Available via license: CC BY-NC 4.0
Content may be subject to copyright.
International Journal of Global Economics and Management
ISSN: 3005-9690 (Print), ISSN: 3005-8090 (Online) | Volume 6, Number 2, Year 2025
DOI: https://doi.org/10.62051/ijgem.v6n2.11
Journal homepage: https://ijgem.org
Content from this work may be used under the terms of CC BY-NC 4.0 licence (https://creativecommons.org/licenses/by-nc/4.0/).
Published by Warwick Evans Publishing.
WEP
Warwick
Evans
Publishing
Patent Intensity and Firm Performance: Evidence from Chinese
Listed Firms
Haicheng Jiang *
School of Economics and Finance, South China University of Technology, Guangzhou, China
*Corresponding Author: Haicheng Jiang
ABSTRACT
Using a large sample of approximately 2,500 listed firms in China over the past thirty years, I examine
the relationship between innovation output and firm performance, as measured by return on assets
(ROA). I document a significant increase in the number of patents granted to Chinese listed firms
domestically. My findings indicate that patent intensity—measured as the number of patents scaled
by total assets—is positively associated with future ROA, even after controlling for various patent-
related covariates, as well as year and industry fixed effects. These results contribute to the existing
literature on the positive impact of innovation on firm performance, which has been well-documented
in Western economies such as the United States and European countries. Furthermore, I identify
heterogeneities in this relationship across industries, patent types, and firm sizes. Specifically, the
effect of patent intensity on subsequent ROA is more pronounced among firms in the manufacturing
sector, those with a higher intensity of invention patents, and those with greater market equity. This
study provides new evidence on the real effects of innovation output on firm performance in the
context of China’s listed firms.
KEYWORDS
Innovation output; Patent intensity; ROA
1. INTRODUCTION
Technological innovation is a key driver of modern economic growth. Numerous studies have
highlighted innovation as the most critical engine of macroeconomic performance (Kogan,
Papanikolaou, Seru, & Stoffman, 2017; Farre-Mensa, Hegde, & Ljungqvist, 2020). In parallel, the
role of corporate innovation in financial markets has become an increasingly important topic,
garnering significant attention from financial economics researchers over the past two decades (He
& Tian, 2018).
The literature on innovation in financial economics primarily follows two main strands. The first
focuses on the determinants of corporate innovation at both the firm and market levels. Various
factors, such as entrepreneurship, ownership structure, corporate governance, institutional quality,
and market liquidity, have been identified as key drivers influencing firms’ incentives and ability to
finance innovation. The second strand examines the impact of corporate innovation on firm
performance. Empirical research suggests that innovative activities—measured by innovation input
(e.g., R&D expenditure), innovation output (e.g., the number of granted patents), or innovation
efficiency (e.g., the number of granted patents scaled by R&D expenditure)—serve as strong
predictors of future firm performance indicators, including return on assets (ROA) and stock returns
(Gu, 2005; Hsu, 2009; Chan, Chen, Hong, & Wang, 2015; Hou, Hsu, Watanabe, & Xu, 2016;
Hirshleifer, Hsu, & Li, 2013, 2018).
92
This paper examines the impact of innovation on firm performance. To be more specific, I utilize
patent data from more than 2,500 listed firms in China over the past three decades to examine the
relationship between innovation and subsequent ROA. The number of granted patents, scaled by total
assets, is used to measure the intensity of firms’ innovative output. Building on prior studies that
establish the predictive power of innovation on firm performance in the U.S. (Hsu, 2009; Hirshleifer
et al., 2013, 2018), I extend this analysis by investigating a large sample of Chinese listed firms. After
controlling for a wide range of firm characteristics, patent-related covariates, as well as year and
industry fixed effects, I find that patent intensity is positively associated with future ROA.
Furthermore, I explore the heterogeneous effects of innovation on firm performance. To capture
variations in patent intensity across industries, I disaggregate the full sample into manufacturing and
non-manufacturing firms at the first-digit industry classification level and further divide non-
manufacturing firms into second-digit industries, examining the impact of patent intensity on future
ROA within each sub-sample. I also analyze the effects of different types of patents by distinguishing
between invention patents, utility patents, and design patents, assessing their respective impacts on
future ROA separately. Additionally, I investigate firm size as a moderating factor in the relationship
between patent intensity and firm performance by examining this association separately for large and
small firms. Overall, my findings indicate that patent intensity has a stronger predictive power for
future ROA in manufacturing firms, invention patents, and larger firms.
This study closely relates to the literature examining the impact of innovation on various firm
performance indicators. For instance, Griffin, Hong, and Ryou (2018) find that corporate innovation
efficiency gradually enhances credit ratings, although this effect is influenced by firm-level
characteristics such as downside risk, financial constraints, and sales or cash flows. Wu and Chung
(2019) show that firms with greater innovation output and R&D investment are more likely to be
acquired, receive unsolicited bids, and attract multiple bidders, with this effect being stronger when
competition among bidders is higher, when acquiring firms operate in competitive product markets,
and when technological proximity in the acquiring firms’ industry is lower. Using instrumental
variable regressions and a difference-in-differences approach, Hsu, Lee, Liu, and Zhang (2015) find
a negative relationship between firms’ default probabilities and the size of their patent portfolios.
They further demonstrate that bonds issued by more innovative firms tend to have lower issuance
premiums and lower realized excess returns. Similarly, Ben-Nasr, Bouslimi, and Zhong (2019)
document a significant negative association between the number of patents and stock price crash risk,
with this effect being more pronounced in firms with weak corporate governance and high
information opacity.
The rest of the paper is organized as follows. Section 2 reviews the literature and corporate innovation
in China over the last three decades, Section 3 describes the data and variable definitions, Section 4
presents the empirical results about the association between patent intensity and ROA, and finally
Section 5 concludes.
2. LITERATURE REVIEW
In this section, I first discuss the impact of innovation on firm performance in the previous literature.
I then introduce the achievement of corporate innovation in China over the last three decades. Finally,
I summarize the previous empirical analysis and point out the research gap.
