveness of firms’ preventive control strategies and increase the risk premiums demanded by external investors, all of which impair firm value [20]. As mentioned above, the increase in investment cost will aggravate the problem of corporate financing constraints, so environmental uncertainty will inhibit the total factor productivity of enterprises by improving the degree of financing constraints.

Based on the above viewpoints, the following hypothesis is proposed.

H3. 

Environmental uncertainty inhibits the degree to which internal controls can contribute to a firm’s total factor productivity.

3. Study Design

3.1. Data Sources and Descriptive Statistics

Considering the importance and standardization of internal control theory in China in recent years and the remarkable results of high-quality economic development in China, the actual data of Chinese listed companies are representative. Therefore, this paper is targeted to select Chinese A-share listed companies from 2009 to 2019 as the research sample, and data of the years that correspond to the financial crisis and COVID-19 pandemic are excluded to reduce the impact of extreme values. The data related to internal control are mainly from the DIB database and supplemented by the CSMAR database. The data are further processed: (1) excluding ST and *ST listed companies; (2) excluding financial and insurance listed companies; (3) removing some listed companies with serious data omission. A total of 22,058 sample observations were obtained after processing, and ±1% tailing was applied to all continuous variables.

3.2. Variable Selection and Description

3.2.1. Explained Variable: Total Factor Productivity

Referring to related studies [21], this paper adopts the LP method to calculate enterprise total factor productivity and uses the results obtained from the OP method as a basis for robustness testing to alleviate the endogeneity problem and selectivity bias. The LP method is a combination of the Cobb-Douglas (C-D) production function and semiparametric estimation to estimate the values of variables by establishing a simple linear regression of the production function, but in order to correct the simultaneity bias and sample selectivity bias problem, the LP method introduces intermediate inputs as proxy variables to reduce the loss of sample size, and it can more accurately represent the production surplus from inputs other than the capital and labor factors.

3.2.2. Explanatory Variables: Internal Control

This paper selects the “DIB Internal Control Index” from the influential DIB Internal Control and Risk Management Database as a reference indicator. The Internal Control Index is designed by DIB based on the degree of achievement of internal control objectives and consideration of the five elements of internal control. The internal control index score is designed to measure the efficiency and effectiveness of corporate internal control practices. In this study, it can describe the quality of internal control of enterprises scientifically. It is also standardized considering the appropriateness of the regression coefficients as a specific indicator of the final measure of internal control quality, and the larger the index is, the higher the quality of internal control of the enterprise.

3.2.3. Moderating Variable: Environmental Uncertainty

Environmental uncertainty represents the unpredictability and complexity of the external environment, and changes in the external comprehensive environment will have an impact on the firm’s operations. Therefore, this paper refers to the calculation method of Shen (2012) [18] and adopts the least squares method (OLS) to construct model (1), which uses the firm’s performance indicators to measure environmental uncertainty; specifically, the standard deviation of the company’s abnormal sales revenue for the last five years is forecast.

Sale = φ₀ + φ₁Year + ε

(1)

In this model, Sale is sales revenue, Year is the year and ε is abnormal sales revenue. The model is calculated by dividing the standard deviation of abnormal sales revenue for the last five years by the corresponding average for the same period, reflecting the unadjusted environmental uncertainty of the last five years. In order to more accurately measure the indicator, the model excludes sales revenue due to stable company growth and finally divides each company’s unadjusted environmental uncertainty by the industry environmental uncertainty (the median of unadjusted environmental uncertainty of all companies in the same industry in the same year) to reflect the industry-adjusted environmental uncertainty. A larger value indicates greater environmental uncertainty.

3.2.4. Main Control Variables

In order to explore the causal relationships between the main variables more accurately, this article refers to related studies, and the asset–liability ratio (Det), firm age (Age), proportion of independent directors (Lnd_r), firm size (Size), growth rate (Growth), board size (Board) and whether the CEO is also the chairman (Duality) were selected as control variables among the firm factors that may affect the total factor productivity of the firm. All are shown in

Table 1. Definition of variables.

