Technological Forecasting & Social Change 135 (2018) 132–144

Contents lists available at ScienceDirect

Technological Forecasting & Social Change

journal homepage: www.elsevier.com/locate/techfore

The time-varying impacts of government incentives on innovation T

a a,b,⁎
JingJing Zhang , Jiancheng Guan
a
School of Economics & Management, University of Chinese Academy of Sciences, 100190 Beijing, China
b
School of Business Administration, South China University of Technology, 510640 Guangzhou, China

A R T I C LE I N FO A B S T R A C T

Keywords: This paper advances extant theoretical research through focusing on the time-varying effects of government
Innovation fiscal incentives on the innovative performance of firms. Findings show that direct government subsidies favor
Subsidy firms in the short-term, but hinder them in their long-term innovation performance. Indirect tax credit, on the
Tax credit other hand, is favorable to a firm's short-term and long-term innovation performance. Importantly, combining
Resource slack
resource-based theory and social capital theory, we suggest that two dimensions of resource endowment —
financial and human slack — should be considered in evaluating the effectiveness of incentives. Specifically, our
analysis utilizes a unique panel of data from Chinese high-tech companies based in Beijing Zhongguancun
Science Park, allowing for our analysis of interior firm variation over time. The study shows solid empirical
support for the effects of government incentives over time and offers the interesting result that financial slack
positively moderates the relationships between government incentives and firms' innovation performance, while
human slack has the opposite moderating role. As such, our paper contributes to the present debate on gov-
ernment incentives in generating innovation, not only by investigating how effects of specific fiscal instruments
on innovation performance vary with time, but also by attempting to incorporate firms' resource endowment
conditions as contingent factors. Future research directions, implications for innovation research, and policy
implications are discussed.

1. Introduction
How do government financial incentives affect a firm's innovation
performance? Researchers have tried to answer this question, but their
studies yield mixed findings (Mahmood and Rufin 2005; Hall and
Lerner 2010; Carboni, 2017). Some scholars have claimed that having
more government financial support is often perceived to be better than
having less (Lerner 2000). However, other scholars suggested that
government financial incentives have several drawbacks, such as their
role in the substitution of a firm's own innovation expenditures (Zhang
and Wu, 2014).
Researchers have long struggled to reconcile these conflicting per-
spectives in at least two different ways. Firstly, the different results
from prior research suggest that the efficacy of government support for
private innovation activities may vary with respect to some contingent
factors. Specifically, studies have argued how industry conditions,
country institutional factors, and firm characteristics (e.g., tenure, size
and ownership and so on) may intervene in the relationships between
government financial incentives and innovation performance (Lach
2002; Mani 2002; Huergo et al., 2016). Secondly, scholars have illu-
minated that the innovation performance effect of government financial
incentives may depend upon the level of that incentive. Specifically,
government financial incentives are positively associated with innova-
tion performance until a certain threshold value, above which their
positive effect will gradually diminish and eventually turn negative,
expressed as an inverted U-shaped associations between government
financial incentives and innovation outcomes. Guan and Yam (2015)
claimed that direct subsidies failed to improve, and even negatively
affected, firms' innovative performances. Although quantitative re-
search has been investigating the role of government incentives in firms'
innovation performances, the findings are not universal, and this topic
is in urgent need of further work. Our study broadens these perspectives
by arguing that different types of government incentive tools play
distinct roles in short-term and long-term innovation outcomes.
Our study uses a different approach to reconcile present conflicts
and contributes to the ongoing work on the judgment of government
fiscal policy. Firstly, government policy regarding financial incentives is
considered to be a crucial element in firms' development. Governments
utilize a wide variety of incentive instruments to promote innovation,
such as tax credit, R&D subsidy, and special loans (Guan and Yam 2015;
Lee and Cin 2010; Borrás and Edquist, 2013). Evaluating the effects of
subsidies and tax credit is meaningful, because they are primary but

⁎
Corresponding author at: School of Economics & Management, University of Chinese Academy of Sciences, 100190 Beijing, China.
E-mail addresses: guanjianch@ucas.ac.cn, guanjianch@fudan.edu.cn (J. Guan).