2.1. Innovation and Firm Operating Performance
The exploration of the relationship between innovation and firm performance originates from
Schumpeter’s innovation theory, which highlights the role of new combinations and creative
destruction in the survival and growth of firms, particularly in competitive industries (Schumpeter,
1934). Firms can benefit from innovation by enhancing their future performance through several key
93
mechanisms. First, engaging in innovative activities gradually strengthens firms’ absorptive capacity,
facilitating the accumulation of technical knowledge and experience (Hu, Jefferson, & Jinchang,
2005). A rich history of innovation fosters the development of high-quality, technologically advanced
products, increasing firms' competitiveness and profitability. Second, innovation enhances expected
profitability and productivity by reducing investment costs and enabling firms to realize potential
economies of scale in production (Griffith, Harrison, & Van Reenen, 2006; Adams & Jaffe, 1996).
Third, innovation allows firms to temporarily establish a monopoly position in the market, leading to
higher profits (Schumpeter, 2013). Technologically advanced products attract customers while
limiting direct competition, thereby granting firms a competitive advantage. However, as high profits
attract new market entrants, the profitability of innovative products tends to decline over time
(Lieberman & Montgomery, 1988). Although the high returns from innovation may be short-lived,
firms that continuously introduce new products can sustain a series of monopoly profits, thereby
improving long-term performance (Montgomery, 1995).
A substantial body of empirical research has investigated the relationship between innovation and
firm operating performance. Early studies primarily measure firms’ innovative activities using
innovation input proxies, such as R&D expenditure, since R&D investment marks the beginning of
the innovation process. Ehie and Olibe (2010) find a positive relationship between R&D investment
and firm performance in both the U.S. manufacturing and service industries, after accounting for firm
size and leverage. Their results further indicate that the positive impact of R&D investment was more
pronounced in the manufacturing sector before 9/11, whereas the service sector became more
sensitive to R&D investment in the post-9/11 period. Hou et al. (2016) provide evidence that R&D
intensity has a significant positive effect on future operating profitability, based on a large sample of
firms from 21 countries. Similarly, using a sample of leading Chinese electronics manufacturing firms,
Gui-long, Yi, Kai-hua, and Jiang (2017) argue that R&D-based innovation constitutes a critical long-
term investment for the growth and development of high-tech manufacturing firms.
However, R&D expenditure, as a measure of innovation activities, may not be the most appropriate
proxy for firms’ innovation intensity. Subsequent studies have explored the relationship between
innovation and firm performance from the perspective of innovation output. Compared to R&D
expenditure, patents and citations are preferred measures of corporate innovation because they
represent the tangible outcomes of innovation investment, effectively capturing the extent to which
firms utilize their R&D input (Hall, Jaffe, & Trajtenberg, 2005). Additionally, Acharya and
Subramanian (2009) argue that R&D expenditure is sensitive to accounting norms, as firms may
strategically choose to capitalize or expense R&D costs for financial reporting purposes. Empirical
evidence further supports the use of patents and citations as indicators of firm innovation. Gu (2005),
in a study of 1,330 U.S. firms, finds that changes in patents and citations are positively associated
with future earnings, particularly in industries such as computing and medical equipment, where the
time lag between technological progress and profit realization is relatively short. Moreover, the
strength of this relationship increases over time and may persist for up to five years. Pandit, Wasley,
and Zach (2011) extend this analysis by relaxing the assumption of constant marginal productivity of
R&D expenditure, demonstrating a positive association between patents and firms' future operating
performance. Additionally, they find that firms with a larger number of patents experience less
volatile future performance. More recent studies introduce the concept of innovation efficiency,
defined as innovation output relative to innovation input, as an alternative measure of firms’
innovative activities. For instance, Hirshleifer et al. (2013) and Mama (2018) provide evidence that
higher innovation efficiency positively influences firms’ future operating performance.
On the contrary, some studies report mixed empirical findings regarding the relationship between
innovation and firm performance. Using a sample of listed firms in the United States, Cui and Mak
(2002) find that while R&D intensity is positively correlated with Tobin’s Q, it is negatively
associated with return on assets (ROA). Similarly, Vithessonthi and Racela (2016) show that although
R&D intensity positively influences firm value, it is negatively correlated with operating performance
94
indicators such as ROA and return on sales (ROS). Their findings further indicate that, when firms
are categorized based on their level of R&D intensity, ROS is negatively affected by high R&D
intensity, whereas the effect is not statistically significant for firms with low R&D intensity. They
emphasize that substantial R&D expenditure may have an adverse impact on short-term performance,
even though it enhances competitiveness and contributes to long-term growth. Furthermore, Artz,
Norman, Hatfield, and Cardinal (2010), employing a longitudinal study, examine the relationship
between innovation output and firm performance. Their results suggest that patents negatively affect
ROA, which they attribute to the rise of strategic patenting practices rather than genuine innovation-
driven performance gains. In another study, Atalay, Anafarta, and Sarvan (2013) investigate the
impact of different types of innovation on firm performance using a sample of 113 firms from the
Turkish automotive supplier industry. Their findings indicate that product and process innovation
positively correlate with firm performance, whereas organizational and marketing innovation exhibit
no significant effects. Overall, the relationship between innovation and firm operating performance
remains inconclusive, suggesting that various factors, such as firm heterogeneity and the type of
innovation activity, may shape this association.
2.2. Innovation in China
China, as one of the largest emerging economies, holds a crucial position in the global business
landscape. While the country has experienced rapid economic growth for more than three decades
following the implementation of the reform and opening-up policy, this achievement has largely been
driven by factor mobilization and technological adoption. Historically, China has served as the
"world’s factory," relying on cheap labor to produce low-cost, low-end products for global markets
(Golder & Mitra, 2018). Between 1980 and 2000, China’s export-oriented economic model fueled its
rapid development, with the country's strong manufacturing sector playing a pivotal role in the global
industrial supply chain. However, as labor costs have risen and the technological gap between China
and advanced economies has narrowed, the sustainability of this growth model has come into question
(J. Y. Lin & Treichel, 2012). Recognizing the limitations of an economy dependent on low-cost
manufacturing and external technology, the Chinese government has increasingly emphasized the
importance of indigenous innovation as a driver of long-term economic transformation. To this end,
a series of policies have been introduced since the early 21st century to foster corporate innovation.
Among these, the most influential is the Outline of the National Program for Long- and Medium-
Term Scientific and Technological Development (2006–2020), promulgated in 2006. This policy
marked a significant shift by elevating independent innovation to a national strategic priority, making
China one of the few countries where the government explicitly mandates intellectual property targets.