Variable Type	Variable Name	Variable Symbol	Definition
Explained variables	Total Factor Productivity	TFP_LP	Measurement used by the LP method
Explanatory variables	Internal Control	IC	DIB Internal Control Index/100
Intermediate variables	Environmental Uncertainty	EU	Environmental uncertainty without industry restructuring/industry environmental Uncertainty as a whole
Control variables	Asset-liability ratio	Det	Total liabilities/total assets
	Corporate Life	Age	Years of establishment
	Percentage of independent directors	Lnd_r	Number of independent directors/Total board members
	Enterprise size	Size	Ln (Total assets + 1)
	Business Growth	Growth	(Operating income—L. Operating income)/L. Operating income
	Board Size	Board	Ln (number of board members + 1)
	Dual Roles	Duality	1 if the chairman is also the CEO; 0 otherwise
	Year	Year	Year dummy variable
	Industry	Ind	Industry dummy variables

.

3.3. Model Construction

To explore the relationship between total factor productivity and internal control (H1), the following model (2) is constructed in this paper.

TFP_LP = α₀ + α₁IC + α₂Det + α₃Age + α₄Lnd_r + α₅Size + α₆Growth + α₇Board + ∑Year + ∑Ind + ε

(2)

TFP_LP is the explanatory variable, representing total factor productivity and IC is the explanatory variable, representing internal controls. α₁ -α₇ denotes the coefficient of each control variable, Year denotes the year, Ind is the industry fixed effect and ε is the estimated residual. If the coefficient α₁ in Model (2) is positive, it reflects that internal control promotes total factor productivity of firms.

To further study the role that the environmental uncertainty plays to moderate the internal control and total factor productivity of firms (H2), model (3) was created based on model (2).

TFP_LP = β₀ + β₁ EU + β₂EU × IC + β₃IC + β₄Det + β₅Age + β₆Lnd_r + β₇Size + β₈Growth + β₉Board + ∑Year + ∑Ind + ε

(3)

β₂, the coefficient of between internal control and environmental uncertainty, is mainly considered in this model. If it is negative, it means that environmental uncertainty inhibits the interaction between internal control and total factor productivity of the firm.

4. Analysis of the Empirical Results

4.1. Descriptive Statistics

The results of descriptive statistics of the variables are shown in

Table 2. Descriptive statistics of variables.

Variables	Sample Size	Average	Standard Deviation	Median	Minimum Value	Maximum Value
IC	22,058	6.438	1.409	6.728	0	8.86
TFP_LP	22,058	6.909	2.143	6.248	3.751	10.43
Size	22,058	22.117	1.296	21.948	19.538	26.07
Det	22,058	0.438	0.212	0.431	0.052	0.94
Age	22,058	17.292	5.195	17	7	32
Growth	22,058	0.206	0.501	0.119	−0.57	3.48
Board	22,058	2.255	0.178	2.303	1.792	2.77
Lnd_r	22,058	0.374	0.053	0.333	0.333	0.57
Duality	22,058	0.26	0.439	0	0	1

. The minimum value of TFP_LP is 3.751, the maximum value is 10.432, the mean value is 6.909 and the standard deviation is 2.143, indicating that the total factor productivity values of the sample companies differ greatly, and at the same time, it means that the residual efficiency of output other than capital and labor factor inputs in the sample companies is generally moderate; the minimum value of IC is 0, the maximum value is 8.866, the mean value is 6.438 and the standard deviation of 1.409, indicating that the extreme values of the internal control index of the sample companies differ greatly; the minimum value of enterprise size (Size) is 19.538, indicating that the sample companies are all large; the minimum value of gearing (Det) is 0.052, the maximum value is 0.94 and the mean value is 0.431, indicating that the values of gearing of the sample companies vary widely, but the overall leverage level of Chinese companies is moderate; the mean value of enterprise life (Age) is 17, indicating that the enterprise life of the sample companies is generally large; the minimum value of enterprise growth (Growth) is −0.57 and the maximum value is 3.48, indicating that the extreme values of growth of the sample companies vary greatly; the minimum value of board size (Board) is 1.792 and the maximum value is 2.77, with a mean value of 2.555, indicating that the board size of the sample companies is generally moderate; the minimum value of independent director ratio (Lnd_r) is 0.333 and the mean value is 0.374, which is in line with the requirement of the China Securities Regulatory Commission that the proportion of independent directors should not be less than one-third, indicating that the proportion of independent directors is not less than one-third and the corporate governance is relatively sound. The mean value of two-job integration (Duality) is 0.26, which indicates that the proportion of two-job integration in Chinese sample companies is low, at only 26%.

