Plewnia 2017

J Manag Control
DOI 10.1007/s00187-017-0252-y
ORIGINAL PAPER
The benefits of doing good: a meta-analysis of corporate

philanthropy business outcomes and its implications
for management control
Frederik Plewnia1 · Edeltraud Guenther1
© Springer-Verlag GmbH Germany 2017
Abstract Numerous studies have investigated the business benefits of corporate phi-
lanthropy (CP), which can be considered as part of a company’s belief systems.
However, it remains unclear under which conditions and to what extent philan-
thropic activities are associated with corporate financial performance (CFP). This
study sets out to shed light on the relationship between CP and CFP by means of
a meta-analysis. A total of 183 effect sizes from 45 empirical studies are statisti-
cally integrated and analyzed within several subgroups. Particular attention is paid to
time-lagged variable measurements that allow interpretation for causality. The general
relationship between CP and CFP is found to be significantly positive. Furthermore,
results reveal that CP is positively related to subsequent CFP across a wide range
of different study designs, including different operationalizations of CFP and CP,
different sample criteria, and the control for moderating variables. This paper goes
beyond the question of if CP affects CFP and provides a status quo for a more
detailed discussion on how and when CP influences CFP. Consequently, integrat-
ing research on mediating variables and adding perspectives of management control
literature to the CP research field are suggested as promising avenues for future
research.
Keywords Corporate philanthropy · Corporate financial performance · Corporate

giving · Corporate social responsibility · Corporate social performance · Meta-analysis
B Edeltraud Guenther
ema@mailbox.tu-dresden.de
Frederik Plewnia
frederik.plewnia@tu-dresden.de
1 Faculty of Business and Economics, Chair of Environmental Management and Accounting,

Technische Universitaet Dresden, Dresden, Germany
123
F. Plewnia, E. Guenther
JEL Classification M14 · M40
1 Introduction
Against the argument of Friedman (1970) that the only social responsibility of a
corporation is to generate profit, the contemporary understanding is that companies
are closely intertwined with and dependent on the society around them (Porter and
Kramer 2002) and should be controlled accordingly. Consumers, employees, and
other stakeholders expect business organizations to meet legal requirements and min-
imize damage inflicted on society and, moreover, they want them to actively engage
in solving societal problems (Carroll 1991) and expect management attention to be
directed towards these objectives. The associated general concept of business inter-
acting with society is referred to as corporate social responsibility (CSR; Dahlsrud
2008), whereas the combination of CSR principles, processes, and outcomes is defined
as corporate social performance (CSP; Wood 1991). Corporate philanthropy (CP),
understood as the voluntary donation of firm resources to society, forms an impor-
tant part of this concept and is often used as a proxy when financial outcomes of
CSR activities are evaluated (Wang et al. 2015). However, CSP is a multidimen-
sional construct, including aspects such as fair employee treatment, legal compliance,
environmental performance, or community involvement (Rowley and Berman 2000).
Each of these specific dimensions has its own characteristics and their outcomes need
to be controlled (Chenhall 2003) and evaluated separately (Wang et al. 2016). Par-
ticularly for CP, where doing ‘good’ for society is not directly associated with the
main business purpose of a company (Kourula and Halme 2008), there is an ongoing
discussion as to whether corporate donations only inflict costs on a company or if
they may also lead to financial benefits (Liket and Simaens 2015; Gautier and Pache
2015).
Much research has focused on the question of whether philanthropic activities,
besides being socially desirable, can be associated with economic success of com-
panies and has empirically investigated the relationship between CP and corporate
financial performance (CFP). Various studies have used stakeholder theory to argue
that CP would indirectly be connected with financial benefits because it can improve
the reputation of a company, foster customer relations, and increase employee moti-
vation, thereby enhancing the firm’s overall business performance (Seifert et al. 2004;
Wang et al. 2008; Lim 2010; Muller and Kräussl 2011a). While some empirical results
supported these arguments (Tian 2012; Qiu 2013; Iatridis 2015), others did not (Seifert
et al. 2003, 2004). Again others found a negative relationship (Balabanis et al. 1998;
Gao et al. 2012), supporting arguments of agency theory that CP is used by man-
agers for their personal benefit and has a negative relationship with a firm’s financial
performance (Brammer and Millington 2005). Although past meta-analyses have dis-
tinguished CP as a specific dimensions of CSP (Orlitzky et al. 2003; Allouche and
Laroche 2005; Margolis et al. 2009; Wang et al. 2015), these studies did not pay
particular attention to different study designs, which would have allowed for causal
interpretation within the CP subsample, and were unable to resolve confusion sur-
rounding contradictory results.
123
The benefits of doing good: a meta-analysis of corporate…
In consequence, recently conducted literature reviews (Liket and Simaens 2015;

Gautier and Pache 2015) came to the conclusion that the question of whether CP
leads to higher CFP has not yet been answered. Gautier and Pache (2015) found that
“many variations appear across firms, industries, and periods, as robust and widespread
conclusions remain to be drawn” (Gautier and Pache 2015). At the same time, Liket
and Simaens (2015) stated that “when studying the relationship between CP and CFP
in a direct way, the evidence seems to paint a mixed picture.” In this context, our
study intends to advance the discussion on business benefits of CP by providing a
comprehensive empirical and theoretical analysis of subgroups of studies investigating
this phenomenon.
Using a meta-analysis approach (Hedges and Olkin 1985; Hunter and Schmidt
2004), we integrated the results of 45 empirical studies, including 183 effect sizes, on
the relationship between CP and CFP. Based on a contingency perspective, we explore
how different study design factors, such as variable measurement, sample characteris-
tics, as well as controlled moderating variables, influence the results. Particularly, we
focus on the implementation of temporal sequences in study designs, i.e. time-lagged
measurement of variables, as only these results are valid for the interpretation of causal
relationships (Mitchell and James 2001). As a consequence, our study helps to estab-
lish a status quo of a positive relationship between CP and subsequently measured
CFP across a range of different study designs. More importantly, it moves the dis-
cussion away from the question if a relationship between CP and CFP exists towards
the question how and when CP positively affects CFP by identifying aspects that beg
closer examination. Furthermore, we highlight how the levers of control framework
of Simons (1994) might help to explain the causal link between CP and CFP and adds
another intersection to be explored by scholars of both CP and management control
research fields.
After the introduction, the paper is set up as follows. The “theoretical background”
section provides definitions of the analyzed variables, outlines the fundamental inter-
relations between them, and formulates research questions to investigate the influence
of time-lagged variable measurement, study design factors, and methodological mod-
erators on the CP–CFP relationship. The third section briefly explains the employed
method with respect to the data sample, coding, and meta-analytical integration of
study results. In the “results”, meta-analysis outcomes are presented for the gen-
eral relationship between CP and CFP, and for the particular relationship of CP and
subsequent CFP with respect to study design factors and moderating variables. The
“discussion” section sums up findings regarding the posed research questions. Further-
more, limitations of this study and recommendations for further research are presented.
2 Theoretical background
2.1 The relationship between CP and CFP: Doing well by doing good?
Before discussing the relationship between CP and CFP, it is important to clarify what
these terms represent in this study. Very few scholars which have addressed the topic
of CP have offered clear definitions about their understanding of what CP exactly
123
is (Gautier and Pache 2015). This might cause confusion about which activities are
considered as part of CP. Madden et al. (2006) define CP as “the voluntary business
giving of money, time or in-kind goods, without any direct commercial benefit, to one
or more organizations whose core purpose is to benefit the community’s welfare”.
This definition is consistent with definitions used by several other authors (Schwartz
1968; Godfrey 2005) and is taken as the basis of the research conducted within this
study.
In the context of the CP literature, the question has been raised by many if ‘doing
good’ in a societal sense is related to ‘doing well’ in financial terms (Griffin and Mahon
1997; Margolis et al. 2009). Doing well financially is commonly expressed by the term
CFP (Hamann et al. 2013). CFP captures the “social and economic outcomes result-
ing from the interplay among an organization’s attributes, actions, and environment”
(Combs et al. 2005) and is represented by indicators such as profit, sales growth, or
market returns (Hamann et al. 2013).
There are many reasons that suggest a positive relationship between CP and CFP.
On the one hand, with a healthy enterprise that generates considerable profits each
year, decision makers can afford to give money to charity. If most capital is earmarked
for retaining the enterprise’s profitable status and avoiding bankruptcy, there might be
few resources, in the form of time or money, left to engage in philanthropic activities.
For this reason, slack resource theory proposes that strong CFP represents an essential
prerequisite of CP, suggesting a causal effect of CFP on CP (Adams and Hardwick
1998; Buchholtz et al. 1999; Seifert et al. 2003). On the other hand, several theoretical
arguments which are mainly derived from stakeholder theory (Freeman 1984), and
can also be based on ideas of Simons’ levers of controls framework (Simons 1994),
propose that companies can benefit from engaging in CP, indicating a positive causal
effect of CP on CFP.
Engaging with society by donating cash, in-kind goods, or employee worktime
can improve an organization’s performance on several internal and external levels.
Externally, philanthropic giving can enhance reputation. This might help to establish a
more loyal customer base (Luo 2005), ensure the license to operate within communities
(Lim 2010), and serve as a reputational insurance in case of negative events (Godfrey
2005). Furthermore, governmental and financial institutions might be more willing
to support companies which are associated with strong community support (Ye and
Zhang 2011).
Organizations also benefit from CP on the internal level. As Porter and Kramer
(2002) describe, a business is unable to thrive without the right environment. This
includes not only a community’s good will and infrastructure, but also its attractiveness
for possible employees of the company. Companies need to attract a highly qualified
and motivated workforce. Organizational values and belief systems (Simons 1994),
which can be shaped and reinforced by mission statements, company CSR reports, as
well as by corporate donations, can enhance employee identification and provide an
enterprise with a competitive advantage to secure their access to these high-quality
workers (Turban and Greening 1997; Balakrishnan et al. 2011; Arjaliès and Mundy
2013). Moreover, employees within the company show higher job commitment and job
performance when they feel that they are part of a company that believes in giving back
to society (Ke et al. 2015). Finally, value systems supported by CSR and philanthropic
123
activities can provide opportunities for growth and innovation as, “in the absence
of problems, belief systems motivate individuals to search for new ways of creating
value” (Simons 1994).
Despite the aforementioned rationales for a positive relationship between CP and
CFP, there are also some arguments that propose a negative association of CP and
CFP. When comparing different CSR activities for their strategic motivation and the
associated business success, Schaltegger and Burritt (2015) argue that CP is a reac-
tionary form of CSR. As such, CP is “neither expected nor implemented to generate
profits but to keep critics at bay” (p. 10) and will mainly cause costs instead of leading
to economic benefits (Schaltegger and Burritt 2015). Furthermore, even when assum-
ing that well-intended and strategically directed donations might have positive effects
on a firm’s financial performance, agency theory suggest that managers (i.e. agents)
may show opportunistic behavior when deciding over CP activities (Iatridis 2015).
When not properly controlled by the shareholders (i.e. principals), managers may use
firm resources to donate to causes which enhance their personal image instead of the
company’s reputational capital (Navarro 1988; Amran et al. 2007; Cowan et al. 2013).
These agency costs may outbalance potential benefits of CP and may lead to an overall
negative association between CP and CFP.
However, the existence of a positive or a negative relationship between CP and
CFP does not have to be exclusive. Wang et al. (2008) as well as Chen and Lin (2015)
proposed and empirically verified an inverted u-shaped relationship. They suggest
that negative effects of CP, as implied by agency theory, from a certain point on are
greater than the benefits generated by gratifying the stakeholders. As a consequence,
the financial net effect of charitable donations for companies is first positive, than
levels out, and eventually turns negative (Wang et al. 2008).
While various studies have found empirical support for a positive relationship
between CP and CFP (Hall and Rieck 1998; Su and He 2010; Tian 2012; Qiu 2013;
Iatridis 2015), others have not (Seifert et al. 2003, 2004), and again others have even
found negative relationships (Balabanis et al. 1998; Gao et al. 2012). Although the
main purpose of this study is not to reaffirm the general link between CP and CFP
but to take a closer look at different circumstances and study designs which affect this
relationship, it is necessary to establish a starting point of our inquiry. This is why our
research question 1 (RQ1) examines the existence of a generally positive relationship
between CP and CFP across all study designs integrated within our meta-analysis.
RQ1: Is there a positive relationship between CP and CFP?
2.2 Investigating causality with time-lagged measurements
When investigating the relationship between CP and CFP, it is important to keep in

