* This study answers the questions of why firms bribe government
officials and why some firms pay higher bribes than other firms. Using
insights from residual control theory, we examine how governments
exercise residual rights of control through regulation or state
ownership of firms, and how these rights affect the payment and size of
bribes by firms.
* We argue that firms vary in their exposure and vulnerability to
residual rights of control by government officials, depending on the
firms' characteristics and circumstances. Differences in
firms' exposure and vulnerability to corruption affect their threat
point (i.e. ability to walk away) and thus affect which firms pay bribes
and bribe size.
* Our results show that, at the firm level, bribe size depends on
how much a government can exercise residual rights of control and the
firm's threat point. At the same time, at the country level, the
type of corruption matters; pervasive corruption is positively related,
while arbitrary corruption is negatively related, to bribes paid.
Keywords: Residual control theory * Corruption * Bribery *
Pervasive and arbitrary corruption
Bribes have been paid at least since 3400 BC, according to
archaeologists who found an Assyrian tomb listing the names of
"employees accepting bribes" (Martin 1999, p. 1). Despite the
longevity of bribery (Martin 1999), our understanding as management
scholars of bribery is still limited. For example, how do firm
characteristics affect bribes? How does the existing pattern of
corruption in a country affect an individual firm's propensity to
bribe? These are the questions we address in this paper.
Following past research (Doh et al. 2003), we define corruption as
the abuse or misuse of public power for private (personal) benefit (1).
Our goal is to "lift the veil on corruption" by developing and
testing a management perspective on bribery that incorporates firm
heterogeneity. To do this, we use insights from residual control theory
(Grossman and Hart 1986; Hart and Moore 1990; Tirole 1999) to examine
how governments exercise residual rights of control through regulation
or state ownership of firms, and how the firm's threat point (i.e.,
its ability to walk away) affect the level of bribes paid to government
The residual control theory of the firm (Grossman and Hart 1986;
Hart and Moore 1990) argues that actual behavior of a firm depends on
who owns the residual rights to control the firm's assets. Residual
rights are "the rights to determine the uses of assets under
circumstances that are not covered by contractual terms" (Foss and
Foss 1999, p. 4). Politicians or government officials can wield their
residual rights by imposing rules and regulations on firms. This is why
Johnson et al. (1998, p. 387) argue "[i]n most countries
politicians maintain property rights in firms, typically in the form of
residual control rights ..."
Based on their characteristics and circumstances, firms vary in
their exposure and vulnerability to residual ownership by government
officials. (2) We argue that a firm's exposure (in the sense of the
number of stressors or pressures placed on firms by exogenous events) to
public corruption should vary depending on the pervasiveness of national
corruption and the frequency with which a firm's activities brings
it into contact with government officials. A firm's vulnerability
(in the sense of its ability to resist these stressors or pressures) to
corruption should also vary, depending on the resources (financial,
political or otherwise) firms have at their disposal, which make them
better able to resist these pressures. Residual control theory suggests
that, the greater firms' exposure and vulnerability to corruption,
the more likely are they to bribe government officials (3).
This suggests that residual control theory is an appropriate
theoretical lens for understanding bribery. In our paper, we examine how
firm- and country-level characteristics affect an individual firm's
size of bribes paid (Clarke and Xu 2004; Shleifer and Vishny 1993;
Svensson 2003). We argue that, at the country level, the magnitude of
the perceived level and ambiguity of public sector corruption in an
economy, which affects the firm's exposure to corruption, will be
an important predictor of an individual firm's bribes. At the firm
level, we argue three characteristics of firms are important predictors
of the firm's vulnerability to corruption, and thus of the
magnitude of bribes paid to government officials: foreign ownership,
government regulation, and state ownership. Our paper proceeds as
follows. In Sect. 2, we develop a residual control theory of bribery
incorporating insights from management literature. Section 3 empirically
tests our model. Section 4 discusses the results and concludes the
When it comes to bribery, we argue that while it is important to
understand the environment a firm is in (which affects the firm's
exposure to corruption), it is also important to understand that even in
the same environment, depending on how much bargaining power a firm
possesses vis-a-vis government officials (which affects the firm's
vulnerability to corruption), firms differ in how much they bribe. In
other words, examining firm bribing behaviors from either only firm or
institutional level can be misleading. For example, even when the
majority of the firms give up potential business opportunities in a
country because of the high level of corruption, those firms with high
bargaining power may actually do well. To the best of our knowledge, our
study is the first in the management literature to use the theoretical
lens of residual control theory to examine why firms bribe government
Residual Control Theory
The residual control theory of the firm (Grossman and Hart 1986;
Hart and Moore 1990) argues that the actual ownership of the firm
depends on who owns the residual rights to control assets; that is
"the rights to determine the uses of assets under circumstances
that are not covered by contractual terms" (Foss and Foss 1999, p.
4). The central assumption behind this theory is that real world
contracts are incomplete because the allocation of control rights cannot
be fully specified in advance. Due to the incompleteness of contracts,
firms, more often than not, rather than writing comprehensive contracts
among parties, decide which party owns the residual rights and the owner
of the residual rights decides how the assets are used that are not
specified under the contract.