https://doi.org/10.1016/j.techfore.2018.04.012
Received 23 January 2017; Received in revised form 8 March 2018; Accepted 8 April 2018
Available online 23 April 2018
0040-1625/ © 2018 Elsevier Inc. All rights reserved.
J. Zhang, J. Guan
distinct policy tools (Qiu and Tao 1998). Tax incentives are often
considered better than direct subsidies, because tax tools allow firms to
freely fund R&D projects, increase their own private R&D expenditure,
and sustain their long-term growth (OECD). In this study, we focus on
the distinct roles of subsidy and tax credit in influencing firms' in-
novation performance. Support from external resource holders (e.g.
governmental agencies) helps firms to complement a significant
amount of resources, such as knowledge-based resources, and con-
tribute critically to their innovative capacity. The heterogeneity of re-
sources and innovative capabilities would explain innovators' innova-
tion performance differences (Kazadi et al., 2016; Qian et al., 2017;
Guan and Yan, 2016). Grounded on resource-based and organizational
dynamics literature, which depict time as a crucial factor (Richard et al.
2007), we investigate how government fiscal incentives exert effects on
firms' short- and long-term innovation performance.
Secondly, although scholars have considered certain characteristics
of firms as critical contingency variables in the relations between
government incentives and performance, no thorough and systematic
research has studied the role of internal resources in the in-
centive—performance relationship. Resource-based views argue that
firms' resources enable them to generate innovation capabilities. As
such, we set out to integrate social capital and resource-based theory
literature, and thus posit that the impacts of subsidies and tax credit on
innovation performance can be truly understood by considering the
nature of the firms' available resources. Furthermore, resource-based
theory argues that innovators, like firms, are aggregations of hetero-
geneous and idiosyncratic bundles of resources and capabilities (Barney
1991; Wright et al. 2001; Guan et al., 2015a) and put forward the no-
tion of the firm's slack resources relative to other firms (Sharfman et al.
1988). A firm's needs are more important than the absolute levels of its
resources (Mishina et al. 2004). A series of previous studies analyzed
the role of financial slack, which is the level of financial assets ex-
ceeding those needed for basic organizational operating expenses in
innovation processes (Nohria and Gulati 1996). Recently, a small but
growing body of literature has begun to pay more attention to human
resource slack, which implies that the level of employees exceeds that
needed for organizational basic demand (Mishina et al. 2004). In our
paper, we focus on these two resource slacks, namely, financial and
human resource slacks, because both are essential for the growth and
development of firms (Vanacker et al. 2013).
Thirdly, we tested our hypotheses on high-tech firms in Beijing ZSP,
which is the pioneer of Chinese S&T Parks. The context of China is
interesting because its high-tech industries have boomed in recent
decades, and it is always a prior task for the government to foster do-
mestic innovative activities. As the world's largest developing economy
and country, China uses several popular innovation policy instruments
to support its developing firms. The Chinese government has put in-
novation and progress at the top of the agenda for firms in the Science
Park. Most OECD, and some emerging economies, have provided R&D
incentives on firms' R&D expenditure (OECD). However, most of this
incentive research has concentrated on advanced countries, such as the
Europe or the United States (Wallsten 2000; Czarnitzki et al. 2011).
Solid empirical research on the innovation consequences of Chinese
fiscal incentives in Science Park contexts is yet to be performed.
Table 1
Comparison of subsidy and tax credit.
Subsidy
Incentive measure Subsidy is a direct fiscal measure
Characteristic Subsidy is unneutral regarding the firm characteristic or
industry
Incentive time Beforehand subsidy
Incentive object Firms that participate in R&D projects selected by government
Project decisions Government chooses funding projects
Equality degree May distort equality
133
Technological Forecasting & Social Change 135 (2018) 132–144
To sum up, our study offers three important contributions to extant
literature. Firstly, our study examines the distinct roles played by
subsidies and tax credit in the short- and long-term of innovation per-
formance. Some studies have noticed the dynamic effects of govern-
ment incentives, but the majority of them focus on the time pattern of
the effects of tax credits or direct subsidies on R&D investment.
Recently, the effects of the government's financial incentives on firms'
innovation performances in the context of China have received a great
deal of attention (Guan and Yam, 2015; Jiao et al., 2015). To our
knowledge, no other study has attempted to test the time-varying im-
pacts of government incentives on innovation. Secondly, our study also
considers the roles of the resource slack in the incentive—performance
relationship. By exploring the influence of the internal resources of a
firm, along with its external policy, we extend the understanding of
government subsidies into a new realm. Thirdly, our study is among the
first to examine the effects of subsidies and tax credit simultaneously for
all Chinese high-tech companies in the Science Park. Fourthly, based on
a unique dataset of Beijing ZSP during 2005–2014, we conducted OLS
regressions with fixed effects to test the time-varying impacts of gov-
ernment incentives on innovation. To test the robustness of our find-
ings, we adopted several alternative approaches (such as GEE, GMM,
PSM and random effect models) and used multiple measurements of
innovation performance. The results remain consistent. Furthermore,
we tested the moderating roles of financial and human slack in the
above mechanisms. According to recent literature, in our robustness
tests, we measured the slacks using expectation models. We began this
study with a brief literature review of relevant research. In Section 2,
we will develop our theoretical model and propose our main hy-
potheses. After that, we will describe our empirical context, and present
the analyses and results. Finally, we will end with a conclusion and a
deep discussion of the findings, and then provided implications for in-
novation research and managerial practice.
2. Theory and Hypotheses
Direct subsidies and indirect tax credit are primary instruments of
government support for firms' R&D. Subsidies are direct expenditures
targeted to specific objectives chosen by the government, while tax
credit means the deduction in a firm's income tax system if they meet
certain requirements (Guellec and Van Pottelsberghe 2003). There are
some differences between subsidies and tax credit in both theoretical
and operational aspects. We summarize some key characteristics of
subsidies and tax credit in Table 1.
Slack represents the difference or gap between a firm's total re-
sources and their necessary payments (Cyert and March 1963). Fol-
lowing previous research, we also distinguish resource slack along two
dimensions. Financial slack indicates an excess of uncommitted fi-
nancial resources, while human slack consists of an excess of human
resources (Paeleman and Vanacker 2015). As such, we outline some key
characteristics of financial slack and human slack in Table 2.
2.1. Subsidy, tax credit and innovation performance
Government fiscal incentives, which refer to subsidies and tax credit
Tax credit
Tax credit is an indirect fiscal measure
Tax credit is irrespective of the firm nature, industry, or direction and aim of innovation
activity
Afterward credit
All firms that participate in R&D projects
Firms are free to choose their own projects
Fair and transparent
J. Zhang, J. Guan
Table 2
Comparison of Financial Slack and Human Slack.
Financial slack Human slack
Definition Excess uncommitted Excess of human resource
financial resource on the payroll
Slack type Unabsorbed slack Absorbed slack
Slack characteristic Uncommitted and readily Highly idiosyncratic and
available resources difficult to redeploy
in this paper, can benefit a firm's innovation performance for three
primary reasons. Firstly, both government subsidies and tax credit can
reduce the firm's costs of R&D activities, thus promoting its participa-
tion in R&D and potential innovation outputs (Fischer and Newell 2008;
Hall and Van Reenen 2000). Secondly, firms may not develop some R&
D projects only by themselves, even though these projects have great
innovative potential. Government incentives can help firms to cope
with these technological opportunities or projects that are character-
ized by high risk, uncertainty and high investment (Guellec and Van
Pottelsberghe 2003). Thirdly, both subsidies and tax credit may sti-
mulate a firm's additional R&D investments in innovation (Koga 2003;
Kobayashi 2014). Furthermore, government subsidies can positively
signal the quality and competitiveness of the recipient firm, thereby
encouraging investment and strengthening its collaboration in R&D
activities, which may trigger more innovation.
However, we argue that subsidies tend to only facilitate innovation
outcomes in the short-term, because their detrimental effects on in-
novation emerge in the long-term. The effectiveness of subsidies aimed
at promoting long-term innovation can be challenged for three reasons.
Firstly, government subsidies may partially or even fully crowd out a
firm's own R&D spending in innovation (Dai and Cheng 2015), and
reduce social welfare and its growth (David et al. 2000). Much research
has demonstrated that complementary and substitutive relationships
exist between government funding and private R&D spending (Guellec
and Van Pottelsberghe 2003). Firms usually need time to change their
strategy, and some major strategy changes may even take a long time.
Besides, there is a time lag between government subsidies and firms'
private R&D expenditures (Guellec and Van Pottelsberghe 2003; Levy
and Terleckyj 1983). Thus, the problems brought about the strategy of
decreasing private R&D spending may be more salient in the long-term.
Secondly, allocating direct subsidies or funding may not be as strict as a
firm's owned assets. Lax regulation of subsidies will make firms allocate
money into other dispensable activities, such as raising salaries, hiring
redundant employees, or purchasing unnecessary equipment (David
et al. 2000). In the long run, there will be a long series of cost run-over,
such as potential maintenance savings, and firms must spend further
resources to sustain the project. Thirdly, government subsidies may
distort the patterns and motivations of participation in a firm's R&D
activities, projects and misallocate resources to projects that are not
reliant on market forces (Guellec and Van Pottelsberghe 2003; Kung
et al. 2016). Thus, subsidies are the most costly and inefficient single
policy (Fischer and Newell 2008). Government-oriented R&D projects
not only aim to acquire economic return, but also take social benefits
into account. Thus, firms with government subsidies may not always
adopt a purely market-oriented approach, which is perceived to be one
of the most important antecedents of future organizational innovation
(Han et al. 1998). Some research has revealed similar findings, that
government subsidies have both advantages and disadvantages. For
example, an inverted U-shaped relationship exists between the share of
government R&D subsidies and a firm's innovative performance (Guan
and Yam, 2015).
However, these subsidy problems do not seem to be major concerns
for tax credit for several reasons. Firstly, some research argues that the
crowding-out effect may not be so easily produced, except that the real
R&D inputs do indeed increase (Hall and Van Reenen 2000; Czarnitzki
et al. 2011). Thus, tax credit may not easily crowd out private R&D
134
Technological Forecasting & Social Change 135 (2018) 132–144
investment in innovation like subsidies do. Secondly, compared with
subsidies, tax credit can be seen as one kind of a firm's owned assets,
and their regulation is stricter. Thirdly, a firm's administration of R&D
activities is not subject to arbitrary government interventions, with
respect to fund distribution and the objective of innovation activities
(Czarnitzki et al. 2011). Compared with direct government subsidies or
financing support, tax-based credit is more of a market-oriented tool,
since it leaves more choice about how to manage and develop R&D
projects in the hands of firms (Hall and Van Reenen 2000). Thus, we
propose:
Hypothesis 1a. Firms' subsidies favour short-term innovation
performance, while hindering long-term innovation performance.
Hypothesis 1b. Firms' tax credit favours their short-term and long-term
innovation performance.
2.2. Moderating roles of resource slack
Organizational slack implies the storage of excessive and freely
usable resource and enables organizations to cope with external pres-
sures and environmental changes (George et al. 2005; Yang et al.,
2014). Consequently, the resource configuration of firm influences the
progress of utilizing government incentives. As such, the implications of
government incentives for innovation performance should be con-
tingent on organizational resource endowment situations. In this paper,
we focus on financial and human resources slacks, and further theorize
their distinct moderating roles.
Prior research has demonstrated that organizational strategy and
growth depend on the financial slack available (Lungeanu et al., 2016).
For example, Bradley et al. (2011) suggested that financial slack re-
presents a double-edged sword, which has a negative effect on en-
trepreneurial management but a positive effect on a firm's growth. Yang
et al. (2014) examined the positive moderating roles of organizational
resource slack in reinforcing the changes of organizational routines and
beliefs to promote organizational innovation. In this paper, we argue
that, in the context of Chinese high-tech firms, a firm's financial slack
positively moderates the relationship between government incentives
and innovation performance. We will elaborate on this argument
below. Firstly, financial slack will provide the opportunity for firms to
take risks, facilitate proactive strategic initiatives, shield them from
disruptions and environmental turbulence, and allow them to experi-
ment with risky but promising projects (Moses 1992; Daniel et al.
2004). Thus, financial slack facilitates a firm's innovative and creative
behaviours, such as undertaking government-oriented R&D projects
that are characterized by as high investment and uncertainty. Secondly,
the bountiful financial assets of a firm may send a favourable signal to
potential external investors that firm's managers are prudent and have
sufficient financial resources to optimally invest. As such, this signal
can enhance their quality and competitiveness signals, which are pro-
duced by government supports. Dual signals intend to increase the
amount of external innovation investments and ultimately yield more
innovation. Thirdly, public incentives may lead to the improvement of a
firm's private investment, and these can occur more easily in good fi-
nancial conditions. Besides, financial slack makes the crowding-out
effect much less likely to happen, since firms have sufficient financial
resources to invest optimally (Mousa and Reed 2013). What's more, tax
credit can be applied only after the R&D investment is expended; thus,
insufficient financial funds will reduce the advantages and usefulness of
tax credit, thereby giving rise to the deficiency of a desired positive
effect (Radas et al. 2015).
We further posit that human slack negatively moderates the in-
centive-innovation relationship for the following reasons. Much re-
search has contributed to the insight that human resource slack is not
liquid, sticky, or hard to relocate (Lecuona and Reitzig 2014; Paeleman
and Vanacker 2015). Redundant labours, with their specialized skills
J. Zhang, J. Guan
and knowledge, will fit with firms' existing patterns of strategy (Mishina
et al. 2004). With government incentives, firms may seek new and
potentially risky innovation activities. However, existing human re-
sources may be perceived to be obsolete, ill-suited, or even detrimental
for innovation activities that need new skills and capabilities (Paeleman
and Vanacker 2015). Besides, such problems will be more salient in
high-tech industries in which environments are rapidly changing. For
another thing, holding excess employees will increase several costs,
such as payroll, administrative, and adjusting costs, and exert a burden
on the organizational cost structure (Lecuona and Reitzig 2014). We
argue that the government incentive diversion can easily happen, since
the major effect of incentives may be paying the wages of researchers or
adjusting the composition and size of the workforce to cope with the
innovation activities. Contrarily, human resource constraints will keep
firms vigilant and flexible, and avoid over-optimism and complacency
(Paeleman and Vanacker 2015). At the same time, when firms with
human resource slack want to develop new innovation projects after
they receive government incentives, they can easily pursue alternatives
to access skilled labour or knowledge on demand, like hiring employees
with the required skills, outsourcing, or technology procurement. Thus,
Hypothesis 2a. Financial slack positively moderates the relationships
between government incentives and firms' innovation performance.
Hypothesis 2b. Human slack negatively moderates the relationships
between government incentives and firms' innovation performance.
3. Methods
3.1. Setting and data collection
Our empirical analyses are based upon a unique, large, and annual
survey covering all high-tech companies located in Beijing ZSP during
the period 2005–2014, which was conducted jointly by the
Zhongguancun Management Committee and Beijing Statistics Bureau.
The survey is compulsory for all high-tech companies in the science
park, so the data has high accuracy and reliability. This survey provides
detailed, comprehensive, and longitudinal information. An extensive
data set of high-tech firm characteristics is included in this survey, in-
cluding information on their address, PRC business registration certi-
ficate code (9 numbers), established time, financial performance, in-
novation activities, operation statements, and so forth.
Furthermore, this dataset contains detailed innovation character-
istics of high-tech companies, for instance, R&D investment, R&D em-
ployees, and innovation performance. It also includes government in-
tervention information, such as taxes and subsidies. This dataset
provides a unique numerical ID for every firm using its code, name, and
address to link it over time and improve accuracy. Thus, we can track a
firm's innovation activity through year-to-year matches. Because of
possible missing values in this survey, we used both the firm's name and
business registration certificate code to search on the website of SAIC
(State Administration of Industry and Commerce of China) and com-
plement the information of these firms, such as established year.
Choosing high-tech firms in the ZSP as our sample to assess the
effects of government incentives on the firm's innovation outputs is
important for three reasons. Firstly, the ZSP holds an absolutely
dominant position in Chinese S&T parks, because it was the first to be
established and the largest park so far, now being hailed as the Chinese
“Silicon Valley”. The Chinese government has invested huge amounts of
money and energy to increase the innovation level and effectiveness of
high-tech companies. These high-tech firms also rely on government
fiscal policy to improve both their private and social benefits. Thus, it is
easy and common to observe government support. Secondly, the ZSP
represents a unique context in which to examine our research hy-
potheses regarding the geographical location of our samples. Thus,
choosing the ZSP as our research setting helped us to control for the
135
Technological Forecasting & Social Change 135 (2018) 132–144
confounding impacts of geographical location (Porter and Stern 2001).
Thirdly, the analysis focuses on high-tech firms because they have
particularly high desires and demands for innovation, unlike, for in-
stance, low-tech firms, which have low levels of expenditure on activ-
ities related to innovation. A further clear advantage of focusing on
high-tech firms in the ZSP, aside from guiding policy implication, is that
firms in the ZSP are relatively young, and thus appear to be less con-
strained by prior strategies or decisions that may affect the innovation
behaviours of firms, thereby mitigating concerns of endogeneity.
The aim of this paper is to look deeper into firms' funding processes,
both in the short-term period and long-term period, and more specifi-
cally how firms' internal resource slacks influence these processes, in-
cluding financial and human resource slacks. Thus, a longitudinal de-
sign was utilized to examine the time-varying impacts of government
incentives on innovation performance. We obtained explanatory vari-
ables from the detailed statements, which were reported by every firm
in the focal year (in this paper, year t). Because all firms in the ZSP are
required to record detailed information by law, such data allows us to
measure a firm's innovation performance in the several years sub-
sequent (in this paper, year t + 1, t + 2, t + 3 and t + 4). All variables
used in this study were logged to mitigate heteroscedasticity problems
and reduce the influence of extreme observations, except for dummy
variables (i.e., High-tech in this paper). The Chinese government has
launched several national programs to promote innovation, such as the
“Designation of National High and New Techno-logy Enterprises”. The
High-tech variable is set as a dummy variable; it takes the value of 1 if
this firm is identified as a national high-tech company, and 0 if not.
Finally, we obtained an unbalanced sample consisting of 165,531
firm-year observations, which constitutes 32,431 firms from various
industries across the period 2005–2014. Among those firms, 50.69%
are from the general service industry, 17.53% are from the manu-
facturing industry, 15.99% are from the key service industry, and
11.58% are from the wholesale, retail, hotel, and catering industries.
Other industries, such as construction and financial industries, occupy a
minor ratio. The average age of the firms was 7.06 years, with a range
of 1–115 years. Annual revenues ranged from RMB 0 to RMB 98 mil-
lion. The number of employees ranged from 0 to 28,423, with an
average of 75.26. These ranges suggest our sample constitutes a re-
presentative sample.
3.2. Measurement
3.2.1. Dependent variables: innovation
According to the Oslo Manual, there is currently no consensus re-
garding perfect innovation performance indicators. New products refer
to products for satisfying a new market demand or providing a novel
way of meeting an existing market demand. New products are con-
sidered as the potential commercial value of firms' R&D activities. As
previous research said, most innovations do not have value to firms
until the related new products are put into the market. Therefore,
creating new products is a critical component of a firm's innovative
performance. This study uses new product performance as the innova-
tion performance indicator. Further, measuring new product introduc-
tions can complement other intermediate indicators of a firm's in-
novation performance, such as patents, scientific papers, and R&D
investments. To test the robustness of our findings, we also measured
innovation performance using the number of patents in the robustness
check section. The findings are consistent. New products, considered as
strategic market positioning, may confer on a firm's increased sales,
bigger profits, and competitive strength (Fontana and Nesta 2009).
Thus, product innovation can be also deemed as the firm's quality
signal. The variable, Innovation, is constructed as the firm's new-product
sales revenue in the focal year t, which is one output of the firm's in-
novation process (log-transformed, unit used = one thousand RMB
Yuan).
J. Zhang, J. Guan
3.2.2. Independent variables
Subsidy is the money of government subsidies, which is received and
reported by the firm (unit used = one thousand RMB Yuan). We per-
formed a logarithmic transformation of subsidies to reduce the pro-
blems caused by the skewedness of its distribution. Because of the lack
of specific amount of government subsidies, most existing research
treated this variable as a dummy variable (González and Pazó 2008),
thus unable to capture the many aspects of government subsidies. In
this paper, Subsidy is a continuous variable, thus entailing less loss of
subsidy information than a binary variable would.
Another fiscal policy measure is Tax credit, which is the amount of
credit the focal firm subtracts from the nation. We computed the log of
Tax credit before it enters our regression models, and the unit of Tax
credit is one thousand in RMB Yuan. Under the Chinese tax credits
system, firms can receive tax reductions or even exemption from tax
when they undertake R&D projects via government programs. In China,
tax credit is a treatment, which is special and not obligated by the
common standards of the taxpayers' net income. Tax credit is actually
an incentive to encourage firms towards innovation activities, and the
range or amount of the decreasing tax may vary among firms and de-
pend on the degree of compliance between the firm's projects and the
national technology development directions. To promote innovation
and achieve a firm's technological development, Chinese governments
formulate some fiscal policies and rules to reduce a firm's taxes and
facilitate their innovation and technology development. For instance,
high-tech entrepreneurial firms in the ZGC zone could have 2-year tax-
exemption and enjoy an income tax rate of 15% from the third year (see
the website of The Ministry of Finance of the Government of the
People's Republic of China). The reverse causality may exist in the re-
lationship between a firm's investment choices and tax incentives. To
avoid the causality issue, following previous research, we lagged all
independent variables. According to previous research, this approach
can reduce concerns of reverse causality and avoided simultaneity.
3.2.3. Moderating variable
We tested the moderating roles of two dimensions of resource en-
dowment: financial and human slack. Financial slack shows the firm's
capital and its financial capability, which can be arranged or deployed
to issue low-risk debt or cultivate other internal capabilities. We mea-
sured financial slack using three steps. Step 1: we calculated the money
of equity relative to the debt of each firm in the year t, which is called
unadjusted financial slack. Step 2: we adjusted this measure for in-
dustry norms. The industry median values of ratios of equity-to-debt
were calculated based on all the companies in their specific industries
(according to Chinese Industry Classification Standards). Step 3: we
subtracted the median value of unadjusted financial slack for all com-
panies in the identical industry from the score of the focal firm, which is
our final financial slack measure (Daniel et al. 2004). We used the
following equation to calculate financial resource slack:
equity equity
Financial slackfirm i = −
debt firm i debt industry median (1)
Like financial slack, human slack is excessive human resources,
which enables firms to cope with the external environment. Following
early principles, we measured human slack using three steps. First, we
calculated the expenditure of employee salaries scaled by the sales of
each firm in the year t, which is called unadjusted human slack. Both
employee quantity and quality are considered through the inclusion of
employment salary, since labour of a higher skill or work experience,
which are perceived as more human capital, are always costly to em-
ploy (Vanacker et al. 2013). Second, we adjusted this measure for in-
dustry norms. The industry median values of the ratio of employee
salary-to-sales are calculated based on all the companies in their spe-
cific industries. Third, we subtracted the median value of unadjusted
financial slack for all companies in the same industry from the score of
the focal firm, which is our final human slack measure (Paeleman and
136
Technological Forecasting & Social Change 135 (2018) 132–144
Vanacker, 2015). We used the following equation to calculate human
resource slack:
employee salary
Human slackfirm i =
sales firm i
employee salary
−
sales industry median (2)
3.2.4. Control variable
When estimating the influences of subsidies and tax credit on in-
novation, we added several control variables associated with the focal
firm. We included several financial statements related control variables.
The Total assets variable means the sum of a firm's all current and
noncurrent assets. Research has long recognized the role of the firm's
size in explaining innovation performance. Larger firms might be more
competitive, resourceful, and able to yield more innovations (Luo and
Deng 2009). We controlled for the firm's asset size and logged this
variable in the regression models. The unit of this variable is “one
thousand RMB Yuan”. The Intangible assets variable is the balanced
payments of the intellectual property rights that the firm owns, such as
patents, designs, licenses and contracts (Hall 1993). The log-transfor-
mation of the firm's Intangible asset is controlled in our estimations
(unit used = one thousand RMB Yuan). The Product sale reflects the
firm's profit potential from selling its products or rendering its services
(unit used = one thousand RMB Yuan). The Exports variable is the re-
ported value of the firm's exported products, indicating its export
capability. We also used a logarithmic transformation of the firm's ex-
port sale revenue to measure its export behaviour (unit used = one
thousand RMB Yuan). Researchers claim that exporting firms will focus
on the exploration of diverse knowledge and resources. This strategic
orientation and tendency will make it essential for the firm to place
emphasis on innovation activities, consequently enhancing the firm's
innovativeness level. We also included variables related to the focal
firm's R&D activity. The R&D variable indicates the firm's expenditures
and investment in R&D activity, and these expenditures are perceived
as indicators for innovation input, and is crucial for innovation, since
increased R&D funding can be channeled to the establishment of new R
&D projects or innovative activities. The unit of this variable is “one
thousand RMB Yuan”, and we regressed the estimated value of the log
(R&D expenditure) in our estimations. The High-tech variable is a
dummy variable. We further made an explanation for non-technological
factors of the focal firm's innovation. The Firm age variable is calculated
by the amount of years since its inception. This variable, Firm age,
captures the firm's accumulation of general experience. Some studies
have stated the importance of tenure for explaining innovation activ-
ities. For example, older firms might be reluctant to innovate and are
not inclined to apply for public R&D promotion programs and funds
(Almus and Czarnitzki 2003; Dai and Cheng 2015). This variable was
included as the log (variable) in our regression. We proxied for Em-
ployee the total number of employees at the end of the focal year, and
PHD employee, the total amount of employees with PHD degrees. We
used these two variables to capture both the quantity and quality of
employees. The firm's employee scale and quality are commonly re-
garded as significant factors in explaining their innovation performance
(Rogers 2004; Gruber et al. 2013). We also took their logarithmic forms
into our estimation because of their skewed distribution. The descrip-
tions of all variables are summarized in Table 3.
4. Analysis and results
The past ten years witnessed a boom in the ZSP (Total revenue
reached 3605.76 billion RMB in 2014). Fig. 1, with two y-axes, displays
an overview of the government subsidies and tax credits, providing the
total numbers and proportion of firms awarded government financial
incentives each year. The left y-axis represents the total number of firms
J. Zhang, J. Guan Technological Forecasting & Social Change 135 (2018) 132–144