The policies implemented by the Chinese government have yielded remarkable results. In recent years,
many Chinese firms have evolved from mere adopters and imitators of foreign knowledge to
innovative enterprises capable of challenging the dominance of Western multinational corporations.
From an innovation input perspective, China’s R&D expenditure surpassed that of Japan in 2011,
making it the world’s second-largest spender on R&D. By 2014, China had also overtaken continental
Europe in this regard (McKinsey Global Institute, 2015). From an innovation output perspective, the
country has witnessed a substantial increase in patent applications, particularly following its accession
to the WTO. As of 2018, China had ranked first globally in patent applications for eight consecutive
years. Furthermore, the McKinsey Global Institute (2015) highlights that China has made
comprehensive advancements in science and engineering across various industries. The country’s
well-established industrial production system, supported by a robust manufacturing sector, provides
a strong foundation for corporate innovation across different industries. Additionally, China has
achieved notable success in various forms of innovation, including cost innovation, manufacturing
innovation, and business model innovation, over the past three decades (Williamson & Yin, 2014).
95
2.3. Summary
As evidenced by the aforementioned empirical studies, the majority conclude that innovation is
positively correlated with firm performance, while a few suggest that the relationship remains
inconclusive. However, it is important to note that most of these studies focus on firms in the United
States and European countries, primarily due to the availability of firm-level innovation data.
Consequently, there is a significant gap in research on the impact of innovation on firm performance
in emerging economies such as China, which plays a crucial role in global economic performance
and innovation. Existing studies on China are largely conducted at the industry level or limited to
select firms. However, the continuous development of Chinese databases, particularly those related
to patents, now offers a unique opportunity to examine the relationship between innovation and firm
performance in the Chinese context. Therefore, this study aims to investigate whether the relationship
between innovation capabilities and future operating performance (measured by ROA) in Chinese
listed firms aligns with the patterns observed in Western companies.
3. DATA DESCRIPTION
In Section 3.1, I first outline the data sources and selection criteria for firms included in the empirical
analysis. I then describe the construction of the patent variable, operating performance indicators, and
other control variables. In Section 3.2, I present the descriptive statistics and the simple correlations
among all key variables.
3.1. The Data and Variables
The dataset is sourced from the Chinese Stock Market and Accounting Research (CSMAR) database,
which provides comprehensive firm-level and market-level financial data for all Chinese listed firms
since 1990. These firms are traded on the Main-Board Market (including A-shares and B-shares on
the Shenzhen and Shanghai Stock Exchanges) and the Second-Board Market (Growth Enterprises
Market Board). Over the past thirty years, the database has included more than 3,000 listed firms. To
ensure data reliability, I exclude firm-year observations with significant missing values in key
financial variables (ROA and total assets), invalid total assets or market equity (zero or negative),
instances where the earliest accounting year precedes the firm’s market entry, and firms in the
financial industry. Additionally, the dataset includes patent information from 1992 to 2017. After
merging patent data with operating performance indicators, the final sample consists of approximately
2,500 listed firms.
The CSMAR patent database records two key timestamps for each firm-year observation: the
application date and the grant date. Since patent applications may be rejected by the State Intellectual
Property Office (SIPO), the grant date is a more reliable measure of a firm's innovative activities. To
minimize potential look-ahead bias, I use the grant date as the effective date of each patent and count
the number of patents granted to firm i in year t. Additionally, the database classifies patents into
three types based on SIPO's classification: invention patents, utility patents, and design patents.
Empirical studies typically represent firm-level innovative activities through three main aspects:
innovation input, innovation output, and innovation efficiency. Earlier studies primarily relied on
innovation input (R&D expenditure) or innovation output (the number of patents) to examine the
relationship between innovation and firm performance. A more recent study by Hirshleifer et al.
(2013) introduced a different approach by focusing on innovation efficiency, which is measured as
the ratio of innovation output to input.
Ideally, I would use innovation efficiency—defined as the number of patents granted scaled by R&D
expenditure—to assess firms' innovative efficiency. However, data on R&D expenditure is only
available from 2007, and even within this period, significant missing values exist. Due to these data
96
limitations, I construct patent intensity as a measure of firms' innovative activities. Following Gu
(2005), patent intensity is computed as the number of patents granted scaled by total assets.
A more accurate measure of innovative activities should account not only for the number of patents
but also for patent citations, as suggested by previous studies (Hall et al., 2005; Hirshleifer et al.,
2013; Ben-Nasr et al., 2019). However, the CSMAR patent database does not provide firm-level
patent citation data. As a result, I rely solely on the number of patents granted as a proxy for
innovative activity. Thus, the key explanatory variable, patent intensity, is constructed as follows:
Where denotes the number of patents granted to firm i in year t and denotes total
assets of firm i in year t. The patent intensity variable covers the period between 1992 and 2017.
The operating performance of listed firms is measured using Return on Assets (ROA), which serves
as an indicator of how efficiently a company utilizes its assets to generate profits. ROA reflects a
firm’s profitability relative to its total assets, offering investors insight into how effectively the
company converts its investments into net earnings. A higher ROA indicates greater profitability, as
it signifies that the firm generates more income with fewer resources. In the CSMAR database, ROA
is calculated using the following formula:
Where denotes net profits of firm i in year t, and the denominator is the average of total
assets of firm i in year t and t − 1.
To control for potential confounding factors in the empirical analysis, several financial indicators are
included as control variables. These include capital expenditure, advertising expenditure, R&D
expenditure, and market equity. Additionally, firms in the CSMAR database are categorized into first-
level industries, such as manufacturing, real estate, business, public utilities, and general industries.
Each of these broad categories is further divided into more specific two-digit industries, with a total
of 70 two-digit industries represented in the dataset.
3.2. Descriptive Statistics
Figure 1. The trend of the number of patents in China (1992-2017)
97
Figure 1 illustrates the trend in the total number of patents and the number of patents per firm from
1992 to 2017. Over the past three decades, there has been a substantial increase in patent grants, with
a particularly sharp rise in the most recent ten years of the sample period. For instance, the total
number of patents granted to all listed firms surged from 12,478 in 2007 to 185,720 in 2017, while
the average number of patents per firm increased from 9.3 to 60.1 over the same period.
Table 1 presents the number of observations, pooled mean, standard deviation, 25th percentile,
median, and 75th percentile of patent intensity across first-digit and selected second-digit industries.
There is significant variation in patent intensity both at the first-digit and second-digit industry levels.
At the first-digit level, the manufacturing industry exhibits the highest patent intensity (0.068),
whereas the real estate and business industries have the lowest values (0.0102 and 0.0108,
respectively). Meanwhile, the public utility and general industries fall in the middle range, with patent
intensities of 0.0351 and 0.0325, respectively. The variation in patent intensity is even more
pronounced across second-digit industries. For example, agriculture and forestry record the lowest
patent intensities (0.0054 and 0.0059), while other manufacturing (0.1592), furniture manufacturing
(0.148), and electrical machinery and equipment manufacturing (0.1291) rank among the highest.
Table 1. Patent intensity across different industries
Industries
Obs
Mean
Std
25th
percentile
Median
75th
percentile
Panel A: first-digit industry
Public Utility
4947
0.0351
0.0979
0
0.0004
0.0270
Business
2007
0.0108
0.0585
0
0
0.0040
Manufacturing
23447
0.0680
0.1986
0
0.0164
0.0739
Real Estate
2413
0.0102
0.0335
0
0
0.0060
General
1114
0.0325
0.1732
0
0
0.0196
Panel B: second-digit
industry
General machinery
manufacturing
1122
0.0890
0.1351
0.0130
0.0464
0.1100
Electrical machinery and
equipment manufacturing
2043
0.1291
0.3081
0.0181
0.0694
0.1560
Furniture manufacturing
84
0.1480
0.2181
0.0084
0.0555
0.1606
Pharmaceutical
manufacturing
2295
0.0337
0.0646
0
0.0125
0.0397
Other manufacturing
144
0.1592
0.3359
0
0.0177
0.1311
Forestry
57
0.0059
0.0146
0
0
0
Health care industry
128
0.0153
0.0333
0
0
0.0176
Civil Engineering
Construction Industry
711
0.0195
0.3551
0
0.0066
0.0240
Agriculture
199
0.0054
0.0153
0
0
0.0046
Art industry
37
0.0201
0.4694
0
0
0.0036
Note: This table reports patent intensity across different industries. Patent intensity is defined as the
number of granted patents scaled by total assets. Panel A shows patent intensity across five industries
at the first-digit industry level, while Panel B shows patent intensity of some selected industries at
the second-digit industry level.
Table 2 presents the summary statistics of patent variables, ROA, and other control variables by
dividing the sample into two groups. Panel A includes firm-year observations with zero patents, while
Panel B consists of firm-year observations with at least one patent. The number of firm-year
observations is larger in Panel B than in Panel A across ROA and various patent variables, suggesting
98
that more firms belong to the group with at least one patent. Based on the patent measure in year t,
the average ROA in years t and t+1 is 0.069 and 0.071, respectively, for firms in Panel B. In contrast,
for firms in Panel A, the average ROA in years t and t+1 is 0.067 and 0.059, respectively. These
findings suggest that firms with at least one patent tend to outperform those without patents in terms
of ROA in China.
Table 2. Summary statistics
Obs
Mean
Std
Min
Max
Panel A: No patent granted in year t
One-year Forward ROA
12221
0.05915
0.06866
-0.2048
0.3454
Current ROA
12872
0.06703
0.07369
-0.2048
0.3507
∆ROA
11976
-0.01516
0.3733
-14.34
16.24
Patents
13415
0
0
0
0
Total Asset
13415
409.3
1236
0.134
41923
Patent Intensity
13415
0
0
0
0
ln(1+Patent Intensity)
13415
0
0
0
0
ln(1+CapExp/ME)
11726
3.027
1.507
0
8.944
ln(1+AD/ME)
1473
1.14
1.051
0
4.526
ln(1+RD/ME)
1497
14.82
1.772
0
18.42
Size
13318
4.647
9.628
0.01223
332.5
Panel B: At least one patent granted in year t
One-year Forward ROA
17770
0.07094
0.06659
-0.2047
0.3497
Current ROA
20168
0.06885
0.06539
-0.2047
0.3513
∆ROA
18606
-0.009898
0.3032
-20.87
27.38
Patents
20513
45.86
223.9
1
13394
Total Asset
20513
1301
7068
2.461
240538
Patent Intensity
20513
0.09018
0.2168
0.00001364
12.6
ln(1+Patent Intensity)
20513
0.07734
0.1181
0.00001364
2.61
ln(1+CapExp/ME)
20387
3.192
1.239
-0.003539
7.832
ln(1+AD/ME)
8696
0.9902
1.074
-2.228
5.822
ln(1+RD/ME)
13353
15.84
1.229
0
21.42
Size
20486
11.67
57.34
0.02759
5013
Note: This table shows descriptive statistics of all the variables. The unit of observation is firm-year.
The firms are sorted into two groups according to whether they have new patents granted in year t.
4. THE ASSOCIATION BETWEEN PATENT INTENSITY AND ROA
In Section 4.1, I first examine the relationship between patent intensity and future ROA using all
available Chinese listed firms. In Section 4.2, I explore the heterogeneous effects of patent intensity
across different industries, patent types, and firm sizes. Finally, I perform various robustness checks
to confirm the positive association between patent intensity and future ROA.
4.1. Special Signs
To systematically examine the relationship between patent intensity and subsequent operating
performance, I conduct pooled regressions using variations of the following econometric model:
99
Where is firm i’s one-year forward return on assets in year t+1, the preferred measure of
the future operating performance, is the natural log of one plus patent
intensity,
is the natural log of one plus capital expenditure to market equity,
is the natural log of one plus advertising expenditure to market equity, is the change in
ROA between year t and year t-1,
is the natural log of one plus R&D expenditure to
market equity. Considering the significant difference of ROA across different industries and the
potential unobservable industry-level time-invariant variables, the model accounts for industry fixed
effects. Similarly, to alleviate the bias from external economic and political shocks to ROA in specific
years, the model controls for year fixed effects as well.