In turn, since property rights protect their holders against
expropriation of their investment, the allocation of property rights
determines who holds the control of residual rights. Agents that secure
control of residual rights have greater bargaining power and can
determine "who wins" in the ex post outcome. In addition, the
exercise of property rights is limited by the indispensability of the
second party to the expost production process. Even though one agent
(the government official) controls the residual rights, that control is
limited by the threat point, the point at which the other agent (the
owner of the firm) can walk away from the agreement.
When the firm is indispensable to the agreement, the government
agent cannot extract gains beyond the point where the firm's owner
would decide to give up all ownership claims over the firm's assets
and exit the industry; at that point, exiting as one of the possible
alternatives is better for the firm's owner than staying in the
agreement. Thus, the final bargain depends not only on who has the
residual rights of control, but also on the "ability to walk away
from the table" of the agent without the residual rights of
control. "Walking away from the table" can, of course, involve
a range of actions, the most serious of which would be exiting the
industry or country. Firms can also choose less drastic actions such as
building consortiums, renegotiating terms with government officials, and
even whistle blowing.
In this sense, residual contract theory matches well with
stakeholder theory in management in that it is not only the firm, but
also other agents that have stakes in the firm, that are important in
considering how firms make decision (Donaldson and Preston 1995).
Different stakeholders can have a bigger say in different institutional
environments. For example, while shareholders are the most important
stakeholders in the United States, employees are more important in
corporatist European firms, while managers are more important in Japan
(Economist 1993). In this research, we focus on the role of government
officials in taking stakes in firm management.
The Market for Bribes
In a corrupt society where government officials seek private gains
from their relationships with firms, the bargaining power of officials
can be enormous and bribery is likely to occur (Clarke and Xu 2004).
Seeking private benefits, government officials want and demand bribes
from firms. However, even when corruption is illegal, soliciting and
taking bribes is potentially costly for officials, with the cost
depending on the probability of being caught and the size of the
penalty. Each official therefore weighs the benefits against the costs,
at the margin, in deciding whether and how much to demand in bribes.
Paying a bribe imposes a direct cost in the form of reduced cash
flow. Since bribe paying is costly, the firm's willingness to offer
a bribe and the size of the bribe should depend on its perception of the
likely short-term and long-term benefits provided by the government
official. However, the firm knows that the bribe bargain may not be
sustainable; the official could fail to deliver on his/her commitments
or could come back and request an additional bribe. In either situation,
the firm cannot appeal to a court of law because corruption contracts
are not legally enforceable. Moreover, there can be additional costs on
the firm if the government punishes bribe payers in addition to those
receiving bribes (for example, the US Foreign Corrupt Practices Act
punishes firms paying bribes); this latter cost depends on the
probability of being caught and the size of the punishment. When the
probability of getting caught is low, government officials are more
likely to seek bribes and firms should also be more willing to offer
bribes in exchange for private benefits (4).
The market for bribes therefore brings together demanders
(government officials) and suppliers (firms) of bribes. Since countries
differ in their corruption characteristics, firms will face varying
degrees of exposure to corruption depending on their country location.
Moreover, since firms differ in their firm-level characteristics, they
will be more or less vulnerable to corruption. We argue that both
exposure and vulnerability to corruption will affect the firm's
threat point, and thus affect the size of the bribe paid. Building on
the medical literature (Grzywacz et al. 2004), we define exposure as the
quantitative exogenous stressors or pressures that affect a firm;
whereas its vulnerability depends on its ability to withstand these
stressors/pressures. A firm's exposure to bribery therefore depends
on country-level characteristics such as the pervasiveness and
arbitrariness of corruption in an economy; whereas the firm's
vulnerability to bribery depends on its ability to withstand these
exogenous pressures. We start first with vulnerability and then address
Vulnerability to Corruption: Firm Characteristics and Bribery
Residual control theory suggests that government officials demand
fewer bribes from firms that have greater bargaining power (Hakala et
al. 2005; Svensson 2003). Svensson (2003), for example, argues that the
greater the mobility of capital and the higher the alternative return to
capital in other industries, the lower the firm's threat point and
the smaller the bribe. Bargaining power represents the firm's
ability to withstand the "grabbing hand" of external pressures
from government officials to pay bribes; the greater the firm's
bargaining power the less its vulnerability to corruption.
The MNE-state relationship literature hypothesizes that the
firm's bargaining power rises as its percentage of foreign
ownership increases, and is particularly strong at the time of first
entry (Eden et al. 2005; Vernon 1971). Foreign firms are more likely to
have alternative investment opportunities than local firms, a higher
propensity to exit, and thus a higher threat point (Kogut and Kulatilaka
1994). Compared to domestic firms, multinationals are also less embedded
in the host environment (Zaheer 1995). In addition, foreign firms,
inherently having disadvantages arising from liability of foreignness,
would be more likely to invest abroad when equipped with valuable
capabilities (Zaheer and Mosakowski 1997). Furthermore, foreign firms
from a different cultural background are less likely to know whom to
bribe and how much compared to their domestic counterparts (Rodriguez et
In other words, the higher bargaining power of MNEs provides them
with larger residual control rights vis-a-vis government officials,
leading to smaller bribes paid. For example, Herrera and Rodriguez
(2003) find that foreign firms bribe less than domestic firms, arguing
that the capabilities foreign owners bring to the host country imply
that less government assistance is needed. The International Bribe
Payers Index (Transparency International 2006) also shows that domestic
firms have a much higher tendency to bribe than their foreign
counterparts. For example, foreign owned firms (6.92 out of 10) have
experienced a 0.39 point lower incidence of bribery than locally owned
counterparts (6.53) in the top 10 countries of BPI 2006. If greater
foreign ownership is associated with an enhanced ability to "walk
away", in effect, the firm's threat point increases. Thus, we
Hypothesis 1: The higher the foreign ownership of the firm, the
lower its vulnerability to corruption and the smaller the bribes paid by
the firm to government officials.