Table 3 b = −0.266, p < 0.05; respectively). Further, it is important to see

Variable descriptions.
Variable name Variable description
Dependent variables
Innovation Logarithm of the annual new product sale revenue
Independent variables
Subsidy Logarithm of subsidies the focal firm receives from
government
Tax credit Logarithm of credit the focal firm subtracts from the nation.
Moderating variables
Financial slack The ratio of equity relative to debt
Human slack The ratio of employee salary to sales
Control variables
Product sale Logarithm of the firm's products sale revenue
Exports Logarithm of the firm's exported products sale revenue
Age Logarithm of the number of years from the firm's
establishment to the focal year
Total assets Logarithm of all current and noncurrent assets
R&D Logarithm of the firm's spending in R&D activity
Intangible asset Logarithm of all intangible assets that the firm owns
Employee Logarithm number of the firm's employee
PHD employee Logarithm number of employee with PHD degree
High-tech Dummy variable equal to 1 if the focal firm is a national
high-tech company
in each year, while the right y-axis shows the ratio of firms, which
receive government subsidies accounted or tax credits. The figure il-
lustrates the numbers of firms that peaked in 2007, and then stayed
fairly stable. Between 2005 and 2010, the subsidies and tax credit
proportion fluctuated at around 10% and 22%, respectively. Fig. 1 re-
veals that there has been a marked growth in subsidy and tax credit
proportion since 2011.
Table 4 represents the descriptive statistics for all the study's vari-
ables and the pair-wise correlations among variables. Relatively low to
moderate values of correlations among predicting variables are ob-
served. The biggest VIF (Variance Inflation Factor) in this paper is 2.18,
which is lower than the commonly accepted criterion (Yan and Guan,
2018). Overall, collinearity is not a big threat to our regression ana-
lyses.
Since our Hausman test firmly rejects the random-effects estimator,
we used fixed-effects panel data regression in STATA 12.0 to examine
the hypotheses for new product performance. Table 5 displays regres-
sion results with fixed-effects. Models 1, 3, 5, 7 display the baseline
models with all control variables for innovation performance in the
time t, t + 1, t + 2, t + 3, respectively. Models 2, 4, 6, 8 include both
control and independent variables. All eight models are statistically
significant according to F statistics. We rely on models 2, 4, 6, 8 to test
our hypotheses H 1a and H 1b. Hypothesis 1a predicts that firms'
subsidy favors short-term innovation performance, while hinders long-
term innovation performance. Results from Models 2 and 4 reveal that
firms' subsidy positively and significantly affects their short-term in-
novation performance (b = 0.282, p < 0.01; b = 0.366, p < 0.01,
respectively), while models 6 and 8 show that firms' subsidy negatively
affect their long-term innovation performance (b = −0.106, ns;
that government subsidy may likely exert its negative effect on in-
novation performance from the second year. Thus, H 1a is partially
supported. Hypothesis 1b, which proposes that firms' tax credit favors
both short and long-term innovation performance, is fully supported
from models 2, 4, 6, 8 (b = 0.078, p < 0.01; b = 0.103, p < 0.01;
b = 0.050, p < 0.01; b = 0.038, p < 0.1, respectively).
Furthermore, we did several robustness analyses, and the results are
displayed in Tables 6 and 7. Firstly, we tested our models by using
different samples in Table 6. For example, in Sample 1, we included the
firms whose subsidies and tax credits are both > 0. For Sample 1, the
total number of observations selected is 12,872. The sample proportion
is 7.78% of the total sample. The major distribution of industry is:
general service industry (52.64%); manufacturing industry (21.08%);
key service industry (17.24%); wholesale, retail, hotel, and catering
industries (6.81%). In Sample 2, we exclude firms in the wholesale,
retail, hotel, and catering industries (11.58% of total sample) which
have a low propensity to innovate. For Sample 2, the total number of
observations selected is 146,362. The sample proportion is 88.42% of
the total sample. The major distribution of industry is: general service
industry (57.33%); manufacturing industry (19.82%); key service in-
dustry (18.08%). Table 6 illustrates the changes of coefficient signs of
subsidy in year t + 1 and t + 2, and reveals the positive role of tax
credit in continuous four years.
Secondly, we performed different model specifications in Table 7.
We used the technique of generalized estimating equation (GEE), which
can deal with correlating data and allow for correlation patterns within
clusters. In addition, we used the xtabond2 Stata module to estimate the
System Generalized Method (GMM). We treated all the independent
variables as endogenous and found the results of these models are
mainly consistent with prior findings in Table 5.
Besides, some types of government financial incentives may have
different influences on firms' primary activities. We used the t-test to
compare the means of primary performance between groups with and
without a certain incentive (see Table 8). The sample firms were
grouped into two subsets based on subsidy (has subsidies versus no
subsidies) and tax credit (has tax credits versus no tax credits). Al-
though the results reveal that firms with subsidies (or tax credits)
perform significantly better than firms without subsidies (or tax
credits), we cannot conclude that subsidies/ tax credit lead to more
innovation performance. The primary reason is the selection problem
(Jaffe 2002), which means that recipient firms are not randomly se-
lected, and the selection potentially correlates with other factors, such
as firm characteristics. For example, government agencies are inclined
to follow the “picking the winner strategy”, because larger and more
dominant firms are commonly considered as more potential and pro-
mising (Dai and Cheng 2015). A great deal of research has argued that
firms' heterogeneity should be considered when we assess the effect of
government fiscal incentives (Oh et al. 2009). As for firms' hetero-
geneity, this involves size, tenure, competence, and so forth. We can
observe significant differences between subgroups in product sale, R&
D, assets, and employees, which play important roles in innovation