mind that a mere statistical association of two variables does not allow for any conclu-
sion about causality (Mitchell and James 2001; Van der Stede 2014). The relationship
between CP and CFP could be a result of a causal effect of CFP on CP, i.e. companies
with stronger financial performance give more because they can, or of a causal effect
of CP on CFP, i.e. positive effects of CP on reputation and employee motivation lead
to higher CFP, or both. To find evidence that allows for an inference of causation,
123
researchers can measure variables in a temporal sequence by introducing a time lag.

If, for example, CFP is measured 1 year after the corporate donation amount was mea-
sured (subsequent measurement), a statistical correlation between the two measures
can plausibly be interpreted to imply that CP has an effect on CFP. If both indicators
are assessed at the same time (concurrent measurement), on the other hand, no causal
relationship can be inferred from the results (Bausch and Pils 2009).
To take implemented time lags into consideration, CFP indicators are assessed as
to whether they are measured prior, concurrently, or subsequently to the CP indicator.
The consideration of these time lagged measurements within all study designs and
across all subgroups of effect sizes is a key aspect within this study. The relationship
we want to focus on is between CP and subsequently measured CFP to assess the
causal effect of CP on the latter. The existence of this effect would also suggest that
CP activities need to be integrated into management control systems to exploit positive
outcomes of CP, such as stakeholder reputation and employee motivation, and limit
negative consequences in the form of managerial opportunism.
Besides the temporal sequence of measured events, another important aspect that
needs to be considered is the temporal distance of indicator measurement (Mitchell
and James 2001). As positive reputation and employee motivation, representing the
main mediators for the effect of CP on CFP, can be considered as inducing mid- or
long-term effects, a time frame of at least 1 year seems adequate. Conveniently, when
studies investigate this relationship and consider the temporal sequence for these two
variables, they usually depend on archival data provided on a yearly basis and, thus,
implement a sufficient time lag of at least 1 year.
If we want to lead a more specific discussion on two variables’ relationship includ-
ing causality effects, we need to differentiate empirical studies that considered the
temporal sequence of variable measurement. Therefore, based on RQ1, our research
question 2 (RQ2) focuses on differences in findings for prior, concurrent, and subse-
quent measurement of CFP in relation to CP.
RQ2: Does the relationship hold for subgroups of studies that used different temporal
sequences of variable measurement?
2.3 Study design factors
Even if a relationship between CP and subsequent CFP can be empirically shown to

allow for the inference of causality, the important question of the conditions under
which this relationship holds remains. After all, the existence of a positive or negative
relationship might depend on differences concerning the study design and the context
of respective CP activities. To account for these differences within a meta-analysis,
empirical results can be sorted into subgroups according to their study design attributes
and contextual factors. This kind of subgroup analysis might shed some light on
questions about the conditions under which CP can have a significant influence on
CFP. As a consequence, it needs to be investigated how the relationship is influenced
by different study designs. While RQ3 postulates the general question on this issue,
the following subquestions demand more specific answers for different subgroups of
studies.
123
RQ3: Does the relationship hold for subgroups of studies that differed with regard to
study design?
Operationalization of CFP An important and usual attribute to start with is the mea-
surement of CFP (Combs et al. 2005; Hamann et al. 2013). Other studies investigating
the relationship between CFP and CSP or corporate environmental performance (CEP)
have mainly focused on the differentiation of accounting-based indicators, e.g., return
on assets (ROA) or return on equity (ROE), and market-based measures, e.g., stock
market returns or market risk values (Orlitzky et al. 2003; Allouche and Laroche 2005;
Margolis et al. 2009; Endrikat et al. 2014). Following these approaches, the same classi-
fication is introduced for CFP measures in this study. Earlier meta-analyses have found
that the correlations between CSP and CFP are stronger when using accounting-based
rather than market-based measures of CFP (Orlitzky et al. 2003; Margolis et al. 2009).
The argument which was used to explain these findings is presented by Wood and
Jones (1995), who argue that “there is no theory to explain why stockholders would
or would not prefer a company that gives one percent of pre-tax earnings to charity
[…]”. Building on these findings, research question 3a asks:
RQ3a: Does the relationship hold for subgroups of studies that differed with regard
to the employed CFP indicator?
Operationalization of CP Just as there exist different possibilities to measure CFP, sev-
eral distinct indicators have been used to represent the philanthropic activities of firms.
While some have measured CP by a dummy variable, evaluating the overall existence
of charitable donations (Hall and Rieck 1998; Qiu 2013), others have assessed the
total amount of donations made (Patten 2008; Wang and Qian 2011), and again others
computed the percentage of donated amounts relative to total assets, pre-tax income,
or sales (Jia and Zhang 2014; Chen et al. 2015; Iatridis 2015). Generosity in the form
of relative donations could be more important to customers and other stakeholders than
the absolute amount of donations (Patten 2008) or the mere existence of donations.
Furthermore, according to the theory of an inverted u-shaped relationship between CP
and CFP (Wang et al. 2008; Chen et al. 2015), one could expect that the initial partic-
ipation in donation programs yields higher CFP benefits than ever-increasing sums of
charitable donations. This is why we ask whether a positive relationship persists for
these different types of CP indicators.
RQ3b: Does the relationship hold for subgroups of studies that differed with regard
to the employed CP indicator?
Besides the general indicator measuring CP, there are more aspects to be considered
about the charitable donation indicators that have been assessed for empirical analyses.
The first aspect concerns the donation medium that was captured. According to the
definition of CP, charitable donations can be made in the form of cash, in-kind goods,
or volunteered time (Madden et al. 2006). These different forms of resources donated
to society might be perceived differently by stakeholders. Company volunteering pro-
grams, which temporarily release company paid workers from their business duties
in order to work for social causes, can increase motivation of employees and foster
brand identification of customers (Jones 2010; Veleva et al. 2012; Mattila and Hanks
123
2013). In-kind donations, which include the donation of company products, generate
a marketing effect and simultaneously make it possible to write off excess inventory
(Gao et al. 2012). As a result of these additional benefits of specific kinds of donations,
it is expected that these might influence the found relationship but that the positive
relationship holds for all subgroups.
RQ3c: Does the relationship hold for subgroups of studies that differed with regard
to the donation medium included in the CP measure?
The second aspect that is notably different about the type of CP evaluated within
empirical analyses concerns the time horizon of the donation. Whereas many studies
examined continuous donation efforts over periods of 1 year or longer (Diltz 1995;
Balabanis et al. 1998; Chen et al. 2015; Iatridis 2015), others focused on donations
made directly at or after a specific event, mostly natural disasters (Muller and Kräussl
2011b; Gao et al. 2012; Qiu 2013). The special context and increased media attention
of disaster relief donations might influence the impact on CFP, which is why we intend
to investigate RQ3d.
RQ3d: Does the relationship hold for subgroups of studies that differed with regard
to the time horizon of donations included in the CP measure?
Sample data Other contextual factors that could influence the study outcomes are
the country and the year in which the data was gathered. Because of important cul-
tural, economic, institutional, and legal differences between developed and developing
countries, it is argued that the relationship between CP and CFP might differ within
these two contexts (Su and He 2010; Ye and Zhang 2011). Arguments and respective
findings have gone in both directions. On the one hand, in developing countries with
greater market inefficiencies, CP can help to improve political connections and protect
property rights (Su and He 2010). On the other hand, markets are more transparent
and CSR awareness is said to be higher in developed countries. This leads to higher
visibility of companies among more demanding stakeholders and could increase the
influence of CP efforts on CFP (Ye and Zhang 2011; Wang and Qian 2011). Further-
more, especially in China where political affiliation is a more salient issue than in other
countries, managers might use donations to enhance their political connections on a
personal level instead of benefiting the company. Due to these higher agency costs
assumed in countries like China, we need to ask the question:
RQ3e: Does the relationship hold for subgroups of studies that differed with regard
to sample country?
Regarding the year of the investigation, it can be expected that found relationships
have become stronger over time for several reasons. First, study designs are generally
assumed to improve over time to yield stronger relationships (Combs et al. 2011).
Furthermore, for this particular meta-analysis, as the differentiated analysis of spe-
cific CSR dimensions has become more important over recent years (Wang et al.
2016), more studies are expected to exist for recent decades, implying more robust
meta-analysis results with growing sample sizes. Finally, due to digital technology
and extended non-financial reporting, growing transparency between stakeholders and
companies might have increased the impact of strategic philanthropy for business and
123
society (Porter and Kramer 2002; Saiia et al. 2003; Gao et al. 2012; Rangan et al.
2015)
RQ3f: Does the relationship hold for subgroups of studies that differed with regard
to the year of sample collection?
2.4 Methodological moderators
When investigating the relationship between CP and CFP, it is important to consider