In residual control theory, the stronger the residual rights of
control held by the firm, the greater the firms' bargaining power
relative to the government officials. When firms have stronger residual
rights of control compared to government officials, rather than taking
bribes, government officials may even provide support for the firms that
have stronger residual rights.
For example, national governments in developing countries and in
countries with balance of payments problems value exports highly for
their contributions to foreign exchange and employment (Grosse 1996;
UNCTAD 2006; Vernon 1971). This suggests that export oriented firms have
high national salience and that bureaucrats dampen their bribe demands
as a result to avoid punishment. In addition, the competition among
national governments to attract firms that export also gives export
oriented firms more bargaining power. Finns that are heavily involved in
exports typically receive govemment grants, rewarding them for
exporting. Since the late 1980s, governments have significantly
liberalized their export regulations; as a result, for most countries,
export licensing, permits and taxes are minimal (UNCTAD 2006). In
addition, heightened competition among nations makes it difficult for
the government officials to squeeze bribes from exporting firms. We
therefore hypothesize that export orientation is negatively related to
bribe payments; that is:
Hypothesis 2: The more export oriented is the firm, the lower its
vulnerability to corruption and the smaller the bribes paid by the firm
to government officials.
In residual control theory, the government has complete residual
rights of control when the firm is a state owned enterprise. In the case
of state ownership, however, the manager running the state owned firm is
part of the overall government apparatus and therefore shares similar
goals with other government officials (that is, the manager's goal
is to meet government objectives rather than maximizing firm profits).
Similarity of interests makes it easier to reach a common agreement
(Eden et al. 2005; Grosse 1996). Herrera and Rodriguez (2003) conjecture
that the frequency of bribes decreases if firms have effective recourse
through government channels to obtain proper treatment without making
We argue that government officials are less likely to demand
financial bribes from state-owned firms, relying instead on
feather-bedding activities such as requests to provide jobs for family
members. Such hiring is often possible since state-owned firms tend to
be larger than their private counterparts and face less pressure to
control costs (Boycko et al. 1996). In addition, many retired government
officials are later re-employed by stateowned firms (Krueger 1990). We
therefore expect influence-seeking and feather-bedding demands by state
officials to be more common than requests for financial bribes when the
firm is state owned.
In short, we expect that private sector firms are more likely to
pay higher bribes than state owned finns. In addition, bribes paid may
be when firms are privately owned since they are typically more
efficient and thus possess larger cash flows (Clarke and Xu 2004).
Thus, we argue;
Hypothesis 3: The higher the state ownership of the firm, the lower
its vulnerability to corruption and the smaller the bribes paid by the
firm to government officials.
Exposure to Corruption: Country Characteristics and Bribery
We also argue that, at the country level, the magnitude of the
perceived level and ambiguity of corruption in an economy will be an
important predictor of an individual firm's bribes (Martin et al.
2007). The overall level of corruption in a country determines the
firm's exposure level in the sense of the quantitative stressors
placed on an individual or firm by exogenous events. As Grzywacz et al.
(2004) argue stressors can be either discrete, specific
"on-off" events or chronic and enduring daily pressures.
We argue that public sector corruption can also be seen as discrete
or chronic pressures on firms. Corruption has two characteristics:
pervasiveness and arbitrariness (Rodriguez et al. 2005; Uhlenbruck et
al. 2006). Pervasiveness is conceptualized as "the average
firm's likelihood of encountering corruption in its normal
interactions with state officials", that is, "the proportion
of interactions with the state that will entail corrupt
transactions" (Rodriguez et al. 2005, p. 385). Pervasiveness of
corruption is similar to the frequency or incidence of corruption
relative to the firm's transactions with the state. Higher
pervasiveness implies that a higher proportion of the firm's
activities with government officials will involve corrupt behaviors.
Arbitrariness, on the other hand, refers to the unpredictability or
variability of corruption, more specifically, "the inherent degree
of ambiguity associated with corrupt transactions". A high degree
of arbitrariness implies that "transactions with government
officials are characterized by an enduring uncertainty regarding the
size, target, and number of corrupt payments necessary to obtain an
approval" (Rodriguez et al. 2005, p. 385). Both characteristics
affect firms' exposure to corruption.
Pervasiveness of Corruption
First, at the country level, the pervasiveness of public sector
corruption is likely to affect each official's assessment of the
benefits and costs of demanding bribes (Doh et al. 2003). Residual
control theory suggests that the government official's demand for
bribes depends on his/her net marginal valuation of the received bribe.
This includes the official's assessment of the probability of being
caught and punished for accepting a bribe, and the expected size of the
punishment. This assessment should vary with the pervasiveness of
corruption at the country level. For example, President Suharto of
Indonesia was often referred to as "Mr. Ten Percent" because
it was widely understood that paying 10% of the deal to the government
would secure the business in Indonesia (Wei 2000; Fisman 2001). Thus,
given the high pervasiveness of corruption in Indonesia, government
officials would assess their likelihood of being caught and punished for
seeking bribes as low.