25000 35%
30%
20000
25%
15000 20% Firms

10000 15% Subsidy proporon

10% Tax credit
5000
5%
0 0%
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Fig. 1. The subsidy and tax credit proportion per year, 2005–2014.

137
J. Zhang, J. Guan Technological Forecasting & Social Change 135 (2018) 132–144

performance. Without controlling these factors, we cannot conclude

0.001
1.00
16
how subsidies (or tax credit) are related to innovation. Taking these
into account, we used the Propensity Score Matching (PSM) method to

−0.07
conduct this comparison analysis. PSM tries to answer what the dif-

1.00

0.06
ference is between the actual innovation performance and the innova-
15

tion performance the firm would have yielded if it had not enjoyed
government subsidies or tax credit. By controlling these firms' hetero-

−0.06
1.00
0.34

0.07
geneity factors and constructing a comparable control sample, PSM can
14

eliminate sample selection bias problems and mitigate potential en-

dogeneity (Caliendo and Kopeinig 2008). Based on PSM results in

−0.04
1.00
0.17
0.40

0.11
Table 9, we found that firms with tax credits have higher innovation
13

performances than other firms in time t, t + 1 and t + 2. Firms with

subsidies have a higher innovation performance than other firms in t;

−0.07
1.00
0.26
0.24
0.31

0.17
while, they have worse innovation performance than others in t + 2
12

and t + 3, which further supports our hypotheses.

Innovation performance has a series of indicators, such as patent or
−0.10

invention. To analyze this, we redefined the dependent variable as the

1.00
0.40
0.45
0.26
0.45

0.25
11

number of patents and inventions. Because there is a long lag between

the application and the issue, we chose patent and invention applica-
−0.08

tions rather than patents and inventions issued to measure innovation

1.00
0.32
0.23
0.20
0.18
0.23

0.06
10

performance. The number of patent or invention applications is coun-

table data and suffer from over-dispersion problems. Thus, we adopted
−0.03

fixed-effect negative binomial (NB) models to conduct this robust test.

1.00
0.15
0.32
0.26
0.20
0.12
0.22

0.07

The results, which are represented in Table 10, show that our findings
9

on the positive effects of tax credits are robust. However, we noticed

−0.17

one notable difference between the results of Tables 5 and 10. In

1.00
0.11
0.36
0.35
0.28
0.13
0.19
0.26

0.02

Table 5, the effects of subsidies on firms' new-product sales are positive

8

in the time t and t + 1, while negative in the time t + 2 and t + 3.

However, in Table 10, the effects of subsidies on firms' patent and in-
−0.04

−0.05
1.00
0.15

0.10
0.29
0.09
0.18
0.09
0.11

0.03

vention are always negative in the time t, t + 1, t + 2 and t + 3. We can

7

explain this finding as follows. Patent innovation aims at legally stating

that patent-holders yield some new ideas that make significant con-
−0.06
1.00
0.15
0.21
0.15
0.16
0.33
0.19
0.16
0.09
0.24

0.06

tributions (Jaffe et al. 1993). Because patent applications are not ne-
Note. All correlations coefficients whose magnitude exceed 0.01 are significant at the 5% significance level.
6

cessarily accompanied by market commercial process, combining with

the exclusive patent right and the rewarding principle of “first-to-file
−0.02
−0.07
1.00
0.23
0.10
0.18
0.09
0.31
0.17
0.38
0.15
0.28
0.18

priority” (Friedman et al. 1991), patents are usually applied within a

5

shorter time to let firms earn the following profits. However, new-
product sales require firms to translate their innovative ideas into
−0.03
−0.02
1.00
0.54
0.22
0.09
0.09
0.30
0.15
0.32
0.15
0.30
0.13
0.23

products, and then translate the products into sales. Thus, improve-
4

ments in new product performance may take a long time (Cohen et al.
1996). Based on this idea, patents and inventions may be faster and
−0.03
−0.03

more sensitive to the effects of subsidies; consequently, the negative

1.00
0.77
0.58
0.21
0.07
0.08
0.33
0.13
0.30
0.14
0.27
0.12
0.21
3

effects of government subsidies on the number of patent and invention

applications may come earlier than those on the new-product sales
−0.02
−0.04

revenue.
1.00
0.78
0.66
0.65
0.21
0.09
0.07
0.33
0.12
0.30
0.12
0.24
0.12
0.20
2

Turning to Hypothesis 2, Table 11 shows models adding moderator

variables and the interaction variables. Models 1–4 reveal the models
−0.02
−0.06

with the added moderator variables (financial slack) and the interaction
1.00
0.76
0.64
0.56
0.69
0.20
0.12
0.06
0.35
0.12
0.29
0.11
0.22
0.13
0.20

variables (subsidy * financial slack). Models 5–8 show the models with
1

the added moderator variables (financial slack) and the interaction

1.69

1.15
1.32

1.23
2.18
1.32
1.39

1.03
1.14

variables (tax credit * financial slack). Models 9–12 represent the

VIF

1.2

1.2

1.2
1.5
–
–
–
–

models with the added moderator variables (human slack) and the in-
Means, standard deviations, and correlations.

teraction variables (subsidy * human slack). Models 13–16 show the

1,822,624
576,139
640,874
696,174
687,322
615,293

models with the added moderator variables (human slack) and the in-
57,943

26,255
59,376
400.35

11,287
17,319
7.35

6.03
0.43

0.06
0.08

teraction variables (tax credit * human slack). To capture any curvi-

SD

linear effects (George, 2005; Sharfman et al., 1988), we also included

the squared terms of financial slack and the squared terms of human
165,626
20,538
22,774
25,311
26,323
22,518

slack in the models. Hypothesis 2a predicts that financial slack posi-

75.26
Mean

1450

2164
3319
7.06

0.82
0.24

0.01
0.01
419
908

tively moderates the relationships between government incentives and

firms' innovation performance, while Hypothesis 2b proposes the op-
10 Intangible asset

posite moderating roles of human slack. Coefficients of interaction

12 PHD employee

16 Financial slack
17 Human slack
2 Innovationt+1
3 Innovationt+2
4 Innovationt+3

terms (subsidy * financial slack and tax credit * financial slack) are
5 Innovationt-1

8 Total assets

15 Tax credit
1 Innovationt

11 Employee

13 High-tech

positive, and most of them are significant in Models 1–8. Coefficients of

14 Subsidy
6 Exports

interaction terms (subsidy * human slack and tax credit * human slack)
Table 4

9 R&D
7 Age

are negative, and most of them are significant in Models 9–16, pro-
viding strong supports for Hypotheses 2a and 2b.

138
J. Zhang, J. Guan Technological Forecasting & Social Change 135 (2018) 132–144

Table 5
Results of regression analyses.
T T+1 T+2 T+3

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8

⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎
Innovationt-1 0.354 0.352 0.131 0.127 −0.193 −0.195 −0.232 −0.233⁎⁎
Exports 0.0271⁎⁎ 0.0227⁎⁎ 0.0296⁎⁎ 0.0234⁎ 0.000885 −0.00386 −0.00277 −0.00693
Age −0.421⁎⁎ −0.431⁎⁎ −1.574⁎⁎ −1.588⁎⁎ −1.884⁎⁎ −1.890⁎⁎ −1.394⁎⁎ −1.397⁎⁎
Total assets −0.203⁎⁎ −0.289⁎⁎ −0.344⁎⁎ −0.437⁎⁎ −0.263+ −0.283+ −0.0832 −0.0962
R&D 0.0426⁎⁎ 0.0422⁎⁎ 0.0762⁎⁎ 0.0757⁎⁎ 0.0682⁎⁎ 0.0677⁎⁎ 0.0474⁎⁎ 0.0471⁎⁎
Intangible asset 0.0889⁎⁎ 0.0647⁎⁎ 0.150⁎⁎ 0.112⁎ 0.171⁎ 0.144⁎ 0.0900 0.0727
Employee 0.115⁎⁎ 0.103⁎⁎ 0.183⁎⁎ 0.166⁎⁎ 0.149⁎⁎ 0.140⁎⁎ 0.0780⁎⁎ 0.0692⁎⁎
PHD employee 0.0248⁎ 0.0161 0.0401 0.0268 0.0808⁎ 0.0727⁎ 0.0706+ 0.0621+
High-tech 0.0422⁎⁎ 0.0253+ −0.0727+ −0.0936⁎ 0.150⁎⁎ 0.141⁎⁎ 0.121⁎ 0.113⁎
Subsidy 0.282⁎⁎ 0.366⁎⁎ −0.106 −0.226⁎
Tax credit 0.078⁎⁎ 0.103⁎⁎ 0.050⁎⁎ 0.038+
Constant 6.623⁎⁎ 5.265⁎⁎ 7.186⁎⁎ 5.187⁎ 8.431⁎⁎ 7.812⁎⁎ 5.881⁎⁎ 4.057+
R2 0.406 0.428 0.176 0.188 0.064 0.069 0.046 0.048
F 2531.04 2095.46 643.25 531.37 434.98 356.43 233.51 191.83

Note.
+
p < 0.1.
⁎
p < 0.05.
⁎⁎
p < 0.01.

Table 6
Robust results of regression analyses.
Sample 1 Sample 2

T T+1 T+2 T+3 T T+1 T+2 T+3

⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎
Innovationt-1 0.351 0.127 −0.1950 −0.2335 0.339 0.127 −0.195 −0.233⁎⁎
Exports 0.023⁎⁎ 0.023⁎ −0.0039 −0.0069 0.024⁎⁎ 0.023⁎ −0.004 −0.007
Age −0.431⁎⁎ −1.588⁎⁎ −1.8898⁎⁎ −1.3973⁎⁎ −0.475⁎⁎ −1.588⁎⁎ −1.890⁎⁎ −1.397⁎⁎
Total assets −0.289⁎⁎ −0.437⁎⁎ −0.2834 −0.0959 −0.240⁎⁎ −0.437⁎⁎ −0.283 −0.096
R&D 0.042⁎⁎ 0.076⁎⁎ 0.0677⁎⁎ 0.0471⁎⁎ 0.042⁎⁎ 0.076⁎⁎ 0.068⁎⁎ 0.047⁎⁎
Intangible asset 0.064⁎⁎ 0.110⁎ 0.1443⁎ 0.0689 0.082⁎⁎ 0.112⁎ 0.144⁎ 0.073
Employee 0.103⁎⁎ 0.166⁎⁎ 0.1395⁎⁎ 0.0691⁎⁎ 0.110⁎⁎ 0.166⁎⁎ 0.140⁎⁎ 0.069⁎⁎
PHD employee 0.016 0.027 0.0727⁎ 0.0622 0.019 0.027 0.073⁎ 0.062
High-tech 0.025 −0.094⁎ 0.1407⁎⁎ 0.1131⁎ 0.022 −0.094⁎ 0.141⁎⁎ 0.113⁎
Subsidy 0.293⁎⁎ 0.384⁎⁎ −0.105 −0.161 0.291⁎⁎ 0.366⁎⁎ −0.106 −0.126
Tax credit 0.078⁎⁎ 0.102⁎⁎ 0.050⁎⁎ 0.047⁎⁎ 0.083⁎⁎ 0.103⁎⁎ 0.050⁎⁎ 0.038
Constant 5.185⁎⁎ 5.060⁎ 7.8139⁎⁎ 3.7892 4.410⁎⁎ 5.187⁎ 7.812⁎⁎ 4.057
R2 0.408 0.177 0.069 0.048 0.480 0.157 0.069 0.048
F 2095.693 531.420 356.428 191.872 1837.408 531.365 356.435 191.826

Note.
⁎
p < 0.05.
⁎⁎
p < 0.01.

Table 7
Robust results of regression analyses.
GEE GMM

T T+1 T+2 T+3 T T+1 T+2 T+3

⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎
Innovationt-1 0.790 0.526 0.155 0.097 0.596 0.972 0.733 0.598⁎⁎
Exports 0.025⁎⁎ 0.096⁎⁎ 0.086⁎⁎ 0.092⁎⁎ 0.037⁎⁎ 0.075⁎⁎ 0.070⁎⁎ 0.096⁎⁎
Age −0.034⁎⁎ −0.573⁎⁎ −0.796⁎⁎ −0.561⁎⁎ −0.038⁎⁎ −0.153⁎⁎ −0.126⁎⁎ −0.041⁎
Total assets −0.080⁎⁎ −0.619⁎⁎ −0.385⁎⁎ −0.192 −0.283⁎⁎ −1.124⁎⁎ −1.031⁎⁎ −0.826⁎⁎
R&D 0.045⁎⁎ 0.130⁎⁎ 0.124⁎⁎ 0.096⁎⁎ 0.059⁎⁎ 0.156⁎⁎ 0.182⁎⁎ 0.156⁎⁎
Intangible asset 0.032⁎⁎ 0.116⁎⁎ 0.206⁎⁎ 0.229⁎⁎ −0.020 0.018 0.043⁎⁎ 0.102⁎⁎
Employee 0.033⁎⁎ 0.313⁎⁎ 0.371⁎⁎ 0.322⁎⁎ 0.166⁎⁎ 0.469⁎⁎ 0.457⁎⁎ 0.354⁎⁎
PHD employee −0.051⁎⁎ −0.067⁎⁎ −0.005 0.024 −0.017 −0.098⁎⁎ −0.012 0.018
High-tech 0.115⁎⁎ 0.155⁎⁎ 0.379⁎⁎ 0.479⁎⁎ −0.106⁎⁎ −0.306⁎⁎ 0.210⁎⁎ 0.531⁎⁎
Subsidy 0.130⁎⁎ 0.298⁎⁎ −0.161⁎ −0.188⁎⁎ 0.332⁎⁎ 0.466⁎⁎ −0.470⁎⁎ −0.430⁎⁎
Tax credit 0.056⁎⁎ 0.133⁎⁎ 0.082⁎⁎ 0.079⁎⁎ 0.080⁎⁎ 0.114⁎⁎ 0.113⁎⁎ 0.188⁎⁎
Constant 0.491 2.856⁎ 2.411 −1.873 2.986⁎⁎ 6.012⁎⁎ 5.436⁎⁎ 2.437⁎⁎
Wald ×2 35505⁎⁎ 16525⁎⁎ 6393⁎⁎ 3702⁎⁎ 38786⁎⁎ 31712⁎⁎ 21376⁎⁎ 14929⁎⁎

Note.
⁎
p < 0.05.
⁎⁎
p < 0.01.

139
J. Zhang, J. Guan Technological Forecasting & Social Change 135 (2018) 132–144

Table 8
The results of t-test.
Variables Subsidies No subsidies T Tax credits No tax credits T

1 Innovationt 3.95 1.85 76.96 3.69 1.63 95.31

2 Innovationt+1 3.55 1.70 61.91 3.37 1.47 80.80
3 Innovationt+2 3.44 1.72 50.88 3.26 1.49 67.67
4 Innovationt+3 3.17 1.59 42.41 2.98 1.37 55.95
5 Innovationt-1 2.91 1.42 75.44 2.69 1.11 89.11
6 Exports 0.97 0.39 43.92 0.95 0.31 60.09
7 Age 2.06 1.80 50.84 1.93 1.80 31.01
8 Total assets 14.99 14.95 42.62 14.97 14.94 32.67
9 R&D 3.65 2.26 58.33 3.73 2.02 90.51
10 Intangible asset 9.57 9.44 59.39 9.52 9.43 47.80
11 Employee 3.98 2.70 1.3e + 2 3.76 2.59 1.5e + 2
12 PHD employee 0.48 0.23 63.11 0.45 0.20 79.85
13 High-tech 0.58 0.19 1.3e + 2 0.55 0.14 1.8e + 2

Table 9 Financial slack = α1 Innovationt − 1 + α2 Exports + α3 Age + α 4 Total assets + α5 R&D

The results of Propensity Score Matching (PSM). + α 6 Intangible asset + α7 Employee + α8 PHD employee
Variable: new product performance + α 9 High − tech + ε1

Year Subsidies No subsidies Difference S.E. T Human slack = β1 Innovationt − 1 + β2 Exports + β3 Age + β4 Total assets + β5 R&D
T 3.96 3.74 0.22⁎⁎ 0.05 4.20
T+1 3.63 3.61 0.02 0.06 0.34 + β6 Intangible asset + β7 Employee + β8 PHD employee
T+2 3.50 3.65 −0.16⁎⁎ 0.07 −2.36 + β9 High − tech + ε2
T+3 3.16 3.32 −0.16⁎⁎ 0.07 −2.16
Year Tax credits No tax credits Difference S.E. T We thus redefine the moderating variables as Financial slacknew and
T 3.61 3.29 0.32⁎⁎ 0.04 7.50
Human slacknew. The Hausman test rejects the random-effects esti-
T+1 3.35 3.13 0.22⁎⁎ 0.05 4.63
T+2 3.21 3.07 0.14⁎⁎ 0.05 2.69 mator, thus we chose a fixed-effects panel data regression. The results
T+3 2.89 2.84 0.05 0.05 0.96 are represented in Table 12. Although we found that some product
terms decrease in significance levels compared to the results in Table 11
in our paper, the positive moderating effects of these alternative vari-
Some scholars suggested that resource slack can be considered as ables are robust.
the resources beyond what is necessary to maintain a firm's operation
and activities. According to this view, we can use the regression spe- Financial slack = 0.001Innovationt − 1 − 0.0002Exports − 0.002Age
cification to estimate expected resources (Richardson 2006). The fitted − 0.071Total assets − 0.0001R&D
value for the regression model is our estimate of the firm's expected
(0.0002∗∗) (0.0003) (0.0008∗)
resource level, and the unexplained portion that is frequently called
regression residual, is the estimate of excessive resources. In this paper, (0.003∗∗) (0.0001)
we constructed two following expectation models for financial slack − 0.004Intangible asset − 0.0007Employee
(equity/debt ratio) and human slack (salary/sale ratio). The residual − 0.002PHD employee − 0.001High − tech + 2.353
value from these expectation models is my new estimate of Fi-
(0.001∗∗) (0.0004)
nancial slacknew and Human slacknew, respectively. A firm with positive
residuals in these expectation models is likely to have excessive fi- (0.0008∗∗) (0.001)
nancial or human resources. The expectation models are:
R2 = 0.370. T-statistics are in parentheses. +
p < 0.1.⁎p
< 0.05.⁎⁎p < 0.01.

Table 10
Robust results of regression analyses.
Dependent variable Patent application Invent application

T T+1 T+2 T+3 T T+1 T+2 T+3

Innovationt 0.017⁎⁎ 0.017⁎⁎ 0.014⁎ 0.012 0.005 0.008 −0.005 0.002

Exports 0.014⁎⁎ 0.014⁎⁎ 0.011⁎ 0.014⁎ 0.012⁎ 0.017⁎⁎ 0.007 0.011
Age 0.218⁎⁎ 0.014 −0.008 0.032 0.276⁎⁎ 0.087⁎⁎ 0.053 0.114⁎⁎
Total assets −0.115 −0.125 −0.144 −0.201⁎ −0.094 −0.047 −0.124 −0.211⁎
R&D 0.052⁎⁎ 0.033⁎⁎ 0.015⁎⁎ 0.005 0.055⁎⁎ 0.032⁎⁎ 0.011⁎⁎ 0.005
Intangible asset 0.115⁎⁎ 0.114⁎⁎ 0.120⁎⁎ 0.106⁎⁎ 0.090⁎⁎ 0.082⁎⁎ 0.106⁎⁎ 0.123⁎⁎
Employee 0.080⁎⁎ 0.069⁎⁎ 0.081⁎⁎ 0.064⁎⁎ 0.058⁎⁎ 0.041⁎⁎ 0.048⁎⁎ 0.011
PHD employee 0.120⁎⁎ 0.116⁎⁎ 0.138⁎⁎ 0.082⁎⁎ 0.140⁎⁎ 0.106⁎⁎ 0.154⁎⁎ 0.115⁎⁎
High-tech 0.330⁎⁎ 0.193⁎⁎ 0.183⁎⁎ 0.060 0.323⁎⁎ 0.166⁎⁎ 0.174⁎⁎ 0.026
Subsidy −0.058⁎ −0.052⁎ −0.064⁎ −0.078⁎ −0.005 −0.028 −0.099⁎ −0.117⁎⁎
Tax credit 0.060⁎⁎ 0.044⁎⁎ 0.058⁎⁎ 0.061⁎⁎ 0.062⁎⁎ 0.042⁎⁎ 0.056⁎⁎ 0.058⁎⁎
Constant −2.601⁎⁎ −0.869 −0.086 1.339 0.005 0.008 −0.005 0.002
Wald ×2 2202.89 832.78 521.96 201.66 1547.26 486.56 293.47 123.55

Note.
⁎
p < 0.05.
⁎⁎
p < 0.01.

140
J. Zhang, J. Guan Technological Forecasting & Social Change 135 (2018) 132–144

Table 11
Results of regression analyses with moderator variables.
T T+1 T+2 T+3 T T+1 T+2 T+3

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8

⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎
Innovationt-1 0.352 0.127 −0.196 −0.234 0.351 0.127 −0.197 −0.234⁎⁎
Exports 0.023⁎⁎ 0.024⁎ −0.004 −0.007 0.023⁎⁎ 0.024⁎ −0.004 −0.007
Age −0.428⁎⁎ −1.583⁎⁎ −1.872⁎⁎ −1.376⁎⁎ −0.424⁎⁎ −1.578⁎⁎ −1.869⁎⁎ −1.382⁎⁎
Total assets −0.213⁎⁎ −0.334⁎ −0.172 −0.062 −0.212⁎⁎ −0.333⁎ −0.168 −0.060
R&D 0.042⁎⁎ 0.076⁎⁎ 0.068⁎⁎ 0.047⁎⁎ 0.042⁎⁎ 0.076⁎⁎ 0.068⁎⁎ 0.047⁎⁎
Intangible asset 0.070⁎⁎ 0.120⁎ 0.160⁎ 0.077 0.070⁎⁎ 0.122⁎ 0.162⁎ 0.080
Employee 0.105⁎⁎ 0.170⁎⁎ 0.144⁎⁎ 0.070⁎⁎ 0.105⁎⁎ 0.170⁎⁎ 0.143⁎⁎ 0.069⁎⁎
PHD employee 0.018 0.029 0.077⁎ 0.063 0.018 0.030 0.078⁎ 0.065
High-tech 0.025 −0.093⁎ 0.141⁎⁎ 0.111⁎ 0.025 −0.094⁎ 0.140⁎⁎ 0.109⁎
Subsidy 0.284⁎⁎ 0.372⁎⁎ −0.126 −0.253⁎ 0.287⁎⁎ 0.375⁎⁎ −0.120 −0.237⁎
Tax credit 0.079⁎⁎ 0.105⁎⁎ 0.053⁎⁎ 0.038 0.080⁎⁎ 0.106⁎⁎ 0.058⁎⁎ 0.046⁎
Financial slack 1.555⁎⁎ 2.378 1.686 −3.218 0.901 1.441 0.749 −2.713
Financial slack square −0.075⁎⁎ −0.111 −0.148⁎ −0.045 −0.073⁎⁎ −0.107 −0.128 −0.012
Subsidy* Financial slack 0.329⁎⁎ 0.231⁎ 0.166 0.489⁎⁎
Tax* Financial slack 0.063 0.402⁎ 0.441⁎ 0.534⁎⁎
Constant 2.068 0.486 3.621 7.244 2.828⁎ 1.585 4.733 6.679
R2 0.409 0.177 0.149 0.048 0.401 0.177 0.168 0.048
F 1652.94 418.42 280.48 151.27 2252.96 764.99 451.85 225.86

Model 9 Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 16

T T+1 T+2 T+3 T T+1 T+2 T+3

Innovationt-1 0.349⁎⁎ 0.129⁎⁎ −0.195⁎⁎ −0.234⁎⁎ 0.350⁎⁎ 0.128⁎⁎ −0.195⁎⁎ −0.234⁎⁎

Exports 0.021⁎⁎ 0.022⁎ −0.005 −0.007 0.022⁎⁎ 0.024⁎ −0.004 −0.007
Age −0.439⁎⁎ −1.653⁎⁎ −1.908⁎⁎ −1.398⁎⁎ −0.446⁎⁎ −1.646⁎⁎ −1.902⁎⁎ −1.395⁎⁎
Total assets −0.323⁎⁎ −0.494⁎⁎ −0.328⁎ −0.111 −0.318⁎⁎ −0.473⁎⁎ −0.315⁎ −0.108
R&D 0.042⁎⁎ 0.075⁎⁎ 0.067⁎⁎ 0.047⁎⁎ 0.041⁎⁎ 0.074⁎⁎ 0.067⁎⁎ 0.047⁎⁎
Intangible asset 0.075⁎⁎ 0.115⁎ 0.151⁎ 0.077 0.077⁎⁎ 0.121⁎ 0.155⁎ 0.078
Employee 0.136⁎⁎ 0.193⁎⁎ 0.166⁎⁎ 0.079⁎⁎ 0.137⁎⁎ 0.195⁎⁎ 0.167⁎⁎ 0.079⁎⁎
PHD employee 0.019 0.031 0.074⁎ 0.063 0.021 0.034 0.076⁎ 0.063
High-tech 0.032⁎ −0.093⁎ 0.143⁎⁎ 0.115⁎ 0.031⁎ −0.094⁎ 0.143⁎⁎ 0.115⁎
Subsidy 0.285⁎⁎ 0.367⁎⁎ −0.102 −0.225⁎ 0.279⁎⁎ 0.352⁎⁎ −0.093 −0.222⁎
Tax credit 0.069⁎⁎ 0.095⁎⁎ 0.044⁎ 0.036 0.078⁎⁎ 0.113⁎⁎ 0.054⁎⁎ 0.038
Human slack −0.113 2.480⁎⁎ 1.048 0.358 0.278 2.586⁎⁎ 0.989 0.020
Human slack square 0.123⁎⁎ 0.099⁎⁎ 0.085⁎⁎ 0.028 0.118⁎⁎ 0.087⁎⁎ 0.078⁎⁎ 0.026
Subsidy* Human slack −0.190⁎⁎ −0.459⁎⁎ −0.275⁎⁎ −0.097
Tax* Human slack −0.351⁎⁎ −0.699⁎⁎ −0.396⁎⁎ −0.087
Constant 0.349⁎⁎ 0.129⁎⁎ −0.195⁎⁎ −0.234⁎⁎ 5.276⁎⁎ 3.707 7.425⁎⁎ 4.170
R2 0.414 0.178 0.070 0.048 0.415 0.179 0.070 0.047
F 1698.56 426.09 283.23 151.08 1705.29 427.63 283.49 151.04

Note.
⁎
p < 0.05.
⁎⁎
p < 0.01.

Human slack = ‐0.0026Innovationt − 1 − 0.0001Exports + 0.023Age paper indicates, as expected, that subsidies increase as firm's innovation
performance only in the short-term, while tax credit is always favour-
− 0.016Total assets − 0.00002R&D
able to a firm's innovation performance in the short and long run. We
(0.0002∗∗) (0.0002) (0.0007∗∗) also suggest that the ability of firms to gain leverage over the govern-
(0.003∗) (0.00007) ment incentive toward the production of innovation performance is
+ 0.008Intangible asset + 0.0217Employee contingent on the resources endowment inside firms. Financial slack
positively moderates the incentive-innovation relationship, but human
+ 0.001PHD employee + 0.0052High − tech + 0.848
slack negatively moderates this relationship. These interesting findings
(0.001∗∗) (0.0003∗∗) have important theoretical implications.
(0.0006) (0.0008∗∗) Firstly, what is good in the short-term is not always good in the
long-term. Government subsidies have long been regarded as a sig-
R2 = 0.139. T-statistics are in parentheses. +p < 0.1.⁎p nificant driver of firms' innovation performance. Subsidies encourage
< 0.05.⁎⁎p < 0.01. firms' innovation and development by supporting innovation projects to
Financial slacknew is the residual from the regression model of fi- acquire increasing social returns or bear huge risks. Recently, there has
nancial slack, and Human slacknew is the residual from the regression been a growing research stream which recognizes the negative effects of
model of human slack. government subsidies, such as R&D crowd effects, resource distortion,
and resource diversion (Wallsten 2000). However, prior studies on
5. Discussion and conclusion government subsidies have not fully examined the innovation effects of
subsidy over time. Further, in developing countries, government sub-
This study uses a panel dataset from the Beijing Zhongguancun sidy policies that support firms' innovation activities have become in-
Science Park (ZSP) to examine the impact of government financial in- creasingly popular, and thus research into subsidies is needed in de-
centives on the innovation performance of Chinese high-tech firms. This veloping countries. In this aspect, we complement prior findings by

141
J. Zhang, J. Guan Technological Forecasting & Social Change 135 (2018) 132–144

Table 12
Robust results of regression analyses with moderator variables.
T T+1 T+2 T+3 T T+1 T+2 T+3

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8

⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎ ⁎⁎
Innovationt-1 0.353 0.130 −0.192 −0.232 0.353 0.130 −0.191 −0.232⁎⁎
Exports 0.022⁎⁎ 0.023⁎ −0.004 −0.007 0.022⁎⁎ 0.023⁎ −0.004 −0.007
Age −0.429⁎⁎ −1.585⁎⁎ −1.885⁎⁎ −1.395⁎⁎ −0.429⁎⁎ −1.585⁎⁎ −1.886⁎⁎ −1.396⁎⁎
Total assets −0.309⁎⁎ −0.473⁎⁎ −0.414⁎⁎ −0.168 −0.306⁎⁎ −0.529⁎⁎ −0.445⁎⁎ −0.197
R&D 0.042⁎⁎ 0.076⁎⁎ 0.068⁎⁎ 0.047⁎⁎ 0.042⁎⁎ 0.076⁎⁎ 0.068⁎⁎ 0.047⁎⁎
Intangible asset 0.065⁎⁎ 0.113⁎ 0.149⁎ 0.076 0.065⁎⁎ 0.115⁎ 0.150⁎ 0.077
Employee 0.104⁎⁎ 0.168⁎⁎ 0.142⁎⁎ 0.070⁎⁎ 0.104⁎⁎ 0.168⁎⁎ 0.142⁎⁎ 0.070⁎⁎
PHD employee 0.015 0.025 0.070 0.060 0.015 0.024 0.070 0.061
High-tech 0.024 −0.094⁎ 0.140⁎⁎ 0.112⁎ 0.024 −0.095⁎ 0.140⁎⁎ 0.113⁎
Subsidy 0.286⁎⁎ 0.379⁎⁎ −0.114 −0.233⁎ 0.286⁎⁎ 0.369⁎⁎ −0.112 −0.224⁎
Tax credit 0.079⁎⁎ 0.104⁎⁎ 0.053⁎⁎ 0.039⁎ 0.079⁎⁎ 0.105⁎⁎ 0.054⁎⁎ 0.041⁎
Financial slack 2.310 −29.045 33.656 58.964 0.583 8.616 9.668 9.155
Financial slack square −0.113⁎⁎ −0.134 −0.242⁎ −0.108 −0.105⁎⁎ −0.228⁎ −0.278⁎ −0.134
Subsidy* Financial slack 0.114 3.479+ 3.426+ 6.496⁎⁎
Tax* Financial slack 0.107 0.979⁎ 0.995⁎ 1.242⁎⁎
Constant 5.496⁎⁎ 5.580⁎⁎ 9.596⁎⁎ 5.028 5.443⁎⁎ 6.475⁎⁎ 10.059⁎⁎ 5.521⁎
R2 0.409 0.177 0.069 0.048 0.408 0.181 0.071 0.048
F 1652.86 418.51 280.45 150.90 1652.87 418.53 280.46 150.79

Model 9 Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 16

T T+1 T+2 T+3 T T+1 T+2 T+3

Innovationt-1 0.352⁎⁎ 0.128⁎⁎ −0.194⁎⁎ −0.233⁎⁎ 0.353⁎⁎ 0.128⁎⁎ −0.194⁎⁎ −0.233⁎⁎

Exports 0.023⁎⁎ 0.024⁎ −0.003 −0.007 0.024⁎⁎ 0.024⁎ −0.003 −0.007
Age −0.426⁎⁎ −1.590⁎⁎ −1.899⁎⁎ −1.407⁎⁎ −0.421⁎⁎ −1.589⁎⁎ −1.897⁎⁎ −1.408⁎⁎
Total assets −0.268⁎⁎ −0.432⁎⁎ −0.292⁎ −0.112 −0.261⁎⁎ −0.432⁎⁎ −0.288 −0.110
R&D 0.042⁎⁎ 0.076⁎⁎ 0.067⁎⁎ 0.047⁎⁎ 0.042⁎⁎ 0.076⁎⁎ 0.067⁎⁎ 0.047⁎⁎
Intangible asset 0.066⁎⁎ 0.113⁎ 0.148⁎ 0.074 0.065⁎⁎ 0.112⁎ 0.147⁎ 0.074
Employee 0.104⁎⁎ 0.168⁎⁎ 0.142⁎⁎ 0.069⁎⁎ 0.105⁎⁎ 0.168⁎⁎ 0.142⁎⁎ 0.069⁎⁎
PHD employee 0.018 0.029 0.074⁎ 0.063 0.018 0.029 0.075⁎ 0.062
High-tech 0.026 −0.091⁎ 0.141⁎⁎ 0.112⁎ 0.026 −0.091⁎ 0.141⁎⁎ 0.112⁎
Subsidy 0.260⁎⁎ 0.351⁎⁎ −0.085 −0.319⁎⁎ 0.283⁎⁎ 0.366⁎⁎ −0.099 −0.228⁎
Tax credit 0.070⁎⁎ 0.097⁎⁎ 0.045⁎ 0.036 0.063⁎⁎ 0.094⁎⁎ 0.041⁎ 0.042⁎
Human slack 30.431⁎⁎ 15.961 8.796 −36.062⁎ 2.626⁎⁎ 0.105 0.596 −2.377
Human slack square 0.158⁎⁎ 0.121⁎⁎ 0.113⁎⁎ 0.049 0.149⁎⁎ 0.120⁎⁎ 0.108⁎⁎ 0.049
Subsidy* Human slack −3.613⁎⁎ −1.948 −1.128 4.029⁎
Tax* Human slack −0.699⁎⁎ −0.230 −0.298⁎ −0.324⁎
Constant 5.156⁎⁎ 5.262⁎⁎ 8.131⁎⁎ 3.499 4.872⁎⁎ 5.152⁎ 7.958⁎⁎ 4.231
R2 0.413 0.178 0.070 0.048 0.412 0.178 0.070 0.048
F 1694.69 421.31 282.24 151.33 1694.37 421.21 282.27 151.07

Note.
+
p < 0.1.
⁎
p < 0.05.
⁎⁎
p < 0.01.

Guan and Yam (2015), who examined the effects of different fiscal in-
struments on firms' performance. Guan and Yam (2015) observed that
direct subsidy not only failed to improve a firm's innovative outcomes,
but also even negatively affected it. Our study broadens this perspective
by arguing that subsidies play distinct roles in short- and long-term
innovative outcomes. The findings of this study, however, go beyond
prior research in arguing and empirically finding that firms generally
generate higher benefits from subsidies only in the earlier period. Our
research also finds that it will take two years for direct subsidies to
hinder firms' performance.
Secondly, the goal of tax credit is to encourage numerous interested
firms to participate in R&D activities (Yigitcanlar et al., 2017). Com-
pared with direct subsidy, tax credit is neutral in relation to the firm's
nature, industry, or innovation direction (Czarnitzki et al. 2011).
Meanwhile, firms are unconstrained to use the relief to fund whatever
projects they think will have a high return rate (Hall and Van Reenen
2000). Our findings reveal that tax credit and subsidy are different on
several key points, extending prior research, which estimated and
compared the effects of both types of tools (Radas et al. 2015). The
empirical results that emerged from our theoretical arguments argue
that, for high-tech companies, tax credit is more significant and long-
standing than direct subsidies. However, tax credit is not always a good
thing; for instance, R&D exploratory projects that require high potential
social benefits or investments in research projects may be less moti-
vated by tax credits (Czarnitzki et al. 2011; Hall and Van Reenen 2000).
Firms are inclined to invest tax credits in the projects with high return
rates (David et al. 2000), thus tax credit receivers will be likely to focus
on activities with short-term benefits. We can also observe that the
regression coefficient of tax credit on innovation performance in models
employing a one-year is bigger than a two-year and three-year lag,
suggesting that firms benefit earlier and more quickly from it.
Thirdly, this study proposes that social capital theory and resource-
based theory should be simultaneously considered and combined to
better explain innovation creation. To our knowledge, few studies at-
tempt to combine these two perspectives to account for the relationship
between government policy and innovation generation among high-
tech companies. Innovativeness is driven by both external and internal
factors (Guan et al. 2015b), and we try to show the importance of the
effects external government incentives (Bloom et al. 2002) and the
contingent roles of internal resource slacks (Tan and Peng 2003;
Paeleman and Vanacker 2015). Our results suggest that financial re-
source slack has a strong positive interaction effect with government

142
J. Zhang, J. Guan
incentives. Furthermore, our results imply that human slack negatively
moderates the relationship between government incentives and firms'
innovation performance. We get statistically significant and negative
interaction coefficients for human slack with both types of govern-
mental instruments. The results suggest that high-tech firms with re-
dundant human resources are less likely to benefit from subsidies or tax
credits.
These findings have some important policy implications, particu-
larly for the Chinese government and firms. Firstly, since direct gov-
ernment subsidies favours firms' short-term innovation performance,
while hindering long-term innovation performance, the Chinese gov-
ernment should keep in mind that not all fiscal policies are effective
tools to promote firms' innovation. The Chinese government intervenes
forcefully in the supply demand-side of the market, and firms are
usually passive adopters under government regulation (Guan and Yam
2015). When government policy makers plan to support R&D projects,
they should pay attention to both the supply and demand sides of
markets and try to mitigate the negative effects of the government's
direct interventions and funding. The subsidy tool in China is a kind of
beforehand direct funding. Thus, the Chinese government can change
its funding or evaluation system, such as adopting a stepwise subsidy
evaluation system. In this spirit, we suggest that government can fund a
part of money before the subsidy project starts, and fund the rest of
money to the firm according to their actual performance during dif-
ferent project stages. Second, since we found that tax credit is a better
and more stable instrument than direct subsidies to promote innova-
tion, we suggest that for high-tech firms in China, the Chinese gov-
ernment takes tax credits as the core and direct subsidy to complement.
It has been noticing that the Chinese government does not set an upper
limit for tax credit rates, which are also commonly higher than some
other developed countries. China's tax credit policy has stayed in its
initial stage with many problems. Some developed countries have es-
tablished detailed rules and regulations on tax credits, such as how to
design the tax rate, run the application, and specific credits for parti-
cular industries or firms. Thus, China can learn from the tax credit
policy of developed economies to more efficiently foster and guide R&D
activities. Finally, we found that the effectiveness of the innovation
policy is related to resource slacks. Firms' financial slack furnishes and
strengthens unequivocal signals of the firms' quality, and enables firms
to create more innovation from their incentives. Firms can invest more
money to develop projects which require a potentially high, long-term
investment, with the help of their own strong financial resource.
However, because of the absorbed, sticky and costly nature of human
resources (Voss et al. 2008), redundant human resources are beneficial
for firms' existing routines and incumbent product areas. Thus, in-
novation creation, which focuses on generations of new ideas and
products, is impeded by human resource slack, because it may be hard
for firms to readily reallocate workers in novel government-granted
projects. In this sense, for those firms wishing to achieve better in-
novation performance, the high value of financial slack and low value
of human slack may help the Chinese government to successfully im-
plement the innovation policy.
These findings should be carefully considered because of some
limitations, which also recommend many promising directions for fu-
ture research. First, the data in this paper comes from all high-tech
companies in the ZSP in China. This sample provides a well-defined
context to examine our theoretical hypotheses, while also implies that
the generalizability of these findings to other contexts becomes pro-
blematic. This generalizability is important, since there are major dif-
ferences in cross-country levels of government incentives. Thus, our
theory would then need to be validated by further empirical studies in
other contexts, such as other countries or other firms. Secondly, we
focused on the joint effects of two types of fiscal incentives and internal
resources, claiming the positive effects of financial slack for reaping the
benefits of subsidy and tax credit from the government. Clarifying other
types of contingent factors in the relationships between government
143
Technological Forecasting & Social Change 135 (2018) 132–144
incentives and innovation generation can provide significant insights
into how firms benefit from government incentives. In addition, a fur-
ther direction to extend our findings presented here is to clarify some
mediating factors and evaluate the indirect impact of government in-
centives on firms' innovativeness. Another possible limitation is that
this study cannot fully eliminate concerns of reverse causality.
Although we lagged our independent variables, our data did not allow
us to deny the possibility that governments introduce R&D tax in-
centives because R&D expenditure is below a presumed optimum. A
better understanding of the mutual influence mechanism between R&D
tax incentives and R&D expenditure is needed to examine the causal
direction in this study. In particular, future longitudinal qualitative
research can examine how a firm's R&D expenditure influences R&D tax
incentives from governments.
Acknowledgements
This study is supported by the Grants from National Natural Science
Foundation of China (Nos. 71673261 and 71373254) and from The
Research Team of Natural Science Foundation of Guangdong Province
in China (2016A030312005). The authors are very grateful for the
valuable comments and suggestions from two anonymous reviewers
and the Editors of the Journal, which significantly improved the quality
of the paper. Special thanks to Mr. Ryan Johnson for his English cor-
rections of the paper.
References
Almus, M., Czarnitzki, D., 2003. The effects of public R&D subsidies on firms' innovation
activities: the case of Eastern Germany. J. Bus. Econ. Stat. 21 (2), 226–236.
Barney, J., 1991. Firm resources and sustained competitive advantage. J. Manag. 17 (1),
99–120.
Bloom, N., Griffith, R., Van Reenen, J., 2002. Do R&D tax credits work? Evidence from a
panel of countries 1979–1997. J. Public Econ. 85 (1), 1–31.
Borrás, S., Edquist, C., 2013. The choice of innovation policy instruments. Technol.
Forecast. Soc. Chang. 80 (8), 1513–1522.
Bradley, S.W., Wiklund, J., Shepherd, D.A., 2011. Swinging a double-edged sword: the
effect of slack on entrepreneurial management and growth. J. Bus. Ventur. 26 (5),
537–554.
Caliendo, M., Kopeinig, S., 2008. Some practical guidance for the implementation of
propensity score matching. J. Econ. Surv. 22 (1), 31–72.
Carboni, O.A., 2017. The effect of public support on investment and R&D: an empirical
evaluation on European manufacturing firms. Technol. Forecast. Soc. Chang. 117,
282–295.
Cohen, M.A., Eliasberg, J., Ho, T.-H., 1996. New product development: the performance
and time-to-market tradeoff. Manag. Sci. 42 (2), 173–186.
Cyert, R.M., March, J.G., 1963. A Behavioral Theory of the Firm. Englewood Cliffs.
Prentice-Hall, Inc, New Jersey.
Czarnitzki, D., Hanel, P., Rosa, J.M., 2011. Evaluating the impact of R&D tax credits on
innovation: a microeconometric study on Canadian firms. Res. Policy 40 (2),
217–229.
Dai, X., Cheng, L., 2015. The effect of public subsidies on corporate R&D investment: an
application of the generalized propensity score. Technol. Forecast. Soc. Chang. 90,
410–419.
Daniel, F., Lohrke, F.T., Fornaciari, C.J., Turner, R.A., 2004. Slack resources and firm
performance: a meta-analysis. J. Bus. Res. 57 (6), 565–574.
David, P.A., Hall, B.H., Toole, A.A., 2000. Is public R&D a complement or substitute for
private R&D? A review of the econometric evidence. Res. Policy 29 (4), 497–529.
Fischer, C., Newell, R.G., 2008. Environmental and technology policies for climate mi-
tigation. J. Environ. Econ. Manag. 55 (2), 142–162.
Fontana, R., Nesta, L., 2009. Product innovation and survival in a high-tech industry. Rev.
Ind. Organ. 34 (4), 287–306.
Friedman, D.D., Landes, W.M., Posner, R.A., 1991. Some economics of trade secret law. J.
Econ. Perspect. 5 (1), 61–72.
George, G., 2005. Slack resources and the performance of privately held firms. Acad.
Manag. J. 48 (4), 661–676.
George, G., Wiklund, J., Zahra, S.A., 2005. Ownership and the internationalization of
small firms. J. Manag. 31 (2), 210–233.
González, X., Pazó, C., 2008. Do public subsidies stimulate private R&D spending? Res.
Policy 37 (3), 371–389.
Gruber, M., Harhoff, D., Hoisl, K., 2013. Knowledge recombination across technological
boundaries: scientists vs. engineers. Manag. Sci. 59 (4), 837–851.
Guan, J.C., Yam, R.C., 2015. Effects of government financial incentives on firms' in-
novation performance in China: evidences from Beijing in the 1990s. Res. Policy 44
(1), 273–282.
Guan, J.C., Yan, Y., 2016. Technological proximity and recombinative innovation in the
alternative energy field. Res. Policy 45 (7), 1460–1473.
J. Zhang, J. Guan
Guan, J.C., Yan, Y., Zhang, J., 2015a. How do collaborative features affect scientific
output? Evidences from wind power field. Scientometrics 102 (1), 333–355.
Guan, J., Zhang, J., Yan, Y., 2015b. The impact of multilevel networks on innovation. Res.
Policy 44 (3), 545–559.
Guellec, D., Van Pottelsberghe, B., 2003. The impact of public R&D expenditure on
business R&D. Econ. Innov. New Technol. 12 (3), 225–243.
Hall, B., Van Reenen, J., 2000. How effective are fiscal incentives for R&D? A review of
the evidence. Res. Policy 29 (4), 449–469.
Hall, B.H., Lerner, J., 2010. The Financing of R&D and Innovation, Handbook of the
Economics of Innovation. Elsevier, North-Holland.
Hall, R., 1993. A framework linking intangible resources and capabiliites to sustainable
competitive advantage. Strateg. Manag. J. 14 (8), 607–618.
Han, J.K., Kim, N., Srivastava, R.K., 1998. Market orientation and organizational per-
formance: is innovation a missing link? J. Mark. 62 (4), 30–45.
Huergo, E., Trenado, M., Ubierna, A., 2016. The impact of public support on firm pro-
pensity to engage in R&D: Spanish experience. Technol. Forecast. Soc. Chang. 113,
206–219.
Jaffe, A.B., 2002. Building programme evaluation into the design of public research-
support programmes. Oxf. Rev. Econ. Policy 18 (1), 22–34.
Jaffe, A.B., Trajtenberg, M., Henderson, R., 1993. Geographic localization of knowledge
spillovers as evidenced by patent citations. Q. J. Econ. 108 (3), 577–598.
Jiao, H., Koo, C.K., Cui, Y., 2015. Legal environment, government effectiveness and firms'
innovation in China: examining the moderating influence of government ownership.
Technol. Forecast. Soc. Chang. 96, 15–24.
Kazadi, K., Lievens, A., Mahr, D., 2016. Stakeholder co-creation during the innovation
process: identifying capabilities for knowledge creation among multiple stakeholders.
J. Bus. Res. 69 (2), 525–540.
Kobayashi, Y., 2014. Effect of R&D tax credits for SMEs in Japan: a microeconometric
analysis focused on liquidity constraints. Small Bus. Econ. 42 (2), 311–327.
Koga, T., 2003. Firm size and R&D tax incentives. Technovation 23 (7), 643–648.
Kung, C.-C., Zhang, L., Kong, F., 2016. How government subsidy leads to sustainable
bioenergy development. Technol. Forecast. Soc. Chang. 112, 275–284.
Lach, S., 2002. Do R&D subsidies stimulate or displace private R&D? Evidence from Israel.
J. Ind. Econ. 50 (4), 369–390.
Lecuona, J.R., Reitzig, M., 2014. Knowledge worth having in ‘excess’: the value of tacit
and firm-specific human resource slack. Strateg. Manag. J. 35 (7), 954–973.
Lee, E.Y., Cin, B.C., 2010. The effect of risk-sharing government subsidy on corporate R&
D investment: empirical evidence from Korea. Technol. Forecast. Soc. Chang. 77 (6),
881–890.
Lerner, J., 2000. The government as venture capitalist: the long-run impact of the SBIR
program. J. Private Equity 3 (2), 55–78.
Levy, D.M., Terleckyj, N.E., 1983. Effects of government R&D on private R&D investment
and productivity: a macroeconomic analysis. Bell J. Econ. 551–561.
Lungeanu, R., Stern, I., Zajac, E.J., 2016. When do firms change technology-sourcing
vehicles? The role of poor innovative performance and financial slack. Strateg.
Manag. J. 37 (5), 855–869.
Luo, X., Deng, L., 2009. Do birds of a feather flock higher? The effects of partner similarity
on innovation in strategic alliances in knowledge-intensive industries. J. Manag.
Stud. 46 (6), 1005–1030.
Mahmood, I.P., Rufin, C., 2005. Government's dilemma: the role of government in imi-
tation and innovation. Acad. Manag. Rev. 30 (2), 338–360.
Mani, S., 2002. Government, Innovation and Technology Policy: An International
Comparative Analysis. Edward Elgar Publishing, Cheltenham.
Mishina, Y., Pollock, T.G., Porac, J.F., 2004. Are more resources always better for growth?
Resource stickiness in market and product expansion. Strateg. Manag. J. 25 (12),
1179–1197.
Moses, O.D., 1992. Organizational slack and risk-taking behaviour: tests of product pri-
cing strategy. J. Organ. Chang. Manag. 5 (3), 38–54.
Mousa, F.T., Reed, R., 2013. The impact of slack resources on high-tech IPOs. Entrep.
Theory Pract. 37 (5), 1123–1147.
Nohria, N., Gulati, R., 1996. Is slack good or bad for innovation? Acad. Manag. J. 39 (5),
1245–1264.
Oh, I., Lee, J.-D., Heshmati, A., Choi, G.-G., 2009. Evaluation of credit guarantee policy
using propensity score matching. Small Bus. Econ. 33 (3), 335–351.
Paeleman, I., Vanacker, T., 2015. Less is more, or not? On the interplay between bundles
144
Technological Forecasting & Social Change 135 (2018) 132–144
of slack resources, firm performance and firm survival. J. Manag. Stud. 52 (6),
819–848.
Porter, M.E., Stern, S., 2001. Innovation: location matters. MIT Sloan Manag. Rev. 42
(4), 28.
Qian, C., Wang, H., Geng, X., Yu, Y., 2017. Rent appropriation of knowledge-based assets
and firm performance when institutions are weak: a study of Chinese publicly listed
firms. Strateg. Manag. J. 38 (4), 892–911.
Qiu, L.D., Tao, Z., 1998. Policy on international R&D cooperation: subsidy or tax? Eur.
Econ. Rev. 42 (9), 1727–1750.
Radas, S., Anić, I.-D., Tafro, A., Wagner, V., 2015. The effects of public support schemes
on small and medium enterprises. Technovation 38, 15–30.
Richard, O.C., Murthi, B., Ismail, K., 2007. The impact of racial diversity on intermediate
and long-term performance: the moderating role of environmental context. Strateg.
Manag. J. 28 (12), 1213–1233.
Richardson, S., 2006. Over-investment of free cash flow. Rev. Acc. Stud. 11 (2–3),
159–189.
Rogers, M., 2004. Networks, firm size and innovation. Small Bus. Econ. 22 (2), 141–153.
Sharfman, M.P., Wolf, G., Chase, R.B., Tansik, D.A., 1988. Antecedents of organizational
slack. Acad. Manag. Rev. 13 (4), 601–614.
Tan, J., Peng, M.W., 2003. Organizational slack and firm performance during economic
transitions: two studies from an emerging economy. Strateg. Manag. J. 24 (13),
1249–1263.
Vanacker, T., Collewaert, V., Paeleman, I., 2013. The relationship between slack re-
sources and the performance of entrepreneurial firms: the role of venture capital and
angel investors. J. Manag. Stud. 50 (6), 1070–1096.
Voss, G.B., Sirdeshmukh, D., Voss, Z.G., 2008. The effects of slack resources and en-
vironmentalthreat on product exploration and exploitation. Acad. Manag. J. 51 (1),
147–164.
Yan, Y., Guan, J., 2018. Social capital, exploitative and exploratory innovations: The
mediating roles of ego-network dynamics. Technol. Forecast. Soc. Chang. 126,
244–258.
Yang, K.-P., Chou, C., Chiu, Y.-J., 2014. How unlearning affects radical innovation: the
dynamics of social capital and slack resources. Technol. Forecast. Soc. Chang. 87
(Supplement C), 152–163.
Yigitcanlar, T., Sabatini-Marques, J., da-Costa, E.M., Kamruzzaman, M., Ioppolo, G.,
2017. Stimulating technological innovation through incentives: perceptions of
Australian and Brazilian firms. Technol. Forecast. Soc. Chang. http://dx.doi.org/10.
1016/j.techfore.2017.05.039.
Wallsten, S.J., 2000. The effects of government-industry R&D programs on private R&D:
the case of the Small Business Innovation Research program. RAND J. Econ. 82–100.
Wright, P.M., Dunford, B.B., Snell, S.A., 2001. Human resources and the resource based
view of the firm. J. Manag. 27 (6), 701–721.
Zhang, X., Wu, J., 2014. Research on effectiveness of the government R&D subsidies:
evidence from large and medium enterprises in China. Am. J. Ind. Bus. Manag. 4 (09),
503.
JingJing Zhang is a PhD at Economics & School of Management, University of Academy
of Chinese Sciences. Her research interest is in Innovation Management. She has pub-
lished 6 papers in international journals such as Research Policy, Scientometrics, JOI,
IJTM and TASM.
JianCheng Guan is a Professor at School of Economics & Management, University of
Academy of Chinese Sciences. His research interest is in Innovation Management. He has
published more than 80 papers in international journals such as Research Policy, TFSC,
Technovation, R&D Management, Regional Studies, JOI, JASIST, Scientometrics, EJOR,
TASM, CER, IJTM,…, etc and also published more than 150 papers in Domestic peer
reviewed journals (most of them in Chinese and some of them in English) as well as in
International Conference proceedings (in English). JianCheng Guan has won four re-
search awards as Principal investigator (Ranking first) issued by Chinese Education
Ministry, Chinese Aviation Ministry, Beijing Municipality in the field of innovation
management. JianCheng Guan has continually ranked Top 2 most cited Chinese
Researcher in the discipline of Social Science, nominated and calculated by Elsevier since
2014.