4.2. Univariate Test

The results of the univariate test are shown in

Table 3. Univariate test results.

Variables	Higher Internal Controls			Lower Internal Controls			Differences
	Sample Size	Average	Standard Deviation	Sample Size	Average	Standard Deviation	Average	T-Value
IC	11,030	7.537	2.074	11,028	6.281	2.024	−1.256 ***	−45.507

Note: *** indicates significance at the 1% level.

. The mean of total factor productivity in the sample group with higher-quality internal controls is 7.537, which is higher than those with lower-quality internal controls (6.281) and significant at the 1% level. This indicates that internal controls can increase firms’ total factor productivity when the effects of the other variables are controlled for and tentatively supports hypothesis H1.

4.3. Correlation Analysis

Total factor productivity is significantly and positively related to internal controls at the 1% level, which is consistent with hypothesis H1 (see

Table 4. Correlation matrix.

	IC	TFP_LP	Size	Det	Age	Growth	Board
IC	1
TFP_LP	0.302 ***	1
Size	0.176 ***	0.661 ***	1
Det	−0.112 ***	0.115 ***	0.463 ***	1
Age	−0.114 ***	0.101 ***	0.146 ***	0.144 ***	1
Growth	0.076 ***	0.146 ***	0.041 ***	0.044 ***	−0.008	1
Board	0.076 ***	0.176 ***	0.261 ***	0.146 ***	0.003	−0.022 ***	1
Lnd_r	−0.01	0.017 **	0.015 **	−0.005	−0.026 ***	0.005	−0.512 ***
Duality	−0.014 **	0.097 ***	−0.162 ***	−0.135 ***	−0.061 ***	0.031 ***	−0.178 ***

Note: ***, ** indicate significance at the 1%, 5% levels, respectively.

). Size (Size), enterprise growth (Growth) and board size (Board) are significantly positively correlated with total factor productivity at the 1% level, while asset–liability ratio (Det), firm age (Age) and dual management roles (Duality) show a significant negative correlation with total factor productivity. However, the correlation analysis only studies the numerical relationship between the main variables and ignores the influence of time, region and industry. It also only reflects the correlation between variables. Thus, a multiple regression analysis is needed to draw more accurate conclusions.

In this paper, we also use the variance inflation factor (VIF) to test whether there is multicollinearity in the model; the larger the value, the more serious the multicollinearity problem. If the maximum value does not exceed 10, it is empirically proven that the model does not have multicollinearity problem. The results in

Table 5. Results of correlation analysis.

	Model 2		Model 3
	VIF	1/VIF	VIF	1/VIF
Size	1.4	0.7142	1.40	0.7138
Det	1.29	0.7750	1.25	0.8026
Age	1.04	0.9649	1.03	0.9747
Growth	1.01	0.9945	1.38	0.7235
Board	1.53	0.6541	1.49	0.6725
Lnd_r	1.41	0.7101	1.37	0.7304
Duality	1.06	0.9461	1.04	0.9597
IC			1.17	0.8577
EU × IC			1.43	0.6725

show that the values of model (2) and model (3) are all less than 10, and the maximum values are 1.53 and 1.49 respectively, so model (2) and model (3) do not have multicollinearity.

4.4. Benchmark Regression Analysis

The results of the regression between internal controls and total factor productivity are shown in

Table 6. Base model regression results.

Explanatory Variables	Explained Variables TFP_LP
	(1)	(2)	(3)
IC	0.460 *** (40.449)	0.204 *** (21.33)	0.246 *** (17.74)
Size		1.218 *** (111.731)	1.197 *** (96.812)
Det		−2.406 *** (−38.192)	−2.462 *** (−33.556)
Age		0.019 *** (8.807)	0.026 *** (9.495)
Growth		0.471 *** (19.593)	0.570 *** (18.224)
Board		−0.188 *** (−2.590)	−0.182 ** (−2.214)
Lnd_r		−1.306 *** (−5.885)	−1.279 *** (−4.995)
Duality		0.015 (0.654)	0.052 * (1.797)
EU			0.111 *** (3.65)
EU × IC			−0.030 *** (−6.338)
_cons	3.948 *** (52.333)	−20.046 *** (−74.581)	−19.883 *** (−64.479)
Year	No	Yes	Yes
Ind	No	Yes	Yes
N	22,058	22,058	17,323

Note: ***, ** and * indicate significance at the 1% , 5% and 10% levels, respectively.

. Column (2) of Table 6 describes the regression results of model (2), in which the coefficient between internal control and total factor productivity of the firm is significantly positive at the 1% level, indicating that internal control can effectively improve total factor productivity of the firm with sufficient control variables as well as the year industry, fully validating H1; column (3) describes the regression results of model (3), which shows that the coefficient between internal control and environmental uncertainty is significantly negative at the 1% level, indicating that the greater the uncertainty of the environment in which the firm is located, the more significantly inhibited the positive effect of internal control on total factor productivity of the firm will be, validating H3.

Overall, the coefficients of the control variables firm size (Size) and firm age (Age) are positively significant at the 1% level, thus indicating that total factor productivity is positively influenced by firm size and age. The regression coefficient of the asset–liability ratio (Det) is significantly negative, thus indicating that a higher asset–liability ratio reduces innovation investment, which in turn affects enterprise performance. The coefficient of enterprise growth (Growth) is significantly positive at the 1% level, thus indicating that as the growth rate of operating revenue increases, the total factor productivity of enterprises will also increase.

4.5. Robustness Tests

To confirm the reliability of the results, the following tests were made in this paper.

4.5.1. Change the Measurement of the Total Factor Productivity

The measurement of core variables can affect the reliability of the study results. To control this impact, we adopt the Olley–Pakes method [22] to measure the value expressed by TFP_OP. The results are shown in column (1),

Table 7. Robustness tests.

Explanatory Variables	Explained Variables TFP_LP
	(1)	(2)	(3)
IC	0.251 *** (17.935)	0.172 *** (10.516)	0.202 *** (17.353)
Size	1.198 *** (96.555)	1.060 *** (24.062)	1.237 *** (82.793)
Det	−2.468 *** (−33.572)	−2.408 *** (−14.949)	−2.518 *** (−31.463)
Age	0.026 *** (9.634)	0.012 (0.215)	0.019 *** (6.971)
Growth	0.561 *** (18.128)	0.424 *** (13.869)	0.563 *** (16.165)
Board	−0.156 * (−1.905)	−0.007 (−0.037)	−0.241 ** (−2.543)
Lnd_r	−1.313 *** (−5.093)	0.082 (0.18)	−1.526 *** (−5.380)
Duality	0.057 ** (1.964)	0.004 (0.085)	0.033 (1.219)
EU	0.123 *** (4.052)	0.054 (1.591)	0.118 *** (2.908)
EU × IC	−0.032 *** (−6.649)	−0.012 ** (−2.418)	−0.034 *** (−5.289)
_cons	−19.980 *** (−64.897)	−17.027 *** (−12.975)	−19.547 *** (−54.124)
Year	Yes	Yes	Yes
Ind	Yes	Yes	Yes
N	17,323	17,323	10,554

Note: ***, ** and * indicate significance at the 1%, 5% and 10% levels, respectively.

.

4.5.2. Replacement of Sample Range

As the core industry of China’s economic development, the sample data are more representative, so this paper replaces the sample of Chinese listed manufacturing companies from 2009 to 2019 for the study. The results in column (2) of Table 7 find that the coefficient of internal control is still positively significant at the 1% level and that the coefficient of EU × IC is significantly negative at the 1% level, which is basically consistent with the previous regression results, testing the previous base regression results.

4.5.3. Endogeneity Test

The endogeneity problem mainly arises from measurement error, omitted variable error, reverse causality and selection error. In this paper, in order to mitigate the possible endogeneity problems, a fixed-effects model is used to mitigate the endogeneity problems caused by omitted variable errors, and a two-stage least squares (2SLS) method is used to mitigate the endogeneity problems caused by reverse causality.

Based on the basic regression model, we further control for individual fixed effects. The results in column (3) of Table 7 show that the coefficient of IC is significantly positive at the 1% level, the coefficient of the environmental uncertainty and internal control EU × IC (β₂) is significantly negative at the 5% level and the sign of the control variables does not change. It indicates that changing the model settings does not change the experimental results, which verifies the corresponding conclusions in the previous section.

The 2SLS model continues to be used for regression testing. In this paper, first-order lags L.IC are chosen as the instrumental variables. There is a significant correlation between the instrumental variables and the endogenous variables in the first stage regression, and, in addition, the results of the underidentification test and the weak identification test significantly reject the original hypothesis, indicating that the instrumental variables are selected effectively. The results of the second stage using instrumental variables to correct the bias of the endogenous variables are shown in

Table 8. Endogeneity test (2SLS).

Variables	First Stage	2SLS
	IC	TFP_LP
L.IC	0.456 ***
	(0.014)
IC		0.239 ***
		(0.022)
Det	−1.314 ***	−2.464 ***
	(0.075)	(0.079)
Size	0.221 ***	1.216 ***
	(0.012)	(0.014)
Growth	0.319 ***	0.451 ***
	(0.027)	(0.026)
Board	0.029	−0.228 ***
	(0.065)	(0.080)
Lnd_r	−0.100	−1.400 ***
	(0.204)	(0.246)
Age	−0.003	0.022 ***
	(0.002)	(0.002)
_cons	−0.943 ***
	(0.252)
Underidentification test		588.421 ***
Weak identification test		4271.963 ***
Observation	18,836	18,836
R-squared	0.300	0.481
Ind	YES	YES
Year	YES	YES
F	330.6	2217

Robust standard errors in parentheses, *** p < 0.01.

, and IC still has a significant positive correlation with TFP_LP, which is consistent with the previous findings.

5. Further Research

After exploring the moderating effect of environmental uncertainty on internal control and firms’ total factor productivity, this paper further explores the heterogeneous influences of this effect in different environments. It selects firm innovation and financing constraints, which are related to internal factors of the company as mediating variables to investigate in depth the direct mechanism of action between internal control and firm total factor productivity and to elaborate more clearly the mechanism of action between the main variables.

5.1. Heterogeneity Test

Enterprise Innovation Mechanism Test

The nature of enterprise ownership, the region in which the enterprise is located and the life cycle usually affect total factor productivity. The heterogeneity test in this paper analyzes the results from each of these three perspectives, and the results are shown in

Table 9. Heterogeneity analysis between the ownership structure and the region where the enterprise is located.

Explanatory Variables	Explained Variables TFP_LP
	(1)	(2)	(3)	(4)
	State-Owned	Non-State-Owned	East	Midwest
IC	0.150 *** (6.353)	0.209 *** (10.575)	0.193 *** (8.665)	0.155 *** (6.597)
Size	1.001 *** (13.068)	1.138 *** (22.261)	1.076 *** (19.713)	1.046 *** (14.467)
Det	−3.049 *** (−11.301)	−2.065 *** (−10.766)	−2.424 *** (−11.572)	−2.455 *** (−9.782)
Age	−0.003 (−0.043)	0.091 (1.494)	−0.001 (−0.015)	0.124 (0.786)
Growth	0.466 *** (10.742)	0.371 *** (9.063)	0.448 *** (10.754)	0.367 *** (8.54)
Board	−0.098 (−0.397)	0.065 (0.254)	0.158 (0.706)	−0.289 (−1.018)
Lnd_r	0.385 (0.65)	−0.333 (−0.509)	0.241 (0.404)	−0.149 (−0.212)
Duality	−0.054 (−0.683)	0.044 (0.738)	−0.011 (−0.205)	0.026 (0.304)
EU	0.025 (0.523)	0.101 *** (3.098)	0.100 * (1.91)	0.042 (1.023)
EU × IC	−0.011 (−1.548)	−0.017 *** (−3.395)	−0.022 *** (−2.740)	−0.006 (−0.956)
_cons	−14.818 *** (−7.324)	−19.680 *** (−11.885)	−17.634 *** (−11.093)	−17.345 *** (−6.675)
Year	Yes	Yes	Yes	Yes
Ind	Yes	Yes	Yes	Yes
N	8463	8777	10,955	6368

Note: *** and * indicate significance at the 1% and 10% levels, respectively.

and

Table 10. Heterogeneity analysis between different business life cycles.

Explanatory Variables	Explained Variables TFP_LP
	(1)	(2)	(3)
	Growth	Maturity	Decline
IC	0.259 *** (10.09)	0.203 *** (10.45)	0.105 *** (7.803)
Size	1.237 *** (37.856)	1.205 *** (46.561)	1.208 *** (45.204)
Det	−2.690 *** (−16.258)	−2.680 *** (−17.565)	−2.365 *** (−16.852)
Age	0.027 *** (3.696)	0.043 *** (7.793)	0.035 *** (6.843)
Growth	0.283 *** (7.133)	0.302 *** (6.718)	0.264 *** (6.591)
Board	−0.138 (−0.646)	−0.112 (−0.712)	0.066 (0.488)
Lnd_r	−0.2 (−0.333)	−1.415 *** (−3.053)	−0.245 (−0.613)
Duality	−0.021 (−0.392)	−0.056 (−1.112)	−0.012 (−0.312)
_cons	−20.943 *** (−25.478)	−19.869 *** (−30.700)	−20.743 *** (−35.504)
Year	Yes	Yes	Yes
Ind	Yes	Yes	Yes
N	5023	6935	9810

Note: *** indicate significance at the 1% levels.

.

To explore the impact of different ownership structures on the environmental uncertainty moderating effect of internal controls and total factor productivity, this paper categorizes firms into state-owned and non-state-owned firms. It is found that the coefficient of EU × IC is negatively significant at the 1% level for non-state-owned enterprises, thus indicating that environmental uncertainty has a significant inhibitory effect on the positive effect of internal controls on total factor productivity in non-state-owned enterprises. For state-owned enterprises, the moderating effect of environmental uncertainty is not significant, which may be due to the fact that non-state-owned enterprises do not receive financial support from the government. Furthermore, non-state-owned enterprises are also subject to “credit discrimination” by banks and other financial institutions [23], which leads to increased financing risks and financial constraints. Both factors are further exacerbated in the presence of uncertainty, which significantly affects the total factor productivity of enterprises.

To study the influence of different locations on the moderating effect of environmental uncertainty on the internal controls and total factor productivity of firms, this paper distinguishes between the eastern and central-western regions. The results show that environmental uncertainty significantly suppresses the positive effect of internal controls on total factor productivity in the eastern region, while the moderating effect is not significant for firms in the central and western regions. The reason for this finding may be that firms in the eastern region are more internationalized and locate in a concentrated industrial area [24], so macroeconomic uncertainty will affect their total factor productivity by affecting the degree to which they engage in international trade as well as the prices of production factors.

In this paper, we distinguish three groups of enterprises according to their life span, which we divide into the growth period, maturity period and decline period, and empirically analyze the impact of different enterprise life cycles on the total factor productivity of enterprises. The results show that improving the quality of internal controls can significantly improve the total factor productivity of enterprises in the growth stage group provided that the influence of other factors is effectively controlled for. The reason for this finding may be that for enterprises in the growth stage, the advantages of investment and financing are relatively obvious. Improving the quality of internal controls can effectively alleviate capital constraints, reduce corporate financing risks, improve corporate investment efficiency and thus increase total factor productivity.

5.2. Analysis of Mediating Mechanisms

In this paper, firm innovation and financing constraints are selected as mediating variables to study the mediating effect between internal control and firm total factor productivity.

5.2.1. Enterprise Innovation Mechanism Test

Enterprises promote innovation and improve their performance by improving the internal environment and its management, thus improving corporate total factor productivity. In this paper, we construct a model by referring to the mediating effect of Wen [25]. For the mediating variable of firm innovation, this paper adopts the Patent1 index (number of applications) as a specific measure of enterprise innovation by referring to previous studies. As shown in model (4), the larger the value of this indicator, the greater the enterprise’s innovation ability is.

Patent1 = Ln (utility model + design patent + invention patent + 1)

(4)

To explore the mediating relationship between enterprise innovation on internal control and total factor productivity of firms, the following models (5) and (6) are constructed in this paper.

Patent1 = α₀ + α₁IC + α₂Det + α₃Age + α₄Lnd_r + α₅Size + α₆Growth + α₇Board + ∑Year + ∑Ind + ε

(5)

TFP_LP = α₀ + α₁IC + α₂patent1+ α₃Det + α₄Age + α₅Lnd_r + α₆Size + α₇Growth + α₈Board + ∑Year + ∑Ind + ε

(6)

The results are shown in

Table 11. Tests for the mediating effect.

Explanatory variables	(1)	(2)	(3)	(4)
	patent1	TFP_LP	KZ	TFP_LP
IC	0.093 *** (12.609)	0.195 *** (19.827)	−0.120 *** (−14.358)	0.264 *** (22.838)
Size	0.543 *** (52.915)	1.198 *** (103.23)	−0.385 *** (−38.553)	1.167 *** (101.304)
Det	−0.363 *** (−6.752)	−2.397 *** (−37.744)	5.789 *** (98.139)	−1.320 *** (−17.006)
Age	−0.005 *** (−2.610)	0.018 *** (8.349)	0.014 *** (7.202)	0.019 *** (8.778)
Growth	−0.025 (−1.303)	0.476 *** (19.675)	−0.595 *** (−15.502)	0.330 *** (12.409)
Board	0.085 (1.273)	−0.203 *** (−2.793)	−0.075 (−1.261)	−0.235 *** (−3.276)
Lnd_r	0.134 (0.638)	−1.283 *** (−5.755)	1.090 *** (5.883)	−1.272 *** (−5.819)
Duality	0.104 *** (4.931)	0.011 (0.473)	−0.104 *** (−4.906)	−0.03 (−1.345)
patent1		0.048 *** (6.095)
KZ				−0.249 *** (−30.664)
_cons	−11.802 *** (−46.299)	−19.556 *** (−69.080)	7.213 *** (30.384)	−19.533 *** (−71.048)
Year	Yes	Yes	Yes	Yes
Ind	Yes	Yes	Yes	Yes
N	21,808	21,808	20,784	20,784

Note: *** indicate significance at the 1% levels.

. Column (1) describes the regression results of model (5), and the value of α₁ is 0.093 and is significant at the 1% level, indicating that the improvement of corporate internal control can effectively promote corporate innovation; column (2) describes the regression results of model (6), and the values of α₁ and α₂ are both significantly positive at the 1% level, verifying that internal control can improve the residual efficiency of corporate production by promoting corporate innovation. The result passes the test of mediating effect, validating H2.

5.2.2. Financing Constraint Mechanism Test

To verify whether firms reduce their financing risk by strengthening their internal controls to enhance total factor productivity, Models (7) and (8) are constructed by selecting financing constraints as mediating variables.

KZ = α₀ + α₁IC + α₂Det + α₃Age + α₄Lnd_r + α₅Size + α₆Growth + α₇Board + ∑Year + ∑Ind + ε

(7)

TFP_LP = α₀ + α₁IC + α₂KZ+ α₃Det + α₄Age + α₅Lnd_r + α₆Size + α₇Growth + α₈Board + ∑Year + ∑Ind + ε

(8)

In Model (7), the KZ index represents financing constraints, which are calculated using the method of Kaplan and Zingales (1997) and set as a mediating variable which is shown as model (9). OCF, Dividends and Cash represent net cash flow from operations, dividends and cash holding levels, respectively, and are normalized. Lev and Tobin’sQ represent the asset-liability ratio and Tobin’sQ, respectively. The higher the value, the higher the degree of financing constraints faced by the firm.

KZ = −1.001909 × OCF/Asset + 3.139193 × Lev − 39.3678 × Dividends/Asset − 1.314759 × Cash/Asset + 0.2826389 × Tobin’sQ

(9)

The results show that Column (3) in Table 11 describes the regression results of model (7), and the value of α₁ in column (3) is significantly negative at the 1% level, which indicates that the strengthening of internal controls can effectively reduce financing constraints. Column (4) describes the regression results of model (8), the value of α₁ is significantly positive and the value of α₂ is significantly negative, which verifies that internal control can provide financial support for business operations by alleviating the financing constraint of the business and ultimately increase the return from total output other than factor inputs. All variables pass the mediating effect test, validating H2.

6. Research Conclusions and Implications

6.1. Research Conclusions

How to effectively improve the total factor productivity of enterprises has been the focus of academic discussion in recent years. This paper takes China’s A-share listed companies from 2009 to 2019 as experimental samples and is based on the perspective of the uncertainty of the enterprise external environment to deeply analyze the influence of internal control on the total factor productivity of enterprises and the regulating mechanism and further test the mediating effect of enterprise innovation and financing constraints.

The empirical study on the relationship between enterprise internal control and TFP and its mechanism of action shows the following conclusions: (1) High-quality internal control is beneficial to the improvement of enterprise total factor productivity, and the influence is more significant in growing enterprises. (2) Environmental uncertainty plays a negative role in moderation, that is, environmental uncertainty significantly inhibits the promoting effect of internal control on total factor productivity. At the same time, the moderating effect of environmental uncertainty is different, which is more obvious in non-state-owned enterprises and enterprises in the eastern region. (3) Internal control can effectively stimulate enterprise innovation, alleviate financing constraints, and thus improve the total factor productivity of enterprises.

The research conclusions of this paper deepen the understanding of the influence of internal control on the total factor productivity of enterprises and enrich the theoretical achievements of the influencing factors of the total factor productivity of enterprises. The test from the angle of the moderating effects of environmental uncertainty also provides a new perspective for the study of internal control and total factor productivity of enterprises. In the practical operation level, this paper provides certain enlightenment for enterprises on how to optimize the internal control system, achieve high efficiency operation and enhance enterprise value.

6.2. Implications

Firstly, we must pay attention to internal control systems, establish a sound enterprise governance structure and improve the effectiveness of internal control operation and the efficiency of organization management to promote the growth of enterprise total factor productivity. We can adopt the principle of combining qualitative and quantitative methods, identify and locate internal control defects through the supervision mechanism in time and take targeted measures to solve them, so as to avoid causing major financial defects and thus affecting the company’s business performance.

Secondly, fully consider the negative impact of environmental uncertainties on enterprises, establish dynamic and effective adjustment and communication mechanisms, reserve appropriate financial flexibility and improve enterprises’ ability to withstand risks. Realize real-time supervision and regulation of enterprise investment activities, improve the rationality of capital use, avoid information asymmetry and blind or excessive investment behavior, and realize innovation investment and operation efficiency. We will effectively reduce financing costs, ease financing constraints for enterprises, and address various problems caused by environmental uncertainties in a timely manner, with particular emphasis on enterprises and non-state-owned enterprises in the eastern region.

Thirdly, we should strengthen innovation and research and development, encourage enterprises to make continuous innovation at multiple levels of technology and system and reduce substitutability within the market scope, so as to promote the growth of enterprise value. In the process of enhancing enterprise value, the principal–agent problem in the enterprise should be fully considered, and the conflicts between investors and managers should be balanced by means of executive compensation incentive mechanism and investor supervision and management mechanism, so as to achieve a good investment environment within the enterprise.

In addition, this study also has certain limitations: (1) Only some representative variables are selected as control variables for the study, which may ignore some factors that also affect the explained variables and fail to completely solve the endogeneity problem; (2) China’s A-share listed companies from 2009 to 2019 were selected for the study, and the sample data of other countries with different economic status were not taken into account, so the research results are limited.

Therefore, in the follow-up study, to make the model more reliable, control variables can be expanded, and the role of different variables between internal control and total factor productivity of enterprises can be further explored. In addition to the environmental uncertainty as a moderating variable studied in this paper, the effects of other moderating variables on the two can be further studied, or the mediating path between the two can be further explored. The sample range can also be expanded to improve the accuracy of the conclusions and supplement the differences.