The Moderating Effect of Value Cognition and Market Competition:
A Study of the Relationship between R&D Intensity and Performance
of AI Enterprise
Litian Chen and Yufeng Wang
Zhejiang Gongshang University, School of Business Administration, Hangzhou, Zhejiang, China
Keywords: R&D Intensity of AI Enterprises, Enterprise Performance, Complexity of Value Cognition, Market
Competition.
Abstract: From the perspective of value cognition and innovation theory, using panel data of 75 listed companies in
China’s AI concept stocks from 2011 to 2019 as samples, this paper analyzes and examines the influence of
R&D intensity of AI enterprises on enterprise performance and the regulatory effect of value cognition and
market competition in this process. Studies have shown that the R&D intensity of AI enterprises is positively
correlated with their performance. In addition, the complexity of value cognition and the pressure of market
competition have a negative regulatory effect on the relationship between the R&D intensity and performance
of AI enterprises.
1 INTRODUCTION
In the field of AI, technology is updated rapidly and
R&D investment is high. Except for a few leading
enterprises with sufficient funds, it is difficult for
business managers and science and technology
decision-makers to balance the survival and
development of start-ups between market efficiency
and R&D investment (Xu 2021
). AI enterprises are
high-tech companies; R&D and innovation are the
keys to maintaining their market competitiveness.
However, this issue is rarely mentioned in existing
studies.
First, there are controversies surrounding the
relationship between R&D intensity and enterprise
performance. In some researches, it is believed that
the R&D investment intensity of an enterprise is
positively correlated with enterprise performance
(Mudambi 2014). In other researches, it is believed
that the two are negatively correlated or have a
nonlinear relationship (Racela 2016, KANG 2013).
However, in the existing researches, there are few
discussions about the relationship between the R&D
intensity and performance of AI enterprises.
Furthermore, value cognition and market
competition affect the relationship between R&D
intensity and performance of AI enterprises. This is
because the R&D and innovation process of
enterprises belongs to the value creation process, and
the decision-making of R&D is affected by the
cognitive level of business managers. It is generally
believed that the more complex the manager’s value
cognition, the more exploration opportunities the
enterprise has, the higher the enterprise’s exploration
capabilities (Daniella 2018, Stabell 1978), and the
greater the possibility that the enterprise will increase
R&D investment. However, when the value cognition
is very complex, enterprises should place resources in
multiple links of the innovation value chain, resulting
in resource fragmentation and resource waste.
Enterprise performance cannot be improved in a short
time. Finally, in most existing studies, it is believed
that under the high pressure of market competition,
leading enterprises will maintain their competitive
advantage by increasing R&D investment. However,
for AI enterprises in the early stages of development,
there is currently no research on whether R&D and
innovation can ensure the growth of enterprise
performance.
In order to make up the research gap, from the
perspective of value cognition and innovation theory,
with 75 listed companies in AI concept stocks from
2011 to 2019 as the objects of study, this paper
studies the relationship between the R&D intensity
and performance of AI enterprises and the
moderating effect of the complexity of value
762
Chen, L. and Wang, Y.
The Moderating Effect of Value Cognition and Market Competition: A Study of the Relationship between RD Intensity and Performance of AI Enterprise.
DOI: 10.5220/0011290300003440
In Proceedings of the International Conference on Big Data Economy and Digital Management (BDEDM 2022), pages 762-768
ISBN: 978-989-758-593-7
Copyright
c
2022 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
cognition and market competition on this relationship
to reveal the boundary conditions of the influence of
the R&D intensity of AI enterprises on enterprise
performance.
2 THEORETICAL BASIS AND
HYPOTHESES
2.1 R&D Intensity and Performance of
AI Enterprises
According to the innovation theory of Joseph Alois
Schumpete, science and technology are the driving
force for the development of enterprises, and the core
competitiveness of enterprises is formed through
R&D activities. The investment in innovation of an
enterprise is mainly reflected in the intensity of R&D
investment (SUN 2019). In the field of AI,
technology is updated rapidly and R&D investment is
high. Through R&D and innovation, AI enterprises
can obtain heterogeneous resources, promote the
launch of high-quality products, help enterprises to
seize a larger market share, and develop lasting
competitive advantages. In addition, according to
some scholars, by increasing innovation investment,
high-tech enterprises can improve their future
business performance (Pand 2011, Ciftci 2011). As
AI enterprises are high-tech enterprises, their R&D
expenditures are in line with the needs of the
enterprise’s long-term development strategies. The
successful transformation of R&D achievements will
ultimately lead to the economic growth of enterprises.
On this basis, Hypothesis 1 was proposed.
H1: The R&D intensity of AI enterprises has a
positive influence on enterprise performance.
2.2 The Influence of Value Cognition
on the Relationship between R&D
Intensity and Performance of AI
Enterprises
The complexity of value cognition refers to the
breadth of the knowledge covered by the managers’
knowledge structure (Walsh 1995, Nadkarni 2008).
From the perspective of value creation, it is the
number of links in the value chain that managers paid
attention to. If the complexity of value cognition is
high, it means that managers recognize various core
concepts (SHANG 2014) and ideas about the
technology field comprehensively in the process of
R&D investment, thereby increasing the possibility
of enterprises obtaining long-term value through
R&D investment. However, the ideas are so
comprehensive that they will use up more resources.
The more complex the value cognition in the R&D
investment process of AI enterprises, the more value
chains they pay attention to. This will lead to the
reallocation of resources distributed in multiple value
chains, which does not help with the concentration of
enterprise resources and may cause higher sunk costs
(CHEN 2021). Enterprise performance cannot be
improved. Therefore, the negative influence of the
increased R&D investment on performance in the
short term will not be reduced due to the higher
complexity of value cognition. On this basis,
Hypothesis 2 was proposed.
H2: The complexity of value cognition has a
negative moderating effect on the relationship
between R&D intensity and performance of AI
enterprises.
2.3 The Influence of the Pressure of
Market Competition on the
Relationship between R&D
Intensity and Performance of AI
Enterprises
The R&D intensity and performance of AI enterprises
are affected by both internal and external factors. The
R&D strategies of the enterprises are affected by the
intensity of market competition (JIAN 2017), and the
competition in China’s AI industry is intense. Once
the competitor has launched new products or offered
the same products at lower prices, management is
forced to make targeted R&D decisions (ZHENG
2018). With the increasingly intense market
competition, AI enterprises continue to increase R&D
investment in order to maintain their competitive
advantages. However, in the early stage of enterprise
R&D activities, the operating pressure of funds was
relatively high, and it was difficult to obtain returns
in a short time. At the same time, the risks of R&D
increase, and the possibility that the innovative
achievements of an enterprise are imitated or being
surpassed by follow-suitors increased (JIANG 2021).
On top of that, the AI industry in China is still at an
early stage of development. If enterprises in the
growth stage invest too much, their future
performance will be poor (Fedyk 2017).
On this basis,
Hypothesis 3 was proposed.
H3: The pressure of the market competition
pressure has a negative moderating effect on the
relationship between R&D intensity and performance
of AI enterprises.
The Moderating Effect of Value Cognition and Market Competition: A Study of the Relationship between RD Intensity and Performance of
AI Enterprise
763
3 METHODS
3.1 Samples and Data Sources
Sample selection: The research samples of this paper
were selected from Tonghuashun AI concept stocks
listed companies from 2011 to 2019. At the same
time, considering the lag effect of R&D intensity of
enterprises on enterprise performance, in this paper,
1 year was considered as the lag phase. The data from
2011 to 2018 were used as independent variables,
regulated variables, and control variables, whereas
the data from 2012 to 2019 were used as the
dependent variables. After excluding ST and *ST
companies, there were 75 listed companies in the
final samples. STATA15 software is used to process
the above-mentioned data.
Data source: Annual reports and social
responsibility reports of listed companies originated
from cninfo.com.cn, combined with the enterprises’
official websites for supplementary verification.
Most of the data on R&D intensity, enterprise
performance, and control variables are sourced from
CSMAR, where some R&D intensity research data
have been supplemented through the enterprises’
annual reports. By identifying and manually
collecting key sentences in the enterprise’s annual
report and social responsibility report, the complexity
of value cognition was obtained.
3.2 Variable Definition and Metric
1) Dependent variable: enterprise performance
(TBQ).
Based on the reference (HE 2021), in this paper,
we decided to measure the enterprise performance
with the profit index. In other words, we measure the
enterprise performance with the ratio of market value
to total assets.
2) Independent variable: R&D intensity (RD).
R&D intensity refers to the intensity of
enterprises investing limited resources in R&D. For
the measurement of R&D intensity, refer to the
practices of Barker & Mueller (Barker, Mueller 2002)
and Lv Diwei et al. (Diwei 2018). This paper
measures the R&D intensity of an enterprise with the
ratio of its R&D expenditure to total sales in year. The
greater the value of this variable, the higher the R&D
investment intensity of the enterprise.
3) Moderator variable: complexity of value
cognition (NC).
The coding research design of Nadkarni and
Narayanan (Nadkarni, Narayanan 2007) and Wu
Dong (Wu 2011) were used for reference and
integration. In this research, the text analysis method
was used to describe the complexity of cognition of
AI enterprises in the process of R&D and innovation.
There are mainly the following two steps.
Step 1: The statements were identified. The
annual reports and social responsibility reports of the
companies were studied, and according to the coding
vocabulary in the link of value creation summarized
by Wu Dong (Wu 2011) in his research, the sentences
showing that the enterprises have considered the
factors in each link of value creation in the innovation
strategy planning in each annual report were
scientifically identified and recorded.
Step 2: In the selected sentences, the number of
links (such as R&D, production, market, manpower,
and operation) in the chain of value creation
considered was determined and recorded as the value
of complexity of value cognition. The greater the
value of complexity, the more links of value creation
are considered in the cognition process of decision-
making of the enterprise.
4) Moderator variable: Market competition
pressure (HHI).
By using previous researches as a reference, this
paper measures the pressure of market competition of
the industry in which the enterprise is based with the
Herfindahl-Hirschman Index (HHI). The HHI is a
comprehensive index used to measure industrial
concentration. The smaller the index, the lower the
market concentration of the industry, and the more
intense the competition. It can be calculated as
follows.
HHI=
x
x
In which, x
represents the operating revenue of
the i enterprise, “x” represents the sum of the
operating revenues of all the enterprises in the
industry,
represents the market share of the “i”
enterprise, and “n” represents the total number of
enterprises in the industry.
5) Control variable
This paper selects the size of the enterprise (Size),
whether the chairman serves as the general manager
(CEO), industry (IND), asset-liability ratio (LEV),
the nature of enterprise ownership (SOE), and slack
resources (SLK) as control variables.
BDEDM 2022 - The International Conference on Big Data Economy and Digital Management
764
4 RESULTS
4.1 Narrative Statistics and Analysis
The descriptive statistics and correlation analysis of
the variables are shown in Table 1. The average
performance of AI enterprises is 2.593, which
indicates an average performance from the AI
enterprises in China, whereas the average R&D
intensity is 0.110. In other words, the R&D
investment accounts for more than 10% of operating
revenue. This shows that AI enterprises in China
attach importance to R&D investment. The average
complexity of value cognition is 4.410, which
indicates that many links of value creation are
evaluated by AI enterprises in China in the process of
R&D and innovation, and there are big differences.
The average value of HHI is 0.187, which indicates
that the pressure of the market competition of
enterprises in different industries is relatively high. It
can be seen from the table that R&D intensity is
significantly positively correlated with enterprise
performance, preliminarily supporting H1.
Table 1 Pearson correlation analysis and Descriptive statistics of all variables
Variabl
es
1 2 3 4 5 6 7 8 9 10 11
TBQ
1
RD
0.206**
*
1
NC
-
0.180**
*
-0.0290 1
HHI
0.0390 -
0.108**
*
-0.075* 1
Size
-
0.142**
*
-
0.127**
*
0.361**
*
0.0330 1
LEV
-
0.235**
*
-
0.219**
*
0.395**
*
-0.0410 0.609**
*
1
SOE
-0.0170 -0.0520 0.149**
*
0.115**
*
0.293**
*
0.083** 1
CEO
-0.0360 -0.0110 0.0470 0.00100 0.311**
*
0.229**
*
0.120**
*
1
SLK
-0.0110 0.281**
*
-0.00600 -0.0340 -
0.272**
*
-
0.416**
*
-0.0420 -
0.236**
*
1
IND
0.245**
*
0.224**
*
-
0.254**
*
-0.072* -
0.207**
*
-
0.325**
*
-
0.119**
*
0.077** -0.0530 1
Year
0.0180 0.00200 0.078** 0.0250 0.382**
*
0.205**
*
-0.00800 0.146**
*
-
0.164**
*
0.035
0 1
Mean
2.593 0.113 4.410 0.187 22.09 0.310 0.269 1.615 5.063 2.942 2,01
5
S.D
1.766 0.125 1.823 0.212 1.275 0.175 0.444 0.487 11.02 1.963 2.58
4
Note:*p<0.1，**p<0.05，***p<0.01
4.2 Hypothesis Testing
In order to verify the hypothesis proposed above, this
paper analyzed data with stata15.0. Since the original
hypothesis was rejected by the Hausman test, a fixed-
effect model was used.
1) Test on the main effect of R&D intensity of
AI enterprises and the enterprise performance
As shown in Table 2, Model (1) is a regression
model that only contains control variables, and in
Model (2) R&D intensity of the enterprise, an
independent variable is added. It can be seen from
Model (2) that the R&D intensity of an enterprise is
significantly positively correlated with enterprise
performance, supporting H1. In other words, the
higher the R&D intensity of an enterprise, the better
the enterprise performance.
2) Test of the moderating effect of the
complexity of value perception on the relationship
The Moderating Effect of Value Cognition and Market Competition: A Study of the Relationship between RD Intensity and Performance of
AI Enterprise
765
between the R&D intensity of AI companies and
the enterprise performance
As shown in Table 2，in Models (3), it can be
seen that the interaction terms of R&D of an
enterprise and complexity of value cognition are
significantly negatively correlated, which indicates
that the complexity of value cognition has a negative
moderating effect on the relationship between the
R&D intensity of an enterprise and enterprise
performance, supporting H2.
3) Test of the moderating effect of the pressure
of the market competition on the relationship
between the R&D intensity of AI companies and
the enterprise performance
In Models (4), it can be seen that the interaction
terms of R&D intensity of an enterprise and the
pressure of the market competition are significantly
negatively correlated, which indicates that the
pressure of market competition has a negative
moderating effect on the relationship between R&D
intensity of an enterprise and enterprise performance,
supporting H3.
Table 2: Stratified regression analysis results.
M(1) M(2) M(3) M(4)
Size -0.131 -0.202 -0.154 -0.233
(-0.83) (-1.15) (-0.86) (-1.38)
LEV -1.914*** 0.296** 0.289* 0.277*
(-5.83) (2.38) (2.14) (2.13)
SOE 0.182* -1.190*** -0.747 -1.032**
(2.02) (-3.54) (-1.73) (-2.46)
CEO 0.211 0.208** 0.240*** 0.193**
(1.82) (3.04) (3.73) (2.69)
SLK -0.009*** -0.012*** -0.013*** -0.018***
(-6.89) (-4.96) (-8.70) (-4.33)
Constant -382.989 5.003 4.326 5.555
(-1.56) (1.39) (1.15) (1.63)
RD 2.537* 3.535** 2.622*
(2.01) (3.35) (2.05)
NC
-0.130***
(-3.77)
RD_NC
-0.578***
(-4.29)
HHI
0.529**
(3.17)
RD_HHI
-8.921***
(-5.89)
Year Controlled Controlled Controlled Controlled
IND Controlled Controlled Controlled Controlled
Δ
R
0.136 0.361 0.379 0.380
F
15.51 16.74 15.92 15.62
Notes:* p＜0.1, **p＜0.05, ***p＜0.01.
4.3 Robustness Test
In order to improve the robustness of the conclusion,
this paper tested the robustness with the following
method. In previous related literature, the enterprise
growth rate was used as the control variable instead
of the asset-liability ratio (XIAO 2016). Therefore,
this paper replaced the asset-liability ratio (LEV) with
the enterprise growth rate (Growth) (the difference
between the main operating revenue of the current
period and the previous period/the main operating
revenue of the previous period). All test results are
consistent with the original results.
BDEDM 2022 - The International Conference on Big Data Economy and Digital Management
766
5 RESEARCH CONCLUSION
AND INSPIRATION
5.1 Research Conclusion
From the perspective of innovation theory and value
cognition, this paper empirically analyzes and tests
the influence of R&D intensity of an enterprise on
enterprise performance and the regulatory effect of
the complexity of value cognition and market
competition, and comes to three conclusions. The
R&D intensity of AI enterprises has a positive impact
on enterprise performance; the complexity of value
cognition has a negative regulatory effect on the
relationship between the R&D intensity and
performance of AI enterprises; the pressure of market
competition has a negative regulatory effect on the
relationship between R&D intensity and performance
of AI enterprises.
5.2 Theoretical Contribution and
Practical Significance
Firstly, in the field of AI, there are few studies on the
relationship between R&D intensity of an enterprise
and enterprise performance. This paper explored the
relationship between them through the innovation
theory and expanded the application of the innovation
theory. It provided some theoretical basis for the
improvement of enterprise performance by AI
enterprises through R&D and innovation.
Secondly, there is no research in which the
complexity of value cognition is considered as a
situational variable to study its influence on the R&D
intensity and performance of AI enterprises. To make
up this gap, in-depth research from the perspective of
value cognition was done. In the context of the high
complexity of value cognition, it is difficult for AI
enterprises to gather resources for R&D and
innovation, so the enterprise performance cannot be
improved in the short term. Therefore, when making
R&D decisions, enterprise managers should prevent
resources from being too fragmented and solve the
problem of resource waste caused by excessive
attention to the value chain.
Thirdly, AI enterprises are emerging technology
enterprises. Under the high pressure of market
competition, managers should not blindly increase
R&D costs because of the decisions of competitors,
which is a short-sighted behavior. If it is separated
from the actual situation of the enterprise, the
substantial increase in R&D costs will not only
increase sunk costs, but also will not help with the
improvement of enterprise performance.
5.3 Limitations and Future Research
The following limitations of this study can be used as
a reference for future research. First of all, since there
were few listed companies of AI concept stocks from
2011 to 2019, in the end, only 75 enterprises were
considered as samples, which is a relatively small
sample size. In the future, we can conduct research
based on more enterprise samples. Next, we only
measure the R&D intensity of AI enterprises with
financial resources, but other resources such as the
R&D personnel and technology may also affect
enterprise performance. The influence of different
types of resources on enterprise performance should
also be discussed in future researches. Last but not
least, in this study, the text analysis method was used
to measure the complexity of value cognition of
managers of AI enterprises. There will be subjective
factors of researchers.
ACKNOWLEDGMENT
This research was substantially supported by research
grants from the Zhejiang philosophy and social
science planning project (20NDJC099YB), National
Natural Science Foundation of China (71972170),
Zhejiang Provincial Natural Science Foundation of
China (LY18G020003).
REFERENCES
CHEN Litian, ZHANG Meimei. Strategy of Changes of
Enterprise Innovation Ability Under the Guidance of
the Efficiency of Value Creation [J]. Studies in Science
of Science, 2021, 39(5): 951-960.
Ciftci M, Cready WM. Scale Effects of R&D as Reflected
in Earnings and Returns [J]. Journal of Accounting and
Economics, 2011, 52(1): 62-80.
Daniella LM, Brusoni S. Cognitive Flexibility and
Adaptive Decision‐making: Evidence from a
Laboratory Study of Expert Decision Makers [J].
Strategic Management Journal, 2018, 39(4): 1031-
1057.
Haushalter D, Klasa S, William FM. The Influence of
Product Market Dynamics on a Firm’s Cash Holdings
and Hedging Behavior. Journal of Financial
Economics, 2007, 84(3): 797-825.
HE Yubing, ZHANG Si. Technological Diversification,
Internationalization, and Enterprise Performance [J].
The Moderating Effect of Value Cognition and Market Competition: A Study of the Relationship between RD Intensity and Performance of
AI Enterprise
767
Studies in Science of Science, 2021, 39(08): 1437-
1447.
JIAN Ze, TAN Liping, LV Daguo, FU Tong. Creativity,
Destructiveness, and Technology Upgrade of Market
Competition [J]. China Industrial Economics, 2017(5):
16-34.
JIANG Zhangsheng. The Impact of Open Innovation on
R&D Input of Manufacturing Enterprises-the
Regulating Effect of Government Subsidies and Market
Competition [J]. Science & Technology Progress and
Policy. 2021, 38(09), 100-108.
KANG Zhiyong. Financing Restrictions, Government
Support, and R&D Input of Local Enterprises in China
[J]. Nankai Business Review, 2013, 16(5): 61-70.
KONG Xu, HAO Feiyan, LIU Peipei, ZHANG Lianbiao.
R&D Input, Innovation Strategies, and Enterprise
Performance in the AI Industry: An Empirical Study
Based on 112 Listed Enterprises [J]. Science and
Technology Management Research, 2021, 41(08): 101-
107.
LV Diwei, LAN Hailin, CHEN Weihong. Research on the
Relationship Between the Inconsistency of
Performance Feedback and R&D Intensity [J]. Nankai
Business Review, 2018, 21(04): 50-61.
Mudambi R, Swift T. Knowing When to Leap:
Transitioning between Exploitative and Explorative
R&D [J]. Strategic Management Journal, 2014, 35(1):
126-145.
Nadkarni S, Barr PS. Environmental Context, Managerial
Cognition, and Strategic Action: An Integrated View
[J]. Strategic Management Journal, 2008, 29(13): 1395-
1427.
Nadkarni S, Narayanan VK. Strategic Schemas, Strategic
Flexibility, and Firm Performance: The Moderating
Role of Industry Clock Speed [J]. Strategic
Management Journal, 2007, 28(3): 243-270.
Pand ITS, Wasley CE, Zach T. The Effect of Research and
Development (R&D) Inputs and Outputs on the
Relation between the Uncertainty of Future Operating
Performance and R&D Expenditures [J]. Journal Of
Accounting, Auditing and Finance, 2011, 26(1): 121-
144.
Racela OC, Mathur I, Booth GG. Short- and Long-run
Effects of Internationalization and R&D Intensity on
Firm Performance [J]. Journal of Multinational
Financial Management, 2016(34): 28-45.
SHANG Hangbiao, LI Weining, HUANG Peilun. Research
on the Relationship between Management Cognition
and Strategic Change in Two Kinds of Environment [J].
Science and Technology Management Research, 2014,
11:167-175.
Stabell CB. Integrative Complexity of Information
Environment Perception and Information Use: An
Empirical Investigation [J]. Organizational Behavior &
Human Performance,1978, 22(1): 116-142.
SUN Ziyuan, WANG Ling, LI Xiuzhi, ZHAO Shaojuan.
Research on the Dynamic Relationship between R&D
Input and Enterprise Performance: The Regulating
Effect Based on the Effectiveness of Internal Control
[J]. Soft Science, 2019, 33(07): 51-57.
Tatiana Fedyk, Natalya Khimich. R&D Input Decisions of
IPO Firms and Long-term Future Performance [J].
Review of Accounting and Finance. 2017, 17(1): 50.
Vincent L. Barker; George C. Mueller. CEO
Characteristics and Firm R&D Spending [J].
Management Science, 2002, 48(6):
Walsh JP. Managerial and Organizational Cognition: Notes
from a Trip Down Memory Lane [J]. Organization
Science, 1995, 6(3): 280-321.
WU Dong. Research on Strategic Planning, Industrial
Transformation, and Entry Mode of Outward Foreign
Direct Investment [D]. Zhejiang University, 2011.
XIAO Hua, ZHANG Guoqing, LI Jianfa. Institutional
Pressure, Executive Characteristics, and Environmental
Information Disclosure of Company [J]. Business
Management Journal, 2016, 38(03): 168-180.
ZHENG Haiyuan, LI Xingjie. R&D Input, Market
Competition, and Enterprise Performance [J].
Communication of Finance and Accounting, 2018(18):
38-42.
BDEDM 2022 - The International Conference on Big Data Economy and Digital Management
768