Since the distribution of patent intensity is highly skewed and often zero, I use ln(1+PatentIntensity)
in the regression model. Similarly, because capital expenditure, advertising expenditure, and R&D
expenditure can sometimes be zero, I take the natural logarithm of one plus their value scaled by
market equity. Capital expenditure and advertising expenditure are included as standard control
variables when examining the relationship between innovation and firm performance (Pandit et al.,
2011; Hirshleifer et al., 2013). Following Gu (2005) and Hirshleifer et al. (2013), I include ROA in
year t to account for its persistence over time. Additionally, I control for the change in ROA between
year t and t − 1 to capture mean reversion in operating performance (e.g., Fama & French, 2000;
Hirshleifer et al., 2013). Given the emphasis in prior literature on the role of R&D expenditure in
explaining firm performance, I include the ratio of R&D expenditure to market equity to capture its
potential explanatory power. Finally, in line with common practice in financial studies, I winsorize
ROA at the 1% and 99% levels to mitigate the influence of outliers. Standard errors are clustered at
the second-digit industry level.
4.2. Baseline Results
Table 3 presents the baseline results from the variants of the econometric model specified in Equation
(4). To predict one-year forward ROA, Column (1) includes only the key explanatory variable—the
log of patent intensity—along with ROA in year t, the change in ROA between year t and t − 1,
industry fixed effects, and year fixed effects. This parsimonious model indicates a positive and
statistically significant association between patent intensity and one-year forward ROA. Column (2)
introduces capital expenditure as an additional control, while Column (3) further includes advertising
expenditure. The model specification in Column (2) is preferred due to data limitations related to
capital expenditure, advertising expenditure, and R&D expenditure. Column (3) closely follows the
model specification in Equation (3) of Hirshleifer et al. (2013), which examines the relationship
between innovation efficiency and firm performance (i.e., ROA and cash flows). Finally, Column (4)
incorporates R&D expenditure alongside the control variables used in the previous specifications.
Despite the presence of missing values in some control variables, the positive and significant
relationship between patent intensity and future ROA remains robust across the last three
specifications. In terms of economic significance, an increase of one standard deviation in the natural
logarithm of one plus patent intensity raises subsequent ROA by 0.19%, a magnitude comparable to
the coefficient estimates for patent efficiency reported in Hirshleifer et al. (2013).
100
Table 3. The association between patent intensity and the future ROA
(1)
(2)
(3)
(4)
∆ROA
-0.0144***
-0.0158**
-0.0081*
-0.1006***
(0.0048)
(0.0064)
(0.0043)
(0.0136)
Current ROA
0.5764***
0.5771***
0.6053***
0.6617***
(0.0199)
(0.0202)
(0.0199)
(0.0201)
ln(1+Patent Intensity)
0.0203***
0.0191***
0.0167***
0.0145***
(0.0032)
(0.0032)
(0.0033)
(0.0032)
ln(1+CapExp/ME)
-0.0017***
-0.0030***
-0.0029***
(0.0004)
(0.0008)
(0.0008)
ln(1+AD/ME)
0.0020***
0.0020***
(0.0005)
(0.0005)
ln(1+RD/ME)
0.0001
(0.0007)
Year dummy
Yes
Yes
Yes
Yes
Industry dummy
Yes
Yes
Yes
Yes
R-squared
0.3640
0.3662
0.4293
0.4733
Observations
26931
25882
7387
6265
Number of firms
2573
2573
1889
1700
Number of industries
70
70
67
63
Note: This table shows the relation between ROA and patent intensity after controlling some variables
from pooled cross-section regression. The dependent variable is the one-year forward ROA. ∆ROA
is the change in ROA between year t and year t+1. Current ROA is the return on assets in year t. Ln
(1+Patent Inten-sity) denotes the natural log of one plus patent intensity. Ln (1+CapExp/Me) is the
natural log of one plus capital expenditure scaled by year-end market equity. Ln (1+AD/ME) is the
natural log of one plus annual advertising expenditure divided by year-end market equity. Ln
(1+RD/ME) denotes the natural log of one plus R&D expenditure in fiscal year ending in year t
divided by market equity in year t. Industry dummies are based on the CSMAR classification. Year
dummies are included to alleviate the effect of economic and political shocks in some specific years.
The variables have been winsorized at the 1% and 99% levels. Standard errors are clustered at the
second-digit industry level and reported in parentheses. *, **, and *** denote 10%, 5%, and 1%
significance level, respectively.
The coefficients of the control variables are generally consistent with the existing literature.
Advertising expenditure is positively associated with one-year forward ROA, as expected. The
significantly positive coefficient for ROA and the significantly negative coefficient for the change in
ROA confirm both the persistence of ROA and the phenomenon of mean reversion. In contrast,
capital expenditure is significantly and negatively associated with one-year forward ROA, which
contrasts with findings from previous studies.
4.3. Heterogeneity and Robustness Checks
In this subsection, I examine the potential heterogeneous effects of innovation on future ROA across
different industry categories, patent types, and firm sizes. To maintain consistency with the baseline
results, the empirical analyses in this section use the model specification and control variables
outlined in Column (3) of Table 3.
Industry-specific factors significantly influence a firm’s innovative behavior, which in turn affects its
innovation performance (Guan & Pang, 2017). Firms within the same industry are likely to share
similarities in products, markets, technologies, and resource bases, leading them to adopt similar
101
strategies. In contrast, firms operating across different industries may exhibit distinct patterns of
innovation and operating performance. Innovation is expected to create variations in economic
performance both across industries and within firms (Breschi, Malerba, & Orsenigo, 2000; Malerba
& Orsenigo, 1997). As shown in Table 1, patent intensity varies widely across industries. For instance,
at the first-digit level, the patent intensity in manufacturing industries is more than twice that of other
industries. Further evidence in Figure 2 demonstrates that both the total number of patents and the
number of patents per firm are significantly higher in manufacturing firms compared to non-
manufacturing firms. Over the past thirty years, this gap has grown larger, indicating a dramatic
increase in patent grants to manufacturing firms. Manufacturing industries are typically capital-
intensive, and firms in these sectors are generally better equipped with funds and technology, making
them more likely to innovate (Guan & Pang, 2017). Additionally, ROA in listed firms tends to vary
systematically across industries. For example, capital-intensive industries (e.g., machinery
manufacturing) typically have different asset structures than labor-intensive industries (e.g.,
agriculture), leading to distinct patterns in ROA. Empirical studies examining the determinants of
ROA should account for these industry-specific differences.
Figure 2. The trend of the number of patents across industries in China (1992-2017)
Table 4 presents the empirical results by dividing the sample into observations from different
industries at the first-digit level. Columns (1) and (2) use the baseline model specification to examine
the association between patent intensity and ROA for manufacturing and non-manufacturing firms
separately. For the sample of 1,843 manufacturing firms across 36 second-digit industries, the
relationship between patent intensity and future ROA is positive and statistically significant. In
contrast, for the sample of 730 non-manufacturing firms across 34 second-digit industries, the
coefficient of patent intensity remains positive but is statistically insignificant. Moreover, when the
non-manufacturing industry is further divided into the public utility, business, real estate, and general
industries in Columns (3) to (6), the results are mixed. The coefficient for patent intensity in the public
utility and real estate industries is positive and statistically significant at least at the 5% level.
However, the coefficient in the business and general industries is negative, and the economic
significance is marginally small. These findings support the argument that the impact of patent
intensity on subsequent firm performance, as measured by one-year forward ROA, is particularly
strong in manufacturing industries. In contrast, in industries with lower levels of innovation activity,
such as public utility and real estate, the impact of patent intensity is either economically or
statistically insignificant. These results align with F.-J. Lin, Chen, and Lo (2014), who argue that
industrial characteristics are more influential than firm characteristics in determining the long-term
competitive advantages of firms in China.
102
Table 4. The heterogeneous association between patent intensity and the future ROA: different
types of industries
(1)
(2)
(3)
(4)
(5)
(6)
Manufac
turing
Non-
manufacturing
Public
Utility
Business
Real
Estate
General
ln(1+Patent
Intensity)
0.0196**
*
0.0145
0.0169**
-
0.0055**
0.1604**
*
-0.0214
(0.0028)
(0.0128)
(0.0079)
(0.0019)
(0.0241)
(0.0129)
Year dummy
Yes
Yes
Yes
Yes
Yes
Yes
Industry dummy
Yes
Yes
Yes
Yes
Yes
Yes
R-squared
0.3950
0.3007
0.3703
0.2486
0.2169
0.1914
Observations
17861
8021
3705
1594
1883
839
Number of firms
1843
730
401
113
136
80
Number of industries
36
34
18
5
4
7
Note: This table reports the relation between ROA and patent intensity across different types of
industries. To save the space, this table only includes the results of patent intensity. The dependent
variable is one-year forward ROA and ln (1+Patent Intensity) denotes the natural log of one plus
patent intensity. Industry dummies are based on the CSMAR classification. Year dummies are
included to alleviate the effect of economic and political shocks in some specific years. The variables
have been winsorized at the 1% and 99% levels. Standard errors reported in parentheses are clustered
at the second-digit industry level. * denotes 10% significance level, ** denotes 5% significance level,
and *** denotes 1% significance level, respectively.
Patents granted to listed firms at the subcategory level may have heterogeneous effects on the
relationship between patent intensity and future ROA. Within China's patent system, the State
Intellectual Property Office (SIPO) classifies all patents into three subcategories: invention patents,
utility patents, and design patents. As shown in Figure 3, utility patents consistently represent the
largest fraction of patents granted to listed firms. Another important trend is that invention patents
have become increasingly significant across all types of innovation activities. Since invention patents
play a crucial role in enhancing productivity, particularly in manufacturing firms (which constitute
over two-thirds of the listed firms in the sample), the relationship between patent intensity and future
ROA is likely to vary across the three types of patents. Therefore, I examine the impact of patent
intensity on one-year forward ROA separately for invention patents, utility patents, and design patents.
Figure 3. The trend of the number of patents across different types of patents in China (1992-2017)
103
Table 5 presents empirical results using invention patents, utility patents, and design patents to
measure patent intensity in the first three columns. The last column incorporates all three types of
patent intensity, considering the potential additive effect of patent intensity on one-year forward ROA
for listed firms. The results show a strong association between invention patents and one-year forward
ROA, while the patent-ROA relationship is much weaker, both economically and statistically, for
utility patents and design patents. This heterogeneity is likely driven by the fact that a significant
proportion of manufacturing firms, which tend to produce invention patents, are included in the
sample.
Table 5. The heterogeneous association between patent intensity and the future ROA: different
types of patents
(1)
(2)
(3)
(4)
invention patent
utility patent
design patent
All
ln(1+invention Patent Intensity)
0.0367***
0.0280*
(0.0135)
(0.0141)
ln(1+utility Patent Intensity)
0.0232***
0.0188***
(0.0057)
(0.0071)
ln(1+design Patent Intensity)
0.0137**
0.0063
(0.0068)
(0.0070)
Year dummy
Yes
Yes
Yes
Yes
Industry dummy
Yes
Yes
Yes
Yes
R-squared
0.3659
0.3660
0.3657
0.3662
Observations
25882
25882
25882
25882
Number of firms
2573
2573
2573
2573
Number of industries
70
70
70
70
Note: This table reports the relation between ROA and patent intensity across different types of
patents. To save the space, this table only includes the coefficient of patent intensity. The dependent
variable is one-year forward ROA and ln (1+Patent Intensity) denotes the natural log of one plus
patent intensity. Industry dummies are based on the CSMAR classification. Year dummies are
included to alleviate the effect of economic and political shocks in some specific years. The variables
have been winsorized at the 1% and 99% levels. Standard errors reported in parentheses are clustered
at the second-digit industry level. * denotes 10% significance level, ** denotes 5% significance level,
and ***denotes 1% significance level, respectively.
Firm size, as a key characteristic of a firm, has been used to examine its heterogeneous effect on the
relationship between innovation and firm value. Due to the advantages or disadvantages of economies
or diseconomies of scale, firm size can have an opposing effect on the link between innovation and
firm value (Connolly & Hirschey, 2005). Specifically, small firms are more likely to face
diseconomies of scale, which weakens the impact of innovation on firm value. In contrast, large firms
benefit from economies of scale, amplifying the effect of innovation on firm value. The theoretical
influence of firm size on the market-value effect of innovation is mixed. However, empirical studies
generally confirm the positive impact of firm size on the relationship between innovation and firm
value (Connolly & Hirschey, 2005; Pindado, De Queiroz, & De La Torre, 2010). While previous
studies have primarily focused on firms in developed countries such as the US and European nations,
whether the heterogeneous size effect exists in China remains an open question that warrants further
investigation.
Table 6 presents empirical results examining the relationship between patent intensity and ROA by
dividing the full sample into two subgroups: large firms and small firms. To determine firm size, I
use market equity and calculate the cutoff point, grouping firms with firm-year observations larger
than the overall median value of market equity as large firms (1,809 observations) and firms with
104
firm-year observations smaller than the overall median value as small firms (762 observations).
Column (1) illustrates the impact of patent intensity on subsequent ROA for large firms, while
Column (2) examines the same relationship for small firms. As expected, the impact of patent
intensity on one-year forward ROA is stronger in large firms and weaker in small firms. The final
column explores this heterogeneous effect by including firm size, measured by the natural log of
market equity, and its interaction with patent intensity. The results confirm the positive influence of
firm size on the relationship between patent intensity and firm value, revealing a statistically
significant size effect between large and small firms (the coefficient of the interaction term is
statistically significant at the 1% level). These findings are consistent with Hung and Wang (2012)
and Chun, Chung, and Bang (2015), who suggest that firm size has significant impacts on the
efficiency of innovative activities, based on studies of manufacturing firms in Taiwan and South
Korea.
Table 6. The heterogeneous association between patent intensity and the future ROA: size effect
(1)
(2)
(3)
> Median
< Median
interaction effect
ln(1+Patent Intensity)
0.0178***
0.0136**
0.0102**
(0.0031)
(0.0056)
(0.0043)
ln(size)
0.0044***
(0.0006)
ln(1+Patent Intensity)* ln(size)
0.0081***
(0.0024)
Year dummy
Yes
Yes
Yes
Industry dummy
Yes
Yes
Yes
R-squared
0.4962
0.2419
0.4210
Observations
15971
9141
25112
Number of firms
1809
762
2571
Number of industries
66
64
70
Note: This table shows the association between ROA and patent intensity across different sizes of
firms as well as the interaction effect of size and patent intensity. Large firms are defined as firms
that have market equity higher than the median, while small firms are firms that have market equity
lower than the median. Ln (1+Patent Intensity) denotes the natural log of one plus patent intensity in
year t. Ln (Size) is the natural log of market equity in year t. Ln (1+Patent Intensity)*ln (size) denotes
the interaction between patent intensity and firm size. Year dummies are included to alleviate the
effect of economic and political shocks in some specific years. Industry dummies are based on the
CSMAR classification. The variables are winsorized at the 1% and 99% levels. Standard errors
reported in parentheses are clustered at the second-digit industry level. *, **, and *** denote 10%,
5%, and 1% significance level, respectively.
5. CONCLUSION
In this paper, I investigate whether innovation activities, as measured by patent intensity, impact firm
performance. Using data from approximately 2,500 Chinese listed firms over the past three decades,
I find that firms with higher patent intensity exhibit better operating performance, as indicated by
ROA. The empirical results remain robust even after accounting for additional controls that could
influence innovation activities and through various robustness checks. Furthermore, I provide
evidence that the positive relationship between patent intensity and ROA is particularly pronounced
in manufacturing firms, for firms holding invention patents, and for larger firms.
105
The results support the conjecture that patent intensity has a significant positive impact on subsequent
firm performance, not only in developed countries but also in China, a developing country that has
experienced a surge in patent applications and grants in recent decades. An important policy
implication is that firms in developing countries could enhance their performance by fostering
innovation activities.
REFERENCES
[1] Acharya, V. V., & Subramanian, K. V. (2009). Bankruptcy codes and innovation. The Review of Financial Studies,
22(12), 4949–4988.
[2] Adams, J. D., & Jaffe, A. B. (1996). Bounding the effects of R&D: An investigation using matched establishment-
firm data (Tech. Rep.). National Bureau of Economic Research.
[3] Artz, K. W., Norman, P. M., Hatfield, D. E., & Cardinal, L. B. (2010). A longitudinal study of the impact of R&D,
patents, and product innovation on firm performance. Journal of Product Innovation Management, 27(5), 725–740.
[4] Atalay, M., Anafarta, N., & Sarvan, F. (2013). The relationship between innovation and firm performance: An
empirical evidence from Turkish automotive supplier industry. Procedia-social and Behavioral Sciences, 75(3), 226–
235.
[5] Ben-Nasr, H., Bouslimi, L., & Zhong, R. (2019). Do patented innovations reduce stock price crash risk? International
Review of Finance.
[6] Breschi, S., Malerba, F., & Orsenigo, L. (2000). Technological regimes and Schumpeterian patterns of innovation.
The Economic Journal, 110(463), 388–410.
[7] Chan, K., Chen, H.-K., Hong, L.-H., & Wang, Y. (2015). Stock market valuation of R&D expenditures—the role of
corporate governance. Pacific-Basin Finance Journal, 31, 78–93.
[8] Choi, S. B., & Williams, C. (2014). The impact of innovation intensity, scope, and spillovers on sales growth in
Chinese firms. Asia Pacific Journal of Management, 31(1), 25–46.
[9] Chun, D., Chung, Y., & Bang, S. (2015). Impact of firm size and industry type on R&D efficiency throughout
innovation and commercialization stages: evidence from Korean manufacturing firms. Technology Analysis &
Strategic Management, 27(8), 895–909.
[10] Connolly, R. A., & Hirschey, M. (2005). Firm size and the effect of R&D on Tobin’s q. R&D Management, 35(2),
217–223.
[11] Cui, H., & Mak, Y. (2002). The relationship between managerial ownership and firm performance in high R&D
firms. Journal of Corporate Finance, 8(4), 313–336.
[12] Ehie, I. C., & Olibe, K. (2010). The effect of R&D investment on firm value: An examination of us manufacturing
and service industries. International Journal of Production Economics, 128(1), 127–135.
[13] Ertugrul, M., Krishnan, K., & Xu, B. (2019). How does innovation impact product and financial market outcomes?
Available at SSRN 3443991.
[14] Fama, E. F., & French, K. R. (2000). Forecasting profitability and earnings. The Journal of Business, 73(2), 161–
175.
[15] Farre-Mensa, J., Hegde, D., & Ljungqvist, A. (2020). What is a patent worth? Evidence from the US patent “lottery”.
The Journal of Finance, 75(2), 639–682.
[16] Golder, P. N., & Mitra, D. (2018). Handbook of research on new product development. Edward Elgar Publishing.
[17] Griffin, P. A., Hong, H. A., & Ryou, J. W. (2018). Corporate innovative efficiency: Evidence of effects on credit
ratings. Journal of Corporate Finance, 51, 352–373.
[18] Griffith, R., Harrison, R., & Van Reenen, J. (2006). How special is the special relationship? Using the impact of US
R&D spillovers on UK firms as a test of technology sourcing. American Economic Review, 96(5), 1859–1875.
[19] Gu, F. (2005). Innovation, future earnings, and market efficiency. Journal of Accounting, Auditing & Finance, 20(4),
385–418.
[20] Guan, J., & Ma, N. (2003). Innovative capability and export performance of Chinese firms. Technovation, 23(9),
737–747.
[21] Guan, J., & Pang, L. (2017). Industry specific effects on innovation performance in China. China Economic Review,
44, 125–137.
[22] Gui-long, Z., Yi, Z., Kai-hua, C., & Jiang, Y. (2017). The impact of R&D intensity on firm performance in an
emerging market: Evidence from China’s electronics manufacturing firms. Asian Journal of Technology Innovation,
25(1), 41–60.
106
[23] Gunday, G., Ulusoy, G., Kilic, K., & Alpkan, L. (2011). Effects of innovation types on firm performance.
International Journal of Production Economics, 133(2), 662–676.
[24] Hall, B. H., Jaffe, A., & Trajtenberg, M. (2005). Market value and patent citations. RAND Journal of Economics,
16–38.
[25] He, J., & Tian, X. (2018). Finance and corporate innovation: A survey. Asia-Pacific Journal of Financial Studies,
47(2),165–212.
[26] Hirshleifer, D., Hsu, P.-H., & Li, D. (2013). Innovative efficiency and stock returns. Journal of Financial Economics,
107(3), 632–654.
[27] Hirshleifer, D., Hsu, P.-H., & Li, D. (2018). Innovative originality, profitability, and stock returns. The Review of
Financial Studies, 31(7), 2553–2605.
[28] Hou, K., Hsu, P.-H., Watanabe, A., & Xu, Y. (2016). Corporate R&D and stock returns: International evidence.
Recuperado de http://www. cicfconf. org/sites/default/files/paper 413. pdf.
[29] Hsu, P.-H. (2009). Technological innovations and aggregate risk premiums. Journal of Financial Economics, 94(2),
264–279.
[30] Hsu, P.-H., & Chaopeng, W. (2010). Innovations, intellectual property protection, and financial markets: Evidence
from China. Available at SSRN 1629331.
[31] Hsu, P.-H., Lee, H.-H., Liu, A. Z., & Zhang, Z. (2015). Corporate innovation, default risk, and bond pricing. Journal
of Corporate Finance, 35, 329–344.
[32] Hu, A. G., Jefferson, G. H., & Jinchang, Q. (2005). R&D and technology transfer: Firm-level evidence from Chinese
industry. Review of Economics and Statistics, 87(4), 780–786.
[33] Hung, S.-W., & Wang, A.-P. (2012). Entrepreneurs with glamour? DEA performance characterization of high-tech
and older-established industries. Economic Modelling, 29(4), 1146–1153.
[34] Koellinger, P. (2008). The relationship between technology, innovation, and firm performance—empirical evidence
from E-business in Europe. Research Policy, 37(8), 1317–1328.
[35] Kogan, L., Papanikolaou, D., Seru, A., & Stoffman, N. (2017). Technological innovation, resource allocation, and
growth. The Quarterly Journal of Economics, 132(2), 665–712.
[36] Lieberman, M. B., & Montgomery, D. B. (1988). First-mover advantages. Strategic management journal, 9(S1), 41–
58. Lin, F.-J., Chen, Y.-M., & Lo, F.-Y. (2014). The persistence of economic profit. International Entrepreneurship
and Management Journal, 10(4), 767–780.
[37] Lin, J. Y., & Treichel, V. (2012). Learning from China’s rise to escape the middle-income trap: a new structural
economics approach to Latin America. The World Bank.
[38] Malerba, F., & Orsenigo, L. (1997). Technological regimes and sectoral patterns of innovative activities. Industrial
and Corporate Change, 6(1), 83–118.
[39] Mama, H. B. (2018). Innovative efficiency and stock returns: Should we care about nonlinearity? Finance Research
Letters, 24, 81–89.
[40] Montgomery, C. A. (1995). Of diamonds and rust: a new look at resources. In Resource-based and evolutionary
theories of the firm: Towards a synthesis (pp. 251–268). Springer.
[41] Pandit, S., Wasley, C. E., & Zach, T. (2011). 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. Journal of Accounting,
Auditing & Finance, 26(1), 121–144.
[42] Pindado, J., De Queiroz, V., & De La Torre, C. (2010). How do firm characteristics influence the relationship
between R&D and firm value? Financial Management, 39(2), 757–782.
[43] Schumpeter, J. (1934). The theory of economic development. Harvard university press: Cambridge. MA.
[44] Schumpeter, J. A. (2013). Capitalism, socialism and democracy. Routledge.
[45] Vithessonthi, C., & Racela, O. C. (2016). Short-and long-run effects of internationalization and R&D intensity on
firm performance. Journal of Multinational Financial Management, 34, 28–45.
[46] Williamson, P. J., & Yin, E. (2014). Accelerated innovation: The new challenge from China. MIT Sloan
Management Review, 55(4), 27.
[47] Wu, S.-Y. J., & Chung, K. H. (2019). Corporate innovation, likelihood to be acquired, and takeover premiums.
Journal of Banking & Finance, 108, 105634.
[48] Yang, J. (2012). Innovation capability and corporate growth: An empirical investigation in China. Journal of
Engineering and Technology Management, 29(1), 34–46.
[49] Zhou, L. J., & Sadeghi, M. (2019). The impact of innovation on IPO short-term performance–Evidence from the
Chinese markets. Pacific-Basin Finance Journal, 53, 208–235.