that there are also other variables, such as firm size, advertising intensity, or ownership
structures, which might have an impact on the measured relationship. These moderat-
ing variables should be held constant or controlled for when measuring a relationship
and are referred to as control variables. To account for control variables, their effect
can be eliminated by using partial correlations, which hold the moderating variables
constant and measure the ‘true’ relationship between CP and CFP. A number of mod-
erators have been considered by existing studies, including firm size, industry effects,
advertising intensity, R&D spending, past financial performance, or public ownership
(Balabanis et al. 1998; Lev et al. 2010; Chen et al. 2015). Before taking a closer look at
the mentioned control variables individually, it is an interesting question as to whether
the overall relationship between CP and subsequent CFP changes if only partial cor-
relations are considered. A number of studies have already included control variables
into their measurements, but their results can not readily be mixed with bivariate effect
sizes, so these are investigated as a separate sample within our study. Based on the
arguments made above, a significant correlation is still expected when the effects of
other variables are partialed out and the ‘true’ relationship is measured.
RQ4a: Does the relationship hold for studies that calculated partial correlations in
general?
Besides the integration of all partial correlation coefficients controlling for different
sets of control variables, these can also be analyzed separately. For example, numerous
studies controlled for firm size, which has an effect on public visibility and is, thus,
suggested to influence the CP–CFP relationship (Balabanis et al. 1998; Chen et al.
2015; Iatridis 2015). Furthermore, within distinct industries, different levels of CP or
profitability are common and shareholder returns vary, which is why industry effects
should be controlled for (Seifert et al. 2004). Advertising is an important way to
create intangible assets, such as firm reputation or brand identification. As these can
have a considerable impact on future firm performance and at the same time are
associated with CP, advertising intensity is an important variable to be taken into
consideration when evaluating CP business impacts (Wang et al. 2008). Furthermore,
Wang and Qian (2011) argue that stakeholders are unlikely to respond positively to
the CP activities of companies if these failed to fulfil their basic financial needs in
the past. The past financial performance, hence, influences the impact of CP on future
CFP and should be considered as a control variable. Another important moderating
variable is public ownership. Especially in China, where many firms are partly owned
by the state, it is argued that private firms need to be more strategic in directing their
donations to generate public goodwill (Zhang et al. 2010; Qiu 2013). Based on the made
123
arguments, all of the mentioned variables are expected to have a significant impact on
the relationship. Consequently, it should be examined whether the relationship holds
for subgroups of studies that controlled for the respective moderators.
RQ4b: Does the relationship hold for studies that calculated partial correlations con-
trolling for firm size?
RQ4c: Does the relationship hold for studies that calculated partial correlations con-
trolling for industry?
RQ4d: Does the relationship hold for studies that calculated partial correlations con-
trolling for advertising intensity?
RQ4e: Does the relationship hold for studies that calculated partial correlations con-
trolling for past CFP?
RQ4f: Does the relationship hold for studies that calculated partial correlations con-
trolling for public ownership?
3 Method
3.1 Sample
To find studies on the topic of CP, we used the search words (corpor* OR com-
pany* OR business OR enterprise* OR firm*) AND (donat* OR philanthrop* OR
“charitable giving” OR volunteer*). Additionally, wherever possible, we employed
the complementary search terms (data OR empirical OR test OR statistical OR find-
ing* OR result* OR evidence) to narrow down the search results to quantitative studies
(David and Han 2004). This search equation was applied to scan the databases Emerald
Insight and ScienceDirect as well as a number of other databases, namely Academic
Source Complete, Business Source Complete, EconLit with Full Text, E-Journals,
and GreenFILE, which were accessed through the host platform EBSCOhost®. From
the database result lists, only the articles or books which had conducted quantitative
empirical analyses about the relationship between CP and CFP were included in the
sample. The database research for literature was completed by the end of May 2015.
Furthermore, the authors of recently conducted literature reviews (Liket and Simaens
2015; Gautier and Pache 2015) were contacted and kindly provided support in con-
tributing additional articles to the final sample. Including cross references identified
within the found literature, an initial dataset of 69 publications was selected for further
analysis.
To be included in the final sample of studies used for the meta-analysis, the studies
had to report empirical results that were convertible into r correlation coefficients with
the help of respective transformation formulas (Greene 2000; Lipsey and Wilson 2001;
Rosenthal and DiMatteo 2001). Furthermore, as recommended by Dalton and Dalton
(2005), correlations had to rely on a sample size of at least 30 to be included into the
meta-analysis. In case of incomplete data (e.g. no sample size reported), studies were
removed from the sample. After applying these selection criteria, a final sample of 45
publications provided the database for this study. Observations and reported results of
the relevant empirical analyses were distinguished according to the employed sample,
statistical relationship (i.e. bivariate or partial), applied time lag, and operationalization
123
of the CP and CFP variable. Finally, a total of 183 captured effect sizes were used for
the conducted meta-analysis. These effect sizes split into 100 bivariate and 83 partial
correlation coefficients.
3.2 Coding and integration of effect sizes
Besides the captured information regarding effect size statistic and sample size, each
effect size result was coded for the study design factors outlined in the “Theoretical
background” section. The first, and for this study most important, distinction of the
study results was made regarding the time lag that was used for the measurement of
CP and CFP variables. Time lags can be used to imply causal relationships between
two variables. To take implemented time lags into consideration, CFP indicators are
categorized as being measured prior, concurrently, or subsequently to the CP indicator
(Bausch and Pils 2009). Furthermore, according to the employed statistical method,
an effect size was captured as describing either a bivariate or a partial relationship.
The CFP indicator was represented by accounting (e.g. ROA) or market-based
measures (e.g. stock return). Hybrid measures such as Tobin’s Q or earnings per share
(EPS) were assigned to both groups for the subgroup analysis of the CFP indicator. In
all other cases, they entered the analysis only as one effect size. The CP indicator was
assessed by a donation dummy, the total amount of donation, or relative donations
(e.g. percentage of pre-tax earnings). Furthermore, the inclusion of only cash, in-kind,
or volunteered time donations into the CP measure, as well as the time horizon of the
donation (i.e., continuous or disaster relief donations), were coded. The country of
origin and observation year of the sample data was also captured. Additionally, the
partial relationship effect sizes were coded for the moderating control variables they
had included in the analysis. For a more detailed overview of coding categories and
codes, supplementary material from the authors is available upon request.
To integrate the empirical results of the relevant studies, all reported effect sizes
were transformed to r product momentum correlation coefficients using the respec-
tive transformation formulas (Greene 2000; Lipsey and Wilson 2001; Rosenthal and
DiMatteo 2001). Whenever several observations were made within one study, we had
to decide whether the measured effect sizes could be treated as independent from each
other. While treated as independent, each effect size could enter the meta-analysis
individually without introducing a significant bias to the results (Hunter and Schmidt
2004). In case effect sizes from one study could not be considered as independent, e.g.
when observing the same sample and variables with different regression models, they
had to be integrated into a single value or only one of the dependent effect sizes had to
be chosen to be entered into the meta-analysis (Geyskens et al. 2009). Following the
recommendations and examples of other meta-analyses in the field of management,
effect sizes from one study were treated as independent as long as they clearly used
different samples, different time lags, or different operationalization for the CP or
CFP variable (Bijmolt and Pieters 2001; Orlitzky et al. 2003; Dalton and Dalton 2005;
Carney et al. 2011).
However, when a study reported the results of several regression models with chang-
ing moderating variables but identical samples, only the effect size from the model
123
with the highest R2 was used. When a study used summarizing variables, for exam-
ple, by computing the factor of other variables (Balabanis et al. 1998), these values
were not considered in this meta-analysis, but only the individual variables they were
derived from. When a cumulated abnormal stock market return was measured over
several days separately for the same sample, only the cumulated value of the last day
was included into the analysis as it is suggested that the positive effect is stronger over
longer terms (Diltz 1995; Gao et al. 2012).
To consider measurement error, the reliability of the dependent and independent
variable was defined to be 0.8, as recommended by Dalton and Dalton (2005), and
the correlations were adjusted accordingly (Hunter and Schmidt 2004). The corrected
correlations r were integrated using Hedges and Olkin’s (1985) approach of converting
correlations to Fisher’s z and weighting them by their inversed variance. Confidence
intervals were constructed at the 90% level and significance p was reported at the 0.10,
0.05, and 0.01 levels. Homogeneity was tested by calculating Q-statistics, and Q-
between (Q B ) was used to assess if the heterogeneity between subgroups is explained
by the regarding study design factor, in which case a significant p-value is reported
for Q B (Hedges and Olkin 1985). Calculated mean effect sizes rm are all based on the
random effects model (Lipsey and Wilson 2001). Following the recommendation of
Dalton and Dalton (2005), subgroup analysis was only performed with a minimum of
three effect sizes.
When conducting meta-analyses, publication bias, also referred to as the file drawer
problem, is a severe issue which should not be neglected (Geyskens et al. 2009).
While journals tend to focus on articles with significant research results to support or
reject certain hypotheses, file drawers might be full with null results which never get
published (Rosenthal 1979).
To account for possibly unpublished studies which might have altered the results, the
fail-safe k was calculated based on Rosenthal’s (1979) file drawer method. The fail-safe
k indicates the number of studies with a mean effect size of zero that would have been
necessary to yield an insignificant ( p = 0.05) summary effect. All the computations for
integrating effect sizes, identifying confidence intervals, and calculating heterogeneity
indicators were performed with the help of the software program Comprehensive Meta-
Analysis.
4 Results
4.1 The causal relationship between CP and CFP
The analysis of the overall relationship between CP and CFP was based on all the
bivariate effect sizes assessing CP–CFP correlations and yielded a significantly posi-
tive summary effect of 0.094 (k = 100, n = 132, 319, p < 0.001). When taking into
account the file drawer problem by calculating the fail-safe k, these results seemed
robust. An additional number of 59,027 null-effect observations would have been
needed to make the result insignificant ( p = 0.05). This result enables us to answer
RQ1 with a clear yes and establish the positive relationship between CP and CFP as
the starting point for further elaborations of this study (Table 1).
123
Table 1 Results of the bivariate relationship between CP and CFP
Bivariate relationship k n rm 90% CI QB p
CP–CFP (total) 100 132,319 0.094*** 0.067 0.120 0.000

CFP time lag 7.93** 0.019
Prior 19 45,470 0.007 −0.067 0.080 0.882
Concurrent 47 53,549 0.136*** 0.102 0.170 0.000
Subsequent 34 33,300 0.082*** 0.040 0.124 0.002
k number of effect sizes, n total sample size, rm mean product-moment correlation, CI confidence intervals,
Q B between-group Q statistic, p significance level of rm /Q B
*, **, and *** indicate significance of p < 0.1, 0.05, and 0.01, respectively
To take a closer look at this relationship and answer RQ2, the sample of effect sizes
was split up into subgroups according to the temporal sequence of time measurement.
First, it is important to acknowledge that the introduced time lag has a significant
impact on the yielded results (Q B = 7.93, p = 0.019). The integrated effect size
based on concurrent measurements of CP and CFP shows the strongest relationship
between the two variables (rm = 0.136, k = 47, n = 53, 549, p < 0.001). However,
no conclusions on a causal relationship can be drawn from this result. When looking at
subsequent measurement of CFP, we find a significantly positive correlation coefficient
(rm = 0.082, k = 34, n = 33,300, p = 0.002), implying that CP indeed has a
positive impact on subsequent CFP, which would suggest the implementation of CP
into management control systems. When examining the relationship of prior CFP to
CP, on the other hand, no significant relationship can be found (rm = 0.007, k =
19, n = 45, 470, p = 0.882).
Before analyzing further results, it is important to note that correlations based on
concurrent CP and CFP measurements do not imply any causal relationship between
these two variables. For these relationships, it is unclear which indicator influences the
other. Bearing this in mind, research questions are answered based only on relation-
ships which included a time lag in the measurement. Accordingly, the results presented
in Tables 2 and 3 are exclusively derived from subsequent measurements of CFP vari-
ables. However, for the sake of completeness and to set further ground for theoretical
discussions, results including concurrent measurements are reported within the text.
4.2 Study design factors
Operationalization of CFP Results of the performed subgroup analysis according to

different study design factors are presented in Table 2. The first study design fac-
tor which was assumed to influence the relationship between CP and subsequent
CFP was the CFP indicator, represented either by accounting or by market-based
measures (RQ3a). Our meta-analysis results indicate that there is no significant differ-
ence between subgroups which employed either one or the other operationalization of
CFP (Q B = 0.305, p = 0.581). The positive relationship holds for both accounting
(rm = 0.061, k = 14, n = 21, 752, p = 0.028) and market-based measures (rm =
0.085, k = 24, n = 22, 102, p = 0.016). When concurrent measurements were
123
Table 2 Results for the subgroup analysis of the bivariate CP–CFP relationship according to study design
factors
Bivariate relationship k n rm 90% CI QB p
CP–CFP subsequent 34 33,300 0.082*** 0.040 0.124 – 0.002

CFP indicator – – – – – 0.31 0.581
Accounting 14 21,752 0.061** 0.016 0.106 − 0.028
Market 24 22,102 0.085** 0.028 0.143 – 0.016
CP indicator – – – – – 1.46 0.483
Donation dummy 3 373 0.234* 0.040 0.428 – 0.078
Donation amount 17 10,570 0.073** 0.015 0.132 – 0.040
Relative donations 14 22,357 0.070* 0.004 0.136 – 0.085
CP donation medium – – – – – 6.94** 0.031
Cash only 7 492 0.152*** 0.079 0.225 – 0.001
Cash and in-kind 11 9032 0.027 −0.006 0.059 − 0.176
(Incl. volunteering)i (1) – – – – – –
CP time horizon – – – – – 0.00 0.986
Continuous 26 32,202 0.082*** 0.034 0.130 - 0.005
Disaster relief 8 1098 0.083 −0.010 0.176 - 0.148
Sample country – – – – – 0.23 0.632
China 13 23,901 0.055 −0.014 0.124 – 0.193
USA 21 9399 0.077*** 0.043 0.112 - 0.000
Sample year – – – – – 5.31 0.151
1980–1990 5 315 0.155 −0.008 0.318 - 0.141
(1990–2000)i (2) – – – – – –
2000–2010 17 24,309 0.070* 0.008 0.131 – 0.064
k number of effect sizes, n total sample size, rm mean product-moment correlation, C I confidence intervals,
Q B between-group Q statistic, p significance level of rm /Q B
*, **, and *** indicate significance of p < 0.1, 0.05, and 0.01, respectively; i for subgroups with k < 3,
no mean effect size was calculated (in parentheses)
included into the analysis, the relationship for accounting measures (rm = 0.129, k =
48, n = 65, 272, p < 0.001), just as for market measures (rm = 0.074, k = 47, n =
50, 136, p < 0.001), remain significantly positive. In this case, however, there is a sig-
nificant difference between the two subgroups, with accounting measures showing a
stronger relationship (Q B = 3.869, p = 0.049). This result supposes a stronger bidi-
rectional relationship for accounting measures and might explain the findings of earlier
meta-analyses, which suggested that the relationship is stronger when accounting mea-
sures are used (Orlitzky et al. 2003; Margolis et al. 2009). Nevertheless, the latter results
are not to be used to draw a conclusion on the causal relationship between CP and CFP.
Operationalization of CP Just as with the measure of the CFP variable, the indi-
cator representing CP was expected to have an impact on the CP–CFP relationship.
It was questioned whether the positive relationship would hold for the use of a dona-
tion dummy, the absolute donation amount, or the relative donation amount as a CP
measure (RQ3b). Again, our meta-analysis showed that there is no significant dif-
123
Table 3 Results for the subgroup analysis of the partial CP–CFP relationship, including different control
variables
Partial relationship k n rm 90% CI QB p
CP–CFP subsequent 30 58,023 0.171** 0.049 0.293 – 0.029

Firm size – – – – – 0.65 0.420
Controlled 24 41,047 0.139 −0.004 0.282 – 0.128
Not controlled 6 16,976 0.289* 0.060 0.518 – 0.085
Industry effects – – – – – 0.65 0.420
Controlled 24 41,047 0.139 −0.004 0.282 – 0.128
Not controlled 6 16,976 0.289* 0.060 0.518 – 0.085
Advertising intensity – – – – – 0.65 0.422
Controlled 4 13,568 0.098*** 0.027 0.169 – 0.025
Not controlled 26 44,455 0.181* 0.036 0.326 – 0.055
Past CFP – – – – – 0.37 0.541
Controlled 5 8243 0.101 −0.003 0.205 – 0.120
Not controlled 25 49,780 0.176** 0.028 0.325 – 0.045
Public ownership – – – – – 0.25 0.618
Controlled 6 10,834 0.129*** 0.071 0.188 – 0.000
Not controlled 24 50,425 0.180* 0.031 0.329 – 0.062
k number of effect sizes, n total sample size, rm mean product-moment correlation, CI confidence intervals,
Q B between-group Q-statistic, p significance level of rm /Q B
*, **, and *** indicate significance of p < 0.1, 0.05, and 0.01, respectively
ference between the three assessed measurements of CP (Q B = 1.46, p = 0.483).

Regardless of the applied indicator, significantly positive relationships were found
for donation dummy (rm = 0.234, k = 3, n = 373, p = 0.078), donation amount
(rm = 0.073, k = 17, n = 10, 570, p = 0.040), and relative donations (rm =
0.070, k = 14, n = 22, 357, p = 0.085). When concurrent relationships are included,
the result remains the same. There is no significant difference (Q B = 2.93, p = 0.233)
and all indicators show significantly positive relationships between CP and CFP (rm =
0.095, k = 8, n = 5, 095, p = 0.001; rm = 0.152, k = 37, n = 18, 537, p < 0.001;
and rm = 0.090, k = 36, n = 63, 217, p < 0.001, respectively).
Regarding the donation medium included in the donation measure, there was a
limitation of the conducted analysis regarding the assessment of corporate volunteer-
ing. Unfortunately, there were not enough effect sizes which included volunteering
time into their CP measure to calculate a separate effect, maybe because organized
volunteering is a rather new trend. Nevertheless, the relationship was found to be
significantly influenced by the inclusion of in-kind donations in the measurement
(Q B = 6.94, p = 0.031). The effect on the relationship, however, was the reverse
of our expectations. Donating only cash led to a significantly positive relationship
(rm = 0.152, k = 7, n = 492, p = 0.001), while including in-kind donations did
not yield a positive effect (rm = 0.027, k = 11, n = 9, 032, p = 0.176). Adding
concurrent relationships to the results supports these findings in showing a signifi-
cant difference (Q B = 7.84., p = 0.02) and a stronger relationship for only cash
123
(rm = 0.196, k = 23, n = 5, 365, p < 0.001) than for cash and in-kind donations
(rm = 0.060, k = 23, n = 19, 054, p = 0.001).
Similarly, while the subgroup of studies measuring continuous donations shows a
significant positive correlation (rm = 0.082, k = 26, n = 32, 202, p = 0.005), the
assessment of disaster relief donations does not (rm = 0.083, k = 8, n = 1, 098, p =
0.148). When concurrent observations are considered additionally, the picture changes.
There is a significant difference (Q B = 2.98, p = 0.084) and continuous donations
result in a higher correlation (rm = 0.122, k = 68, n = 82, 401, p < 0.001) than dis-
aster relief donations. Nonetheless, the latter relationship is still significantly positive
(rm = 0.066, k = 13, n = 4, 448, p = 0.006). The latter findings can be explained
by another reversed causality reasoning. For example, firms might make continuous
donations depending on their financial situation but will donate in cases of disasters
even if or exactly because they and their environment are in a precarious situation.
Sample data Despite having asked for whether the relationship between CP and
CFP holds for different countries, the only countries with a sufficient sample size
to be compared in our meta-analysis were China and the USA. While data gath-
ered from the USA shows a significantly positive relationship (rm = 0.077, k =
21, n = 9, 399, p < 0.001), samples from China do not show a significant relationship
(rm = 0.055, k = 13, n = 23, 901, p = 0.193), supporting arguments that manage-
rial opportunism might be stronger in China than in the USA (Jia and Zhang 2014).
Including concurrent relationships supports these results by finding a slightly stronger
mean correlation for US-based samples (rm = 0.094, k = 53, n = 25, 762, p < .001)
than for samples from China (rm = 0.056, k = 23, n = 51, 957, p = 0.031).
The last study design factor which was expected to influence the relationship
between CP and subsequent CFP was the year or decade of the data acquisition
(RQ3f). While in the decade 1980–1990 a small sample of effect sizes showed a
positive but insignificant relationship (rm = 0.155, k = 5, n = 315, p = 0.141), for
the decade 1990–2000 there were only two effect sizes captured, which resulted in
omitting this subgroup. For the decade of 2000–2010, however, a significant positive
relationship was found (rm = 0.080, k = 27, n = 24, 947, p = 0.010). When con-
current relationships were added to the analysis, which especially in early years were
often used to assess CFP as a determinant of CP, the relationships found for 1980–
1990, 1990–2000, and 2000–2010 were all significantly positive for this extended
analysis (rm = 0.164, k = 11, n = 1, 071, p = 0.003; rm = 0.078, k = 12, n =
18, 643, p < 0.001; and rm = 0.063, k = 33, n = 56, 729, p = 0.003, respectively).
Summing up the results from the previous section with regard to RQ3, we found that
the positive relationship of CP and CFP holds across a wide range of study designs
and for most investigated subgroups. However, there are several groups in which
the relation did not hold and further examination might be necessary in the future.
Furthermore, we found that the distinction of time-lagged measurement also plays a
crucial role for subgroup analysis and can provide further ground for theorization.
4.3 Methodological moderators
The results in Table 4 show that the average relationship between CP and subsequent
CFP continues to be significantly positive if the effect of other variables is partialed
123
Table 4 Overview of research questions and findings
Research question Research question description Finding of a significant Impact for management control systems (MCS) following the
( p < 0.1) positive levers of control framework of Simons—exemplary pathways for
relationship integration for positive relations (if all subgroups are positive, no
further differentiation is made)
RQ1 Is there a positive relationship between CP Yes CP should get explicit management attention and be
and CFP? integrated in companies’ MCS
RQ2 Does the relationship hold for subgroups of
studies that consider different temporal
sequences of variable measurement?
Prior measurement of CFP (causal No, no relation
interpretation possible)
Concurrent measurement of CFP and CP (no Yes, pos. relation CP should be integrated in diagnostic controls as it is
causal interpretation possible) connected to superior CFP (for specific suggestions
based on subsequent analyses, see below)
Subsequent measurement of CFP (causal Yes, pos. relation CP should be integrated in all four levers as it may be
interpretation possible) connected to new business opportunities

studies that differed with regard to the…
RQ3a …employed CFP indicator? Yes, for all subgroups CP/CFP quota could be part of diagnostic controls and
be used for benchmarking, if appropriate
Accounting-based measures Yes, pos. relation
Market-based measures Yes, pos. relation
123
Table 4 continued
123
RQ3b …employed CP indicator? Yes, for all subgroups CP indicator(s) should be integrated in interactive
controls and specified according to stakeholder
perception
Donation dummy Yes, pos. relation
Donation amount Yes, pos. relation
Relative donations Yes, pos. relation
RQ3c …donation medium included in the CP Only partially
measure?
Cash only Yes, pos. relation Cash donations should be integrated in boundary
systems for subordinate units and corresponding
diagnostic controls could be specified
Cash and in-kind donations No, no relation
Cash, in-kind, and volunteering Insufficient data
RQ3d …time horizon of donations included in the Only partially
CP measure?
Continuous Yes, pos. relation CP integrated in boundary system for subordinate units
Disaster relief No, no relation
RQ3e …sample country? Only partially Integration of CP in MCS should be country specific
China No, no relation
USA Yes, pos. relation
Table 4 continued
RQ3f …year of sample collection? Only partially

1980–1990 No, no relation
1990–2000 Insufficient data
2000–2010 Yes, pos. relation New trend should be communicated to managers via the
MCS
studies that calculated partial
correlations…
RQ4a …in general? Yes, pos. relation CP should be integrated in all four levers as it may be
connected to new business opportunities
RQ4b …controlling for firm size? No, no relation
RQ4c …controlling for industry effects? No, no relation

RQ4d …controlling for advertising intensity? Yes, pos. relation CP should be integrated in MCS for the marketing
department
RQ4e …controlling for past CFP? No, no relation
RQ4f …controlling for public ownership? Yes, pos. relation CP should be integrated in boundary system in order to
avoid managerial opportunism
123
out (rm = 0.171, k = 30, n = 58, 023, p = 0.02). The fail-safe k of 1164 effect
sizes ( p = 0.05) for this result again indicated the absence of a possible publication
bias. When the sample is extended to contain concurrent and subsequent time lag
measurements jointly, this relationship persists (rm = 0.121, k=83, n=96, 687, p =
0.002).
This finding could not be confirmed for all investigated subgroups. However,
for each of the included control variables the calculated mean correlation coef-
ficients remains positive, with significant relationships for advertising intensity
(rm = 0.098, k = 4, n = 13, 568, p = 0.025) and public ownership (rm =
0.129, k = 6, n = 10, 834, p < 0.001), and with insignificant results for firm size
and industry effects, which were both controlled by the same subsample of studies
(rm = 0.139, k = 24, n = 41, 047, p = 0.128), as well as past financial performance
(rm = 0.101, k = 5, n = 8, 243, p = 0.120). When concurrent relationships were
added to the sample, all the subgroups which had controlled for one of these vari-
ables showed significantly positive relationships, except for the subsample which had
controlled for public ownership (rm = 0.083, k = 9, n = 16, 840, p = 0.102).
5 Discussion
5.1 Summary and interpretation of findings
All findings regarding our research questions are discussed in the following sections
and summarized in Table 4. Additionally, following the idea of evidence based man-
agement (Barends et al. 2014), we deduct exemplary consequences for management
control from the results yielded by our meta-analysis. Consequently, for each subgroup
where a significant positive relationship was found, Table 4 outlines conclusions for
the integration of CP into management control systems.
RQ1—general CP–CFP relationship Based on a sample of 100 bivariate effect sizes,
the general existence of a positive relationship between CP and CFP was confirmed
by our meta-analysis results. This relationship formed the foundation of our study and
was tested for its dependency on different study designs.
RQ2—time-lagged measurements The first and most important question that we posed
was whether the positive relationship would hold when temporal sequences of variable
measurement were considered in a way that allows for the implication of a causal rela-
tionship between CP and CFP. The subgroup of effect sizes that measured CFP prior to
CP, which might have shown empirical evidence for the slack resources theory, did not
show a significant positive relationship. Studies that used concurrent and subsequent
measurement of CFP, on the other hand, did show a significant positive relationship.
This finding supports arguments of stakeholder theory and the idea that CP as a part of
organizational belief and value systems can positively influence a company’s financial
performance (Simons 1994). Furthermore, this provides arguments for the stronger
implementation of CP into all four of Simons levers of management control (Simons
1994).
123
RQ3—study design factors Contrary to previous results of other meta-analyses (Orl-

itzky et al. 2003; Margolis et al. 2009), the relationship between CP and CFP was not
found to differ between subgroups using either accounting or market-based measures
(RQ3a). This finding suggests that stockholders either indeed appreciate the reputa-
tional capital built by CP or they perceive giving to charity as a positive signal of
strong financial performance. Either way, stockholders positively acknowledge CP
activities. Similarly, internal accounting performance measures show a positive effect
on CFP.
Besides the different possibilities of measuring CFP, it was also assessed if the indi-
cator used to measure CP changed the results. First of all, the relationship remained
significantly positive regardless of the CP indicator and it made no significant dif-
ference if CP was measured by a donation dummy, absolute donations, or relative
donations (RQ3b). Furthermore, the two lettered subgroups of studies measuring CP
with absolute or relative donations showed about the same strength in its relation-
ship to CFP. A possible explanation for this finding is the limited perception of some
stakeholders, e.g. customers or local communities. These might very well perceive the
absolute amount of donations made by companies, but setting these donation amounts
in relation to total firm resources seems to be beyond the means of the average cus-
tomer. Following this line of thought, large amounts of charitable donations can cause
large amounts of positive public attention and can hence generate a positive CFP
impact on the same level as generosity in terms of relative donations. Finally, support-
ing the idea of an inverted u-shaped relationship between CP and CFP, a quite small
sample indicates that initial participation in CP activities might have a stronger impact
on CFP than ever increasing donations.
With regard to the donation medium included in the CP measurement (RQ3c), and
contrary to our expectations, results indicated that the inclusion of in-kind donations
did yield less business benefits compared to the donation of only cash. A possible
explanation for this finding could be that managerial opportunism, i.e. agency costs,
are a greater problem with in-kind donations because they are less controlled than
philanthropic cash flows. In this context, management control boundary systems might
play a crucial role in directing donations according to the company’s purposes (Simons
1994). Moreover, stakeholders might perceive cash donations to community purposes
as more genuine compared to in-kind donations of company products.
While a positive relationship was empirically shown for studies investigating con-
tinuous CP efforts, the subgroup of studies exploring disaster relief donations did not
show a significant relationship (RQ3d). A reason for the fact that the positive relation-
ship does not hold for effect sizes which measured disaster relief donations might be
that most of the studies in this subsample focused on stock market values a few days
after the incident. Medium and long-term effects of CP cannot adequately be captured
by this approach, which is why no significant relationship was found.
With regard to the country where the data was gathered, results did not confirm
that in developed countries (i.e. the USA) the relationship between CP and CFP
was significantly different from the found relationship in developing countries (i.e.
China; RQ3e). However, while in the USA the relationship was significantly posi-
tive, in China it was not, mainly due to the great heterogeneity of integrated effect
sizes from Chinese samples. A possible explanation for the wide range of results
123
for China could be that within the country there exist regions with large differences
regarding their degrees of development, respective CSR expectations, and market
inefficiencies (Su and He 2010; Wang and Qian 2011). Furthermore, state or pri-
vate ownership of companies might play an important role for the wide spread of
results for China (Zhang et al. 2010; Qiu 2013) as agency costs might be greater
for companies with state affiliations because managers make donations to increase
their personal standing instead of company value (Jia and Zhang 2014). Filtering the
studies according to their time of data acquisition showed that a significant positive
relationship could only be assessed for the last decade, which supports the assump-
tion that the relationship has strengthened over time (RQ3f). Still, in the context of
this finding it should be considered that the majority of empirical research was based
on data from the decade 2000–2010 and very little data was available on previous
decades.
RQ4—methodological moderators Not to include the influence of moderating vari-
ables into such an analysis could easily compromise the validity of any findings
(McWilliams and Siegel 2000). To eliminate the upward bias derived from CP-
correlated variables which also influence CFP, such as advertising intensity or firm
size, the effect of these variables can be controlled by partial correlations. It was found
that, when only partial instead of bivariate correlations were analyzed, the relationship
between CP and subsequent CFP was still significantly positive (RQ4a). However, we
found mixed results when different control variables were considered individually
(RQ4b–f). The relationships of the subgroups of studies that controlled for firm size,
industry effects, or past financial performance, although remaining positive, were no
longer significant. To resolve doubts concerning this matter, future research should
always include the mentioned control variables into the analysis.
Summing up our meta-analysis results, the following can be stated. Regarding
the recent suggestions of Schaltegger and Burritt (2015) and arguments of agency
theory, no evidence was found suggesting that CP would mainly inflict costs and, as
a consequence, negatively influence CFP. On the contrary, a statistically significant
positive relationship between CP and subsequent CFP was found, which allows for
the interpretation of a causal effect of CP on CFP. Additionally, this positive effect
was found to hold for the majority of study designs that were used for its investigation.
Although the relationship was not significantly positive for every subgroup analyzed,
there was not a single case in which a negative relationship was found. Thus, we could
identify no particular context in which CP would harm a company’s financial situation,
but often found evidence for the opposite justifying a stronger attention to CP in the
context of management control systems.
5.2 Limitations and further research
The goal of this study was to provide empirical evidence for the existence of a causal
relationship between CP and CFP and only time-lagged measurements were valid for
this purpose (Mitchell and James 2001). Although the inclusion of concurrent rela-
tionships, which in all cases supported our results, would have considerably enlarged
our sample size, research questions were answered based only on observations which
123
had introduced a time lag into their study designs. As a consequence, significant statis-
tical relationships were found which allowed inferring a positive effect of CP on CFP.
We encourage future research to construct their methodological design according to
the intended findings and to employ time-lagged variables when investigating causal
relationships. This would allow theoretical arguments to be founded on a more solid
ground and future meta-analyses studies to draw on even more extensive samples and
conduct more detailed analyses. Particularly, future research should thrive to close
gaps of knowledge that could not be thoroughly investigated in this study for a lack
of data. Specific attention should be paid to the following aspects.
First of all, relative donations as a possible indicator of CP were often measured
in relation to pre-tax-earnings, sales, or some other accounting variable. However,
donations in comparison to industry standards or competitors as well as donation
growth should be established as alternative CP measures to eliminate cross-lagged
effects between CP and CFP. Although some studies controlled for industry effects
and past CFP within their sample, this should be the standard for all studies investigat-
ing the CP–CFP relationship. Similarly, including donation growth as a CP indicator,
on which almost no data was found within our sample, should become a common
practice. Moreover, motivated by the considerable number of studies which did not
indicate if they included in-kind donations in their measurement, and the almost
non-existence of studies which have addressed the CFP impacts of employee vol-
unteering, it is recommended that more attention is addressed to these issues in future
research.
The most important issue which could not be extensively addressed by our meta-
analysis is a possible inverted u-shape relationship between CP and subsequent CFP,
implying that there is an optimal level of donations (Wang et al. 2008; Chen and
Lin 2015). In this context, theories which propose positive and negative relationships
between CP and subsequent CFP might coexist with one effect influencing the other.
Unfortunately, there were not enough observations to conduct any meta-analytical
analysis to evaluate this u-shape theory. However, this might be an interesting and
important direction of further empirical investigation.
Furthermore, our study design did not allow for the discernment of different kinds
of direct, indirect, and cross-lagged effects between CP and CFP. As a consequence,
although sample subgroups were used to test for the relationship in different contexts
and a starting point for more detailed analyses was established, the theoretical lines
of argumentation remain very presumptive. To go beyond establishing if a causal
relationship exists and explore more soundly how this causal effect is achieved, further
investigation of moderating and mediating variables is necessary. Numerous studies
have already investigated the impact of CP on corporate reputation, on consumer
perception, and on employee motivation. Integrating and advancing research on these
mediators could excellently complement and expand the findings which were made
by our meta-analysis.
Finally, the integration of ideas from management control system literature into
this field of study could provide a fruitful ground for rich discussions on the mech-
anisms of how and under which circumstances CP might have an effect on CFP and
company success. While CP activities can generally be seen to represent company
belief and value systems, which attract and motivate employees, boundary systems
123
might be necessary to restrict managerial opportunism from leading to agency costs

(Simons 1994). Moreover, as stakeholders’ interests represent an increasingly impor-
tant part of a company’s goal dimensions, it seems to be a reasonable step to include
social activities, such as CP, in diagnostic and interactive management control sys-
tems which are used to steer financial and non-financial outcomes and might help
to align both interests (Chenhall 2003). This becomes all the more relevant for
companies where CP, leading to community welfare, is not only a strategic instru-
ment for shared value creation but a desired outcome of their business activities
(Godfrey 2005).
6 Conclusion
At the beginning of this study, we raised the question of whether CP was a form of CSR
that could indeed yield economic benefits for companies. Our meta-analysis results
found that, in addition to a general positive relationship between CP and CFP, there
exists a statistically significant relationship between CP and subsequent CFP, which
allows for the implication of a causal relationship. This finding suggests a stronger
implementation of CP into management control systems to steer and understand the
positive influence of CP on CFP.
This relationship remained significantly positive throughout a variety of subgroup
analyses. Nevertheless, integration of effect sizes in a number of subgroups also
showed insignificant relationships between CP and subsequent CFP. Consequently,
future research should pay careful attention to their study design when comparing
and interpreting findings. Especially time-lagged measurements should be imple-
mented when seeking to investigate causality. Furthermore, control variables should
be included to negate any doubts about the validity of results.
By finding a positive effect of CP on CFP, but pointing out how different conditions
might affect this relationship, this study provides a starting point for investigating
how and under which circumstances CP can positively influence CFP. In this context,
the possible existence of an inverted u-shaped relationship, closer examinations of
mediating variables, and integrating perspectives from management control literature
are suggested as promising new avenues for extending our understanding of this rela-
tionship. Finally, our study has shown that the positive impact of CP on CFP clearly
dominates possible negative effects across a wide range of study designs and sample
contexts. As a consequence, the notion that CP only inflicts costs needs to be finally
left behind as it might mistakenly direct decision-makers to abandon their CP activi-
ties. It is important to make clear that CP can make an important contribution to both
society’s well-being and a corporation’s economic success.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest
123
Appendix 1
Coding on observation level - Effect size data (own illustration)

Effect size data on observation level
Reported effect size statistic - Bivariate correlation

• Correlation coefficient (r )
• Group comparison (F, p, t, Z )
- Partial correlation
• Group comparison (F, p, t, Z )
• Regression coefficient ( p, S E, t)
CP indicator - Donation amount
- Donation dummy
- Generosity (relative donations)
Financial indicator type / Accounting-based CFP measure
Financial indicator - Altman Z-score (hybrid)
- Return on assets (ROA)
- Return on equity (ROE)
- Return on capital employed (ROCE)
- Earnings per share (EPS; hybrid)
- Earnings before taxes (EBT)
(Stock) Market-based CFP measure
- Access to funding
- Market model alpha
- Market risk (Beta)
- Market valuation premium
- Stock Return (abnormal or excess returns)
- Tobin’s Q (hybrid)
Time lag (measurement of CFP vari- - Prior
able in relation to the CP variable) - Concurrent
- Subsequent
Effect size Reported in or converted to correlation coefficient r .
Sample size Reported in firm years per observation
Sign - Negative (2)
- Positive (3)
Note. F = F-value; p = significance level; S E = standard error; t = t-statistic; Z = Z-score corresponding
to respective p-value within standard normal distribution
123
Appendix 2
Coding on observation level - Study design factors (own illustration)

Study design factors on observation level
Donation medium included - Only cash

into CP measurementi - Cash and in-kind
- Including volunteering
Time horizon of donation - Continuous
- Disaster relief (event study)
Country of origin of sample dataii - South Korea
- China
- Malaysia
- UK
- USA
Reference year of sample data First captured individually, then cate-
gorized into time frames/decades
Advertising intensity - Controlled for as moderator
- Not controlled for as moderator
Firm size - Controlled for as moderator
Industry - Controlled for as moderator
R&D spending - Controlled for as moderator
Past CFP - Controlled for as moderator
Public Ownership - Controlled for as moderator
i not assessed were 35 effect sizes (19 concurrent and 16 subsequent) where it was not mentioned which
donation medium was included into the analysis
ii One, two, and two concurrent observations from Malaysia, South Korea, and the UK, respectively, were
neglected from the subgroup analysis because the effect size samples were smaller than three and no
subsequent relationships were reported for these studies
References
Marked references (*) are included in the meta-analysis
*Adams, M., & Hardwick, P. (1998). An analysis of corporate donations: United Kingdom evidence. Journal
of Management Studies, 35, 641–654.
Allouche, J., & Laroche, P. (2005). A meta-analytical investigation of the relationship between corporate
social and financial performance. Revue de Gestion des Ressources Humaines, 57, 18.
*Amato, L. H., & Amato, C. H. (2012). Retail philanthropy: Firm size, industry, and business cycle. Journal
of Business Ethics, 107, 435–448. doi:10.1007/s10551-011-1048-x.
*Amato, L. H., & Amato, C. H. (2007). The effects of firm size and industry on corporate giving. Journal
of Business Ethics, 72, 229–241. doi:10.1007/s10551-006-9167-5.
*Amran, A. B., Ling, L. L., & Sofri, Y. (2007). A study of corporate philanthropic traits among major
malaysian corporations. Social Responsibility Journal, 3, 21–30. doi:10.1108/17471110710840206.
123
Arjaliès, D. L., & Mundy, J. (2013). The use of management control systems to manage CSR strategy:
A levers of control perspective. Management Accounting Research, 24, 284–300. doi:10.1016/j.mar.
2013.06.003.
*Balabanis, G., Phillips, H. C., & Lyall, J. (1998). Corporate social responsibility and economic performance
in the top British companies: Are they linked? European Business Review, 98, 25–44. doi:10.1108/
09555349810195529.
Balakrishnan, R., Sprinkle, G. B., & Williamson, M. G. (2011). Contracting benefits of corporate giving:
An experimental investigation. The Accounting Review, 86, 1887–1907. doi:10.2308/accr-10127.
*Bandeira-de-Mello, R., Marcon, R., & Alberton, A. (2008). Drivers of discretionary firm donations in
Brazil. BAR-Brazilian Administration Review, 5, 275–288.
Barends, E., Rousseau, D. M., & Briner, R. B. (2014). Evidence-based management: The basic principles.
Amsterdam: Center for Evidence-Based Management.
Bausch, A., & Pils, F. (2009). Product diversification strategy and financial performance: Meta-analytic
evidence on causality and construct multidimensionality. Review of Managerial Science, 3, 157–190.
doi:10.1007/s11846-009-0027-4.
Bijmolt, T. H. A., & Pieters, R. G. M. (2001). Meta-analysis in marketing when studies contain multiple
measurements. Marketing Letters, 12, 157–169. doi:10.1023/A:1011117103381.
*Brammer, S., & Millington, A. (2005). Profit maximisation vs. agency: An analysis of charitable giving
by UK firms. Cambridge Journal of Economics, 29, 517–534. doi:10.1093/cje/bei036.
*Brammer, S., & Millington, A. (2008). Does it pay to be different? An analysis of the relationship between
corporate social and financial performance. Strategic Management Journal, 29, 1325–1343. doi:10.
1002/smj.714.
*Brammer, S., & Millington, A. (2004). The development of corporate charitable contributions in the UK:
A stakeholder analysis. Journal of Management Studies, 41, 1411–1434. doi:10.1111/j.1467-6486.
2004.00480.x.
*Brown, W. O., Helland, E., & Smith, J. K. (2006). Corporate philanthropic practices. Journal of Corporate
Finance, 12, 855–877. doi:10.1016/j.jcorpfin.2006.02.001.
*Buchholtz, A. K., Amason, A. C., & Rutherford, M. A. (1999). Beyond resources. Business & Society,
38, 167.
Carney, M., Gedajlovic, E. R., Heugens, P. P. M. A. R., et al. (2011). Business group affiliation, performance,
context, and strategy: A meta-analysis. Academy of Management Journal, 54, 437–460. doi:10.5465/
AMJ.2011.61967812.
Carroll, A. B. (1991). The pyramid of corporate social responsibility: Toward the moral management of
organizational stakeholders. Business Horizons, 34, 39–48. doi:10.1016/0007-6813(91)90005-G.
*Chen, D., Jiang, D., & Yu, X. (2015). Corporate philanthropy and bank loans in China. Pacific-Basin
Finance Journal. doi:10.1016/j.pacfin.2015.03.004.
*Chen, J. C., Patten, D. M., & Roberts, R. W. (2007). Corporate charitable contributions: A corporate social
performance or legitimacy strategy? Journal of Business Ethics, 82, 131–144. doi:10.1007/s10551-
007-9567-1.
*Chen, M.-H., & Lin, C.-P. (2015). The impact of corporate charitable giving on hospitality firm perfor-
mance: Doing well by doing good? International Journal of Hospitality Management, 47, 25–34.
doi:10.1016/j.ijhm.2015.02.002.
Chenhall, R. H. (2003). Management control systems design within its organizational context: Findings
from contingency-based research and directions for the future. Accounting, Organizations and Society,
28, 127–168.
Combs, J. G., Crook, T. R., & Shook, C. L. (2005). The dimensionality of organizational performance and its
implications for strategic management research. Research Methodology in Strategy and Management,
2, 259.
Combs, J. G., Ketchen, J., David, J., Crook, T. R., & Roth, P. L. (2011). Assessing cumulative evidence within
“macro” research: Why meta-analysis should be preferred over vote counting. Journal of Management
Studies, 48, 178–197. doi:10.1111/j.1467-6486.2009.00899.x.
*Cowan, A., Huang, C.-H., Padmanabhan, P., & Wang, C.-H. (2013). The determinants of foreign giving:
An exploratory empirical investigation of US manufacturing firms. International Business Review, 22,
407–420. doi:10.1016/j.ibusrev.2012.05.005.
Dahlsrud, A. (2008). How corporate social responsibility is defined: An analysis of 37 definitions. Corporate
Social Responsibility and Environmental Management, 15, 1–13. doi:10.1002/csr.132.
123
Dalton, D.R., & Dalton, C.M. (2005). Strategic management studies are a special case for meta-analysis. In
D. J. Ketchen, D. D. Bergh (Eds.), Research methodology in strategy and management (pp. 31–63),
vol 2. Emerald Group Publishing Limited.
David, R. J., & Han, S.-K. (2004). A systematic assessment of the empirical support for transaction cost
economics. Strategic Management Journal, 25, 39–58. doi:10.1002/smj.359.
*Diltz, J. D. (1995). The private cost of socially responsible investing. Applied Financial Economics, 5,
69–77.
Endrikat, J., Guenther, E., & Hoppe, H. (2014). Making sense of conflicting empirical findings: A meta-
analytic review of the relationship between corporate environmental and financial performance.
European Management Journal, 32, 735–751. doi:10.1016/j.emj.2013.12.004.
*Fombrun, C., & Shanley, M. (1990). What’s in a name? Reputation building and corporate strategy.
Academy of Management Journal, 33, 233–258. doi:10.2307/256324.
Freeman, R. E. (1984). Strategic management: A stakeholder approach. Boston, MA: Pitman.
Friedman, M. (1970). The social responsibility of business is to increase its profits (pp. 32–33, 122–124).
New York Times Magazine, September 13.
*Galaskiewicz, J. (1997). An urban grants economy revisited: Corporate charitable contributions in the twin
cities, 1979–81, 1987–89. Administrative Science Quarterly, 42, 445–471. doi:10.2307/2393734.
*Gao, F., Faff, R., & Navissi, F. (2012). Corporate philanthropy: Insights from the 2008 Wenchuan Earth-
quake in China. Pacific-Basin Finance Journal, 20, 363–377. doi:10.1016/j.pacfin.2011.11.002.
*Gao, Y. (2011). Philanthropic disaster relief giving as a response to institutional pressure: Evidence from
China. Journal of Business Research, 64, 1377–1382. doi:10.1016/j.jbusres.2010.12.003.
Gautier, A., & Pache, A.-C. (2015). Research on corporate philanthropy: A review and assessment. Journal
Geyskens, I., Krishnan, R., Steenkamp, J.-B. E. M., & Cunha, P. V. (2009). A review and evalua-
tion of meta-analysis practices in management research. Journal of Management,. doi:10.1177/
0149206308328501.
Godfrey, P. C. (2005). The relationship between corporate philanthropy and shareholder wealth: A risk
management perspective. Academy of Management Review, 30, 777–798. doi:10.5465/AMR.2005.
18378878.
Greene, W. H. (2000). Econometric analysis (4th ed.). Upper Saddle River: Prentice Hall., Inc.
Griffin, J. J., & Mahon, J. F. (1997). The corporate social performance and corporate financial perfor-
mance debate twenty-five years of incomparable research. Business & Society, 36, 5–31. doi:10.1177/
000765039703600102.
*Hall, P. L., & Rieck, R. (1998). The effect of positive corporate social actions on shareholder wealth.
Journal of Financial and Strategic Decisions, 11(2), 83–89.
Hamann, P. M., Schiemann, F., Bellora, L., & Guenther, T. W. (2013). Exploring the dimensions of organi-
zational performance a construct validity study. Organizational Research Methods, 16, 67–87. doi:10.
1177/1094428112470007.
*Hapitan, R. B. (2012). An analysis of a signalling model of corporate philanthropy for selected philippine
banks. Journal of International Business Research, 11, 45–56.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Cambridge: Academic Press.
*Hogan, K., Olson, G. T., & Sharma, R. (2014). The role of corporate philanthropy on ratings of corporate
social responsibility and shareholder return. Journal of Leadership, Accountability and Ethics, 11,
108–125.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis (2nd ed.). Thousand Oaks, CA: SAGE
Publications.
*Iatridis, G. E. (2015). Corporate philanthropy in the US stock market: Evidence on corporate governance,
value relevance and earnings manipulation. International Review of Financial Analysis, 39, 113–126.
doi:10.1016/j.irfa.2015.03.004.
*Jia, M., & Zhang, Z. (2014). Donating money to get money: The role of corporate philanthropy in stake-
holder reactions to IPOs. Journal of Management Studies, 51, 1118–1152. doi:10.1111/joms.12090.
*Jia, M., & Zhang, Z. (2013a). Critical mass of women on BODs, multiple identities, and corporate philan-
thropic disaster response: Evidence from privately owned Chinese firms. Journal of Business Ethics,
118, 303–317. doi:10.1007/s10551-012-1589-7.
*Jia, M., & Zhang, Z. (2013b). The CEO’s representation of demands and the corpo-
ration’s response to external pressures: Do politically affiliated firms donate more?
123
CEO . Management and Organization Review, 9, 87–114.

doi:10.1111/j.1740-8784.2012.00297.x.
Jones, D. A. (2010). Does serving the community also serve the company? Using organizational
identification and social exchange theories to understand employee responses to a volunteerism
programme. Journal of Occupational and Organizational Psychology, 83, 857–878. doi:10.1348/
096317909X477495.
Ke, J.-L., Qiu, X.-W., & Jiang, Y.-F. (2015). The effect of corporate philanthropy on organization com-
mitment and job performance of employees: Workplace spirituality as mediating variable. American
Journal of Industrial and Business Management, 5, 466.
Kourula, A., & Halme, M. (2008). Types of corporate responsibility and engagement with NGOs: An
exploration of business and societal outcomes. Corporate Governance: The International Journal of
Business in Society, 8, 557–570. doi:10.1108/14720700810899275.
*Lerner, L. D., & Fryxell, G. E. (1988). An empirical study of the predictors of corporate social performance:
A multi-dimensional analysis. ournal of Business Ethics, 7, 951–959.
*Lev, B., Petrovits, C., & Radhakrishnan, S. (2010). Is doing good good for you? how corporate charitable
contributions enhance revenue growth. Strategic Management Journal, 31, 182–200.
Liket, K., & Simaens, A. (2015). Battling the devolution in the research on corporate philanthropy. Journal
of Business Ethics, 126, 285–308. doi:10.1007/s10551-013-1921-x.
Lim, T. (2010). Measuring the value of corporate philanthropy. New York: Committee Encouraging Cor-
porate Philanthropy.
Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: SAGE Publications.
*Litz, R. A., & Samu, S. (2008). Altruistic by association, altruistic for advantage? Buying groups and
small firm community involvement. Nonprofit and Voluntary Sector Quarterly, 37, 646–667.
Luo, X. (2005). A contingent perspective on the advantages of stores’ strategic philanthropy for influencing
consumer behaviour. Journal of Consumer Behaviour, 4, 390–401. doi:10.1002/cb.19.
Madden, K., Scaife, W., & Crissman, K. (2006). How and why small to medium size enterprises (SMEs)
engage with their communities: An Australian study. International Journal of Nonprofit and Voluntary
Sector Marketing, 11, 49–60. doi:10.1002/nvsm.40.
Margolis, J. D., Elfenbein, H. A., & Walsh, J. P. (2009). Does it pay to be good...and does it matter? A
meta-analysis of the relationship between corporate social and financial performance. Rochester, NY:
Social Science Research Network.
Mattila, A. S., & Hanks, L. (2013). Corporate volunteering programs and consumer perceptions: An infor-
mation processing perspective. Journal of Services Marketing, 27, 572–578. doi:10.1108/JSM-06-
2012-0096.
McWilliams, A., & Siegel, D. (2000). Corporate social responsibility and financial performance: Correlation
or misspccitication? Strategic Management Journal 603–609.
Mitchell, T. R., & James, L. R. (2001). Building better theory: Time and the specification of when things
happen. Academy of Management Review, 26, 530–547.
*Muller, A., & Kräussl, R. (2011a). The value of corporate philanthropy during times of crisis: The sensegiv-
ing effect of employee involvement. Journal of Business Ethics, 103, 203–220. doi:10.1007/s10551-
011-0861-6.
*Muller, A., & Kräussl, R. (2011b). Doing good deeds in times of need: A strategic perspective on corporate
disaster donations. Strategic Management Journal, 32, 911–929. doi:10.1002/smj.917.
Navarro, P. (1988). Why do corporations give to charity? Journal of Business, 61, 65–93.
Nelson, R. L. (1970). Economic factors in the growth of corporate giving. Russell Sage Foundation.
Orlitzky, M., Schmidt, F. L., & Rynes, S. L. (2003). Corporate social and financial performance: A meta-
analysis. Organization Studies, 24, 403–441. doi:10.1177/0170840603024003910.
*Patten, D. M. (2008). Does the market value corporate philanthropy? Evidence from the response to the
2004 tsunami relief effort. Journal of Business Ethics, 81, 599–607. doi:10.1007/s10551-007-9534-
x.
Porter, M. E., & Kramer, M. R. (2002). The competitive advantage of corporate philanthropy. Harvard
Business Review, 80, 56–68.
*Pyo, G., & Lee, H.-Y. (2013). The association between corporate social responsibility activities and
earnings quality: Evidence from donations and voluntary issuance of CSR reports. Journal of Applied
Business Research 29, 945–962. http://journals.cluteonline.com/index.php/JABR/issue/archive
*Qiu, X. (2013). Corporate philanthropic disaster response and post performance: Evidence from China.
International Journal of Management and Marketing Research, 6, 39–51.
123
Rangan, K. V., Chase, L., & Sohel, K. (2015). The truth about CSR. Harvard Business Review, 93(1/2),
40–49.
Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86,
638.
Rosenthal, R., & DiMatteo, M. R. (2001). Meta-analysis: Recent developments in quantiative methods for
literature reviews. Annual Review of Psychology, 52, 59–82.
Rowley, T., & Berman, S. (2000). A brand new brand of corporate social performance. Business & Society,
39, 397–418. doi:10.1177/000765030003900404.
Saiia, D. H., Carroll, A. B., & Buchholtz, A. K. (2003). Philanthropy as strategy when corporate charity
“begins at home”. Business & Society, 42, 169–201. doi:10.1177/0007650303042002002.
Schaltegger, S., & Burritt, R. (2015). Business cases and corporate engagement with sustainability: Dif-
ferentiating ethical motivations. Journal of Business Ethics, 1–19, doi:10.1007/s10551-015-2938-
0.
Schwartz, R. A. (1968). Corporate philanthropic contributions. The Journal of Finance, 23, 479–497. doi:10.
1111/j.1540-6261.1968.tb00821.x.
*Seifert, B., Morris, S. A., & Bartkus, B. R. (2004). Having, giving, and getting: Slack resources, cor-
porate philanthropy and firm financial performance. Business & Society, 43, 135–161. doi:10.1177/
0007650304263919.
*Seifert, B., Morris, S. A., & Bartkus, B. R. (2003). Comparing big givers and small givers: Finan-
cial correlates of corporate philanthropy. Journal of Business Ethics, 45, 195–211. doi:10.1023/A:
1024199411807.
Simons, R. (1994). Levers of control: How managers use innovative control systems to drive strategic
renewal. Brighton: Harvard Business Press.
*Su, J., & He, J. (2010). Does giving lead to getting? Evidence from Chinese private enterprises. Journal
Tian, X. (2012). Empirical Research on Organization Performance and the Influential Factors of Corporate
Philanthropy. International Journal of Digital Content Technology and its Applications, 6(13).
Turban, D. B., & Greening, D. W. (1997). Corporate social performance and organizational attractiveness
to prospective employees. Academy of Management Journal, 40, 658–672. doi:10.2307/257057.
Van der Stede, W. A. (2014). A manipulationist view of causality in cross-sectional survey research. Account-
ing, Organizations and Society, 39, 567–574. doi:10.1016/j.aos.2013.12.001.
Veleva, V., Parker, S., Lee, A., & Pinney, C. (2012). Measuring the business impacts of community involve-
ment: The case of employee volunteering at UL. Business and Society Review, 00453609(117),
123–142. doi:10.1111/j.1467-8594.2011.00395.x.
*Wang, H., Choi, J., & Li, J. (2008). Too little or too much? Untangling the relationship between corporate
philanthropy and firm financial performance. Organization Science, 19, 143–159.
*Wang, H., & Qian, C. (2011). Corporate philanthropy and corporate financial performance: The roles of
stakeholder response and political access. Academy of Management Journal, 54, 1159–1181. doi:10.
5465/amj.2009.0548.
Wang, H., Tong, L., Takeuchi, R., & George, G. (2016). Corporate social responsibility: An overview and
new research directions. Academy of Management Journal, 59, 534–544. doi:10.5465/amj.2016.5001.
Wang, Q., Dou, J., & Jia, S. (2015). A meta-analytic review of corporate social responsibility and
corporate financial performance the moderating effect of contextual factors. Business & Society,
0007650315584317, doi:10.1177/0007650315584317.
Wood, D. J. (1991). Corporate social performance revisited. Academy of Management Review, 16, 691–718.
Wood, D. J., & Jones, R. E. (1995). Stakeholder mismatching: A theoretical problem in empirical research
on corporate social performance. The International Journal of Organizational Analysis, 3, 229–267.
*Ye, K., & Zhang, R. (2011). Do lenders value corporate social responsibility? Evidence from China.
Journal of Business Ethics, 104, 197–206. doi:10.1007/s10551-011-0898-6.
*Zhang, J., Yang, B., Wang, F., & Wang, P. (2012). Corporate philanthropic giving: Active responsibility or
passive ingratiation? Evidence from Chinese family-controlled listed companies. Journal of Applied
Business Research 28, 427–439. http://journals.cluteonline.com/index.php/JABR/issue/archive
*Zhang, R., Rezaee, Z., & Zhu, J. (2010). Corporate philanthropic disaster response and ownership type:
Evidence from Chinese firms’ response to the Sichuan earthquake. Journal of Business Ethics, 91,
51–63. doi:10.1007/s10551-009-0067-3.
123

Plewnia 2017

Uploaded by

Copyright:

Available Formats

Plewnia 2017

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Plewnia 2017

Uploaded by

Copyright:

Available Formats

J Manag Control

The benefits of doing good: a meta-analysis of corporate

Frederik Plewnia1 · Edeltraud Guenther1

© Springer-Verlag GmbH Germany 2017

Keywords Corporate philanthropy · Corporate financial performance · Corporate

1 Faculty of Business and Economics, Chair of Environmental Management and Accounting,

JEL Classification M14 · M40

In consequence, recently conducted literature reviews (Liket and Simaens 2015;

2.2 Investigating causality with time-lagged measurements

When investigating the relationship between CP and CFP, it is important to keep in

researchers can measure variables in a temporal sequence by introducing a time lag.

2.3 Study design factors

Even if a relationship between CP and subsequent CFP can be empirically shown to

2.4 Methodological moderators

When investigating the relationship between CP and CFP, it is important to consider

3.2 Coding and integration of effect sizes

4.1 The causal relationship between CP and CFP

Table 1 Results of the bivariate relationship between CP and CFP

Bivariate relationship k n rm 90% CI QB p

CP–CFP (total) 100 132,319 0.094*** 0.067 0.120 0.000

4.2 Study design factors

Operationalization of CFP Results of the performed subgroup analysis according to

Bivariate relationship k n rm 90% CI QB p

CP–CFP subsequent 34 33,300 0.082*** 0.040 0.124 – 0.002

Partial relationship k n rm 90% CI QB p

CP–CFP subsequent 30 58,023 0.171** 0.049 0.293 – 0.029

ference between the three assessed measurements of CP (Q B = 1.46, p = 0.483).

4.3 Methodological moderators

interpretation possible) connected to new business opportunities

RQ3f …year of sample collection? Only partially

RQ4c …controlling for industry effects? No, no relation

5.1 Summary and interpretation of findings

RQ3—study design factors Contrary to previous results of other meta-analyses (Orl-

5.2 Limitations and further research

might be necessary to restrict managerial opportunism from leading to agency costs

Compliance with ethical standards

Coding on observation level - Effect size data (own illustration)

Reported effect size statistic - Bivariate correlation

Coding on observation level - Study design factors (own illustration)

Donation medium included - Only cash

CEO . Management and Organization Review, 9, 87–114.

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.