Murphy et al. (1993, p.409) argue that an increase in corrupt
activities in a country makes corrupt behaviors more attractive; as the
"strength in numbers" speaks, "the probability of any one
... getting caught is much lower" when more people are stealing.
When most government officials ask for bribes, it is less risky for
another government official to do the same (Blackburn et al. 2004;
Rose-Ackerman 1975); moreover, where bribery is prevalent, the risk
involved in non-compliance increases (Drabek and Payne 2001). This is
well put by Mauro (1998): In a country where everybody steals the
probability of your being caught for stealing too is low and, even if
you are caught, the probability of severe punishment is also low; thus,
you steal, too.
Moreover, when pressures for bribes are repeated and chronic, they
become an "additive and cumulative toll of daily hassles"
(Grzywacz et al. 2004, p. 3) with a stronger impact than the sum of the
individual bribe requests would suggest. Therefore, we argue that the
more pervasively corrupt the country environment, the more it becomes
acceptable for government officials to demand bribes. Therefore, we
Hypothesis 4: The more pervasive is corruption in an environment,
the greater the firm's exposure to corruption and the larger the
bribes paid by the firm to government officials.
Arbitrariness of Corruption
The greater the arbitrariness of corruption, the less predictable
it becomes. Firms do not know when to expect bribery demands, or from
whom, or what size, or if the firm does pay a bribe whether the
government official will deliver the promised service. Arbitrariness
complicates the predictability and planning of firms' bribery and
thus can make bribing more damaging (Rodriguez et al. 2005). In a
situation where there are no norms, we argue that high arbitrariness
should be seen by the firm as less of a threat (reduced exposure) and
should dampen a firm's willingness to pay a bribe.
Residual control theory suggests that firms will only assume the
risk of paying bribes when the rewards are adequate and predictable
(Kauffman et al. 1999). When unpredictability is associated with
corruption, the potential varied interpretation and distortion of
government policies by each government official may make bribery
ineffective from the firm's perspective, thus lessen the bargaining
power of the government officials in underthe-table deals (Levy 1989;
If the probability of gaining preferential treatment in exchange
for bribery is unclear, or if government officials come back and demand
more bribes than originally agreed upon (Klitgaard 1990), corruption is
seen as arbitrary in nature (Rodriguez et al. 2005). When arbitrariness
is high, it is costly for firms to distinguish between government
officials who claim to have, and those who do have, residual rights of
control over the firm (Campos and Lien 1999).
In a sense, when corruption arbitrariness is high, the external
environment is perceived by the firm as an "an ungoverned
space" that the firm must navigate. This is why Martin et al.
(2007) use anomie theory to explain bribery activity of the firms,
defining anomie as "a condition of normlessness and social
disequilibrium where the rules once governing conduct have lost their
savor and force" (Merton 1964, p. 226). Vaaler and Schrage (2009)
also find that firms are less able to cope with the external environment
when the policy environment is unstable. We therefore see high
arbitrariness as equivalent to high opacity of exposure; firms cannot
determine the degree of corruption exposure facing them in a particular
country, industry or activity. As a result, firms may misperceive or
underestimate their exposure to corruption.
Given the increased unmeasurable uncertainty of high arbitrariness,
firms are reluctant to bribe government officials, which in turn lowers
their residual control rights vis-a-vis firms (Doh et al. 2003).
Consequently, as corruption becomes more arbitrary, firms should pay
Hypothesis 5: The more arbitrary the corruption is in an
environment, the more difficult it is for the firm to determine its
exposure to corruption and therefore the smaller the bribes paid by the
firm to government officials.
Methodology and Results
Dataset and Variables
To test our hypotheses, we build a dataset using three World Bank
datasets: The World Business Environment Survey (WBES), the World
Development Indicators, and the Governance Indicators dataset. WBES
dataset contains unique firm level survey data, covering more than
10,000 firms in 81 countries in 2000 (Batra et al. 2003) and has been
used in the past research (i.e., Uhlenbruck et al. 2006). Due to bribery
being a sensitive subject in many countries, the WBES suffers from
missing values especially for from Africa and Middle East regions. Thus,
our final data set consists of firm respondents from five broad regions:
Transition Europe, East Asia, South Asia, Latin America, and the OECD.
Our final data include 61 countries and consist of 5,215 observations.
After screening for sample selection bias (see below), we have 3,119
observations for our analysis.
Our dependent variable is the total amount of bribes paid annually
by a firm to all government officials, measured as a percentage of the
firm's total annual sales. Our proxy is based on the WBES survey
question: On average, what percentage of revenues do firms like yours
typically pay per annum in unofficial payments to public officials? The
score ranges from 1 to 7, which corresponds to a range of 0 to more than
Our independent variables are at two levels: Firm and country. At
the firm level, Foreign Ownership is operationalized as the percentage
of foreign shares in the total ownership of a firm. WBES data, on
average, contains about 15% of firms with some degree of foreign
ownership. Following previous research (Shaked 1986), we measure Export
Orientation as the ratio of a firm's export sales to its total
sales. Past research finds that export orientation is associated with
corruption (Ades and Di Tella 1999). State Ownership is the percentage
of governmental shares in the total ownership of a firm. WBES data, on
average, contains around 13% of firms in which government has some share
of firms' ownership. State ownership has been widely used in past
research on corruption (Hellman et al. 2003; Milovanovic 2002; Shleifer
and Vishny 1993).
At the country level, we have two variables: Pervasiveness and
Arbitrariness of corruption. These variables denote the country average
of the individual firms' perceptions of the pervasiveness and
arbitrariness of corruption in that country. Both variables are
constructed using the same WBES questions and methodology developed in
Uhlenbruck et al. (2006). Our results show that the two latent
variables, pervasiveness and arbitrariness, are independent of each
other and can be used for reflecting two idiosyncratic features of
We control for various country specific and region specific factors
that might influence firms' tendencies to engage in corruption.
First, we include the logged value of Gross Domestic Product (GDP) and
GDP Growth (Habib and Zurawicki 2002; Wei 2000). Both variables come
from the World Development Indicators: GDP for 2000 and the average
annual GDP growth rate for 1996-2000. Regional dummy variables are used
to control other differences among countries. Our region dummies are
Transition Europe, East Asia, South Asia, Latin America and OECD, where
OECD is the referent (see Table I for a detail country list).
We also include two country-level variables related to export,
export promotion and export taxes, which may be compounded with the
effect of export orientation on a firm's size of bribes. Export
promotion may induce a heightened competition for such funds which may
raise the level of bribery. Also, the extent of export taxes may affect
the size of a firm's bribery. Export promotion is measured by the
difficulty of accessing to specialized export finance and Export taxes
capture taxes on exports as a% of total tax revenue.
The WBES survey may suffer either a non-response or an
under-response bias related to country-level political conditions,
whereby firms in countries with little political freedom either do not
respond or underestimate their bribes paid. Vaaler and McNamara (2004)
use an annual average of country political and civil rights from Freedom
House to proxy for the level of political freedom in a country, as a way
to correct for these biases. A lower value means that the people in a
country enjoy more political freedom. Controlling for Political
Rights' may therefore also help correct for any systematic
non-response or under-response bias by country.
Industry dummies are included to correct for any industry level
differences in bribing. Herrera and Rodriguez (2003) show that
manufacturing firms are less prone to bribe than service firms. Four
categories of industries are used: Manufacturing, Service, Construction,
and Agriculture, with Agriculture as the referent. We also include the
number of industry competitors (Competition) as a control variable.
Existing theory argues that political and economic competition reduces
corruption levels; whereas monopolistic markets produce high levels of
corruption (Ades and Di Tella 1999; Shleifer and Vishny 1993).
Lastly, we employ two firm level variables as control variables:
Firm size and firm age. Firm Size is measured by the number of employees
of a firm. This measure is recoded as small (5-50 employees), medium
(51-500 employees), and large (larger than 500 employees) firms. Firm
Age is measured by duration year since foundation (5).
Due to the secret and illegal nature of bribery, it is perhaps not
surprising that the WBES dataset has missing values in its variables
that measure corruption. We therefore impute missing values with new
values using the multiple imputation procedure "ice" in STATA
(Newman 2003; Royston 2005) (6). Multiple imputation has been widely
used in management research (e.g., Glomb and Liao 2003; Katila 2002;
Mitchell 1994; Spell and Blum 2005) (7).
Descriptive statistics and correlation coefficients are presented
in Table 2. Variance inflation factors (VIF) indicate no potential
multicollinearity problems in our data (Chatterjee and Price 1991; Neter
et al. 1996). To correct for any possible heteroscedasticity, we also
use White-corrected (robust) standard errors, with clustering of
identity groups by country.
An analysis of bribe size, excluding observations where no bribe
was paid, might create an endogeneity problem and bias our results
(Shaver 1998). To avoid sample selection bias, we conduct a two stage
analysis: In the first stage we predict whether firms bribe or not, and
in the second stage we predict bribe size where the second stage drops
firms that do not bribe. We follow Heckman's two stage procedure
(Greene 2003, p. 784; Heckman 1979; Sartori 2003) where the first stage
selection model predicting the probability of bribery is estimated by a
probit model. We calculate the inverse Mills ratio (IMR) from the
selection equation and include the inverse Mills ratio in our equation
estimating bribe size. Following Sartori (2003), we include Property
Rights Protection (IPR) only in the first stage selection equation (8).
Table 3 provides the results of our first stage probit model and maximum
The values shown in each block are the unstandardized regression
coefficients B. Standard errors are in parentheses. Dependent variable
for Probit is the incidence of bribery. Dependent variable for Maximum
Likelihood is the amount of bribery money. For 'industry'
control, omitted category (reference) is agriculture. For
'region' control, omitted category (reference) is OECD As
shown in Table 3, Export Orientation, Foreign Ownership, and State
Ownership are significant and negative. This shows that firms with high
level of foreign ownership and state ownership and more export oriented
firms are less likely to bribe. When we examine the two dimensions of
corruption, Pervasiveness and Arbitrariness, high pervasiveness is
associated with high likelihood of bribery while high arbitrariness is
associated with lower likelihood of bribery.
Several control variables are also worthwhile to examine. The
coefficient for GDP(log) is significant and negative, suggesting the
less affluent a country, the more likely firms are to bribe. The
Political Rights variable is positively associated with the likelihood
of bribery; thus firms in countries with stronger political rights are
less likely to pay bribes, as expected. The results also show that firms
in Manufacturing and Construction are more likely to bribe than the base
case Agriculture. Firm Size and Firm Age are also negatively associated
with the likelihood of bribery. Further, the higher are Export Taxes, a
firm is more likely to bribe. Export Promotion (lower values imply
higher export promotion) is positively related to the likelihood of
bribery. In addition, Competition is positively associated with the
likelihood of bribery and significant. Property Rights Protection (lower
values imply higher property rights) has a positive relationship with
the likelihood of bribery.
Table 4 presents the results of our hierarchical regression
analyses of bribe size. Note that these are stage two regressions using
Heckman's (1979) two stage model so they include only those firms
that paid a certain percentage of sales as bribes to government
officials. Model 1 is the baseline regression for control variables at
the country, region, industry and firm levels. In model 2, we add our
firm level independent variables. Model 3 shows the full model, adding
our country level independent variables.
Our results indicate that all models are statistically significant
at the 0.001 level (using a more conservative two-tailed t-test for
statistical significance). The adjusted R squared values range from
0.116 to 0.15, which is similar to previous research on corruption
(Habib and Zurawicki 2002). The Inverse Mills Ratio is negative and
statistically significant at the 0.00l level. The overall fit of the
models improves as we add the firm level and country level independent
variables, as shown by the change in F statistics.
As predicted by Hypothesis 1, the relationship between Foreign
Ownership and Bribe Size is negative and statistically significant
(p<0.00l); that is, for firms that pay bribes, higher levels of
foreign ownership are negatively correlated with bribe size. Foreign
firms have a greater ability to "walk away" and thus show a
higher threat point compared to domestic firms. They are therefore
better able to resist bribe demands from government officials.
Hypothesis 2 predicts that the more export oriented the firm, the
smaller will be bribe size because exporting is a high valued activity
that increases the firm's bargaining power. Table4 shows that the
relationship between Export Orientation and Bribe Size is negative and
statistically significant at the 0.01 level; of firms that pay bribes,
more export oriented firms pay lower bribes on average. Thus, we find a
strong support for Hypothesis 2. Hypothesis 3 predicts a negative
relationship between State Ownership and Bribe Size; the relationship is
negative and significant (p< 0.001) so Hypothesis 3 is supported. The
greater the state ownership of the firm, the smaller the bribe size, for
those firms that pay bribes. Hypotheses 4 and 5 are also supported in
that Pervasiveness of corruption is associated with more bribes, but
Arbitrariness of corruption is associated with fewer bribes paid by the
Our control variables warrant some attention as well. GDP(log) and
GDP Growth are negatively related to bribes paid as past research finds
(Robertson and Watson 2004; Husted 1999); wealthy and growing countries
suffer less from bribery. Political Rights (a reversed measure) is
positively and significantly related to bribes paid in the full
regression model; that is, as expected, stronger rights are negatively
related to bribery. The regional dummy variables are generally not
statistically significant. Of the industry dummy variables, only
Manufacturing is marginally significant and negative, suggesting that
average bribe size is less in manufacturing than in the base case
industry Agriculture. Competition was negative but not statistically
significant. Firm Size and Firm Age are negatively related to Bribe
Size. Overall, these results suggest that, of those firms that pay
bribes, bigger and older firms have a weak tendency to pay more bribes.
Export Promotion (a reversed measure) has a positive relationship with
the size of bribes, suggesting that when access to export finance is
limited, average bribe size rises for those firms that pay bribes.
Higher Export Taxes' are also more likely to induce an increase in
bribe size (10).
Discussion and Conclusion
Management researchers have paid relatively less attention to the
issue of corruption from the firm's perspective. Our study attempts
to fill this void by developing a residual control theory of bribery
that incorporates insights from the management literature. One of the
contributions of our paper is to distinguish between a firm's
exposure and its vulnerability to corruption.
Firms are more or less vulnerable to corruption depending on their
ability to withstand government officials' demands for bribes. We
argue that three firm-level characteristics affect vulnerability, two of
which reflect a firm's international orientation. First, we found
that higher foreign ownership leads to smaller bribe payments to
government officials. Second, we also found a weakly negative
relationship (p < 0.10) between export orientation and bribe
payments. This suggests that greater international orientation of the
firm, whether through foreign ownership or export orientation, is
associated with lower bribes paid. This might also suggest that more
internationally oriented firms have higher resources and capabilities,
and therefore less need for government assistance. Past research also
shows that more internationally oriented firms learn by engaging in
international activities (Salomon and Shaver 2005). This potential
learning advantage can be an incentive to go international for domestic
firms managing in a highly corrupt environment.
The third firm level characteristic affecting vulnerability is
state ownership. We found that higher state ownership is associated with
lower bribe size. The reason for this negative relationship comes not
from the firm's having more bargaining power and a greater ability
to "walk away", but rather from shared relationships and other
alternative ways for government officials to "pluck the goose"
such as featherbedding and inflated costs. In other words, when there is
less conflict over the residual control rights in a firm due to high
government ownership, firms might be able to satisfy government
officials' demands in the ways other than paying bribes. When there
are few agency conflicts, government officials can extract rents from
firms without getting directly asking for bribes. On the other hand,
when extracting rents from the firms is harder, the bargaining power
relationship becomes important and firms need higher bargaining power to
bribe less. This might also explain why government firms are less
efficient. Given that the need to be more competitive (which can
increase the bargaining power of firms) is lower when residual control
is held by government officials, it is not surprising that government
owned firms are less efficient (Meyer and Zucker 1989).
Firms are more or less exposed to corruption depending on a
country's corruption characteristics. We found strong support for
our hypotheses that the two characteristics of corruption, pervasiveness
and arbitrariness, present sharply differential effects on bribe size.
Pervasiveness has a strong, positive impact on the average size of bribe
paid, while arbitrariness negatively affects the average bribe. A highly
pervasive environment exposures firms to strong corruption pressures,
which are difficult for firms to resist. Thus, the probability of bribes
being paid and the average bribe size both rise. These findings support
the arguments made by Shleifer and Vishny (1993) in that organized, more
predictable corruption regimes are likely to extract more bribes than
unorganized ones. On the other hand, when corruption is highly arbitrary
and unpredictable, who, what, when and how much to bribe is unclear;
moreover, firms lack surety that government officials will deliver on
their promises. When corruption pressures are unclear, firms are more
reluctant to pay bribes. Firms perceive lower levels of corruption
exposure due to its opacity and unpredictability, and average bribe size
Our study has managerial implications, hinging on how firms
interact with government agents in paying bribes. The stronger the
residual rights held by government officials, the greater the
officials' ability to demand bribes. Understanding this notion may
help executives decide whether and how the firm should secure its
residual rights. For example, in terms of ownership, a firm may decide
to involve foreign owners in order to increase its bargaining power
relative to government officials. Involvement of the government through
state ownership may also protect the firm from bribe demands.
Understanding where is the firm's threat point--when can and should
the firm walk away--is critical for managers in determining whether and
how much they should pay in bribes. Moreover, the overall corruption
environment, both in terms of level and uncertainty, has implications
for the firm's managers. When pervasiveness is high, firms are
likely to imitate their competitors and pay larger bribes. High
arbitrariness suggests, however, that firms may reduce bribe payments.
Our study suggests that corruption is both a country level and an
industry level phenomenon. Firms' exposure to corruption varies
across industries, as evidenced in our empirical work where we find that
firms in manufacturing are less likely to pay and pay lower bribes to
government officials, whereas firms in construction are more likely to
pay bribes, compared to agriculture. Our results support Herrera and
Rodriguez (2003)'s finding that manufacturing firms are less prone
to bribe than service firms. We also find it not surprising that the
construction industry has a Global Infrastructure Anti Corruption Centre
specifically geared to lessening bribery in the construction industry
Our work suggests that future scholars should look into highly
corrupt nations as a special case. The reason is that in highly corrupt
countries, it is possible that firms that do not bribe might be quite
different from the rest of the firms. In addition, while we find that
high arbitrariness leads to reduced bribes paid to government officials,
it is very possible that when the level of arbitrariness is really high,
firms might have to bribe all the stakeholders, which would make the
firms bribe more, not less. In other words, as stakeholder theory
suggests (Mitchell and Agle 1997), firms might have to bribe all
interested parties given that it would be very costly to find out whom
not to bribe, especially when arbitrariness is very high (11). In
addition, we do not look into a specific country for corruption. Future
study may, however, examine the details of corruption in a specific
country. For example, the United States has a Foreign Corruption
Practices Act (1977) that prohibits firms bribing in foreign countries.
As Rodriguez et al. (2005) argue, US firms may bribe less due to this
formal institutional arrangements.
Our research also suffers from limitations. Data collected from
surveys are prone to problems such as misreporting and missing values.
This especially true when it comes to corruption data, given the secrecy
attached to bribing. This is why we dealt with missing values using the
multiple imputation method. Also, our study is cross-sectional; a panel
study that incorporated bribery levels over time would be an important
addition that would help separate cause from effect when it comes to
patterns of corruption. Future studies should look at the longitudinal
aspects of endogenous corruption. Our study also focuses only on foreign
ownership and export orientation; other forms of international
involvement such as import penetration could also be investigated. (12)
Lastly, our finding that wealthy and growing countries suffer less from
bribery shows the two constructs are correlated, not the direction of
causation. Our paper does not address the issue of whether causation
runs from wealth to corruption or the reverse.
In conclusion, our paper was designed to answer the question: Why
do some firms pay more bribes than others? We found that residual
control theory, supplemented by insights from the international
management literature, offers a useful theoretical lens for analyzing
firms' exposure and vulnerability to corruption.
Ades, A., & Di Tella, D. (1999). Rents, competition, and
corruption. American Economic Review, 89(4), 982-993.
Allison, P. (2002). Missing data. Sage: Thousand Oaks.
Batra, G., Kaufmann, D., & Stone, A. (2003). Investment climate
around the world." Voices of the firms from the worm business
environment survey. Washington: The World Bank.
Blackburn, K., Bose, N., & Haque, E. (2004). Endogenous
corruption in economic development, Research paper 2004-16. UK:
Boycko, M., Shleifer, A., & Vishny, R. (1996). A theory of
privatization. Economic Journal, 106(4), 309-319.
Brownstone, D., & Valletta, R. (2001). The bootstrap and
multiple imputations Harnessing increased computing power for improved
statistical tests. Journal of Economic Perspectives, 15(4), 129-141.
Campos, J., & Lien, D. (1999). The impact of corruption on
investment predictability matters. World Development, 27(6), 1059-1067.
Chatterjee, S., & Price, B. (1991). Regression analysis by
example. New York: Wiley.
Clarke, G., & Xu, L. (2004). Privatization, competition, and
corruption: How characteristics of bribe takers and payers affect bribes
to utilities. Journal of Public Economics, 88(9-10), 2067-2097.
Doh, J., Rodriguez, P., Uhlenbruck, K., Collins, J., & Eden, L.
(2003). Coping with corruption in foreign markets. Academy of Management
Executive, 17(3), 114-127.
Donaldson, T., & Preston, L. (1995). The stakeholder theory of
the corporation: concepts, evidence, and implications. Academy of
Management Review, 20(1), 65-91.
Drabek, Z., & Payne, W. (2001). The impact of transparency on
foreign direct investment. Working paper ERAD-99-02. Geneva: World Trade
Economist. (1993, September 11). The future of capitalism, pp.
Eden, L., Lenway, S., & Schuler, D. (2005). From the
obsolescing bargain to the political bargaining model. In R. Grosse
(Ed.) International business-government relations in the 21st century.
Cambridge: Cambridge University Press.
Fisman, R. (2001). Estimating the value of political connections.
American Economic Review, 91(4), 1095-1102.
Foss, K., & Foss, N. (1999). Understanding ownership: Residual
rights of control and appropriable control rights'. Working paper.
DRUID Working Paper No. 99-4.
Glomb, T., & Liao, H. (2003). Interpersonal aggression in work
groups social influence, reciprocal, and individual effects. Academy of
Management Journal, 46(4), 486-496.
Greene, W. (2003). Econometric analysis. Upper Saddle River, NJ:
Grosse, R. (1996). The bargaining relationship between foreign MNEs
and host governments in Latin America. International Trade Journal,
Grossman, S., & Hart, H. (1986). The costs and benefits of
ownership: A theory of lateral and vertical integration. Journal of
Political Economy, 94(4), 691-719.
Grzywacz, J., Almeida, D., Neupert, S., & Ettner, S. (2004).
Socioeconomic status and health: A micro-level analysis of exposure and
vulnerability to daily stressors. Journal of Health and Social Behavior,
Habib, M., & Zurawicki, L. (2002). Corruption and foreign
direct investment. Journal of International Business Studies, 33(2),
Hakala, K., Norback, P., & Svaleryd, H. (2005). Asymmetric
effects of corruption on FDI Evidence firm Swedish multinational firms.
Working Paper No. 641.
Hart, O., & Moore, J. (1990). Property rights and the nature of
the firm. Journal of Political Economics, 98(6), 1119-1158.
Heckman, J. (1979). Sample selection bias as a specification error.
Econometrica, 47(1), 153-161.
Hellman, J., Jones, G., & Kaufmann, D. (2003). Seize the state,
seize the day: State capture and influence in transition economies.
Journal of Comparative Economics, 31(4), 751-773.
Herrera, A., & Rodriguez, P. (2003). Bribery and the nature of
corruption. Working paper. Darden School of Business, University of
Husted B. (1999). Wealth, culture, and corruption. Journal of
International Business Studies, 30(2), 339-360.
Johnson, S., Kaufmann, D., & Zoido-Lobaton, P. (1998).
Regulatory discretion and the unofficial economy. American Economic
Review, 88(2), 387-392.
Katila, R. (2002). New product search over time past ideas in their
prime? Academy of Management Journal, 45(5), 995-1010.
Kauffman, D., Pradhan, S., & Ryterman, R. (1999). World Bank
finds new ways to diagnose corruption symptoms. Transition, 10(1),
King, G., Honaker, J., Joseph, A., & Scheve, K. (1999).
Analyzing incomplete political science data: An alternative algorithm
for multiple imputation. Cambridge: Harvard University, Mimeo.
Klitgaard, R. (1990). Tropical gangsters. New York: Basic Books.
Kogut, B., & Kulatilaka, N. (1994). Operating flexibility,
global manufacturing, and the option value of a multinational network.
Management Science, 40(1), 123-139.
Knaeger, A. (1990). Government failures in development. Journal of
Economic Perspectives, 4(3), 9-23.
Levy, P. (1989). The Waterfront Commission of the Port of New York:
A history and appraisal. Industrial Labor Relations Review, 42(4),
Martin, A. (1999). The development of international bribery law.
Natural Resources Environment, 14(2), 1-13.
Martin, K., Cullen, K., Johnson, J., & Parboteeah, K. (2007).
Deciding to bribe: Across-level analysis of firm and home country
influences on bribery activity. Academy of Management Journal, 50(6),
Mauro, P. (1998). Corruption causes, consequences, and agenda for
further research. Finance Development, 35(1), 11-14.
Merton, R. (1964). Anomie, anomia, and social interaction. In M.
Clinard (Ed.), Anomie and deviant behavior. New York: Free Press.
Meyer, M., & Zucker, L. (1989). Permanently failing
organizations. Newbury Park: Sage.
Milovanovic, M. (2002). Endogenous corruption in privatized
companies. Working paper.
Mitchell, W. (1994). The dynamics of evolving markets The effects
of business sales and age on dissolutions and divestitures.
Administrative Science Quarterly, 39(4), 575-602.
Mitchell, R., & Agle, B. (1997) Toward a theory of stakeholder
identification and salience: Defining the principle of who and what
really counts. Academy of Management Review, 22(4), 853-886.
Murphy, K., Shleifer, A., & Vishny, R. (1993). Why is
rent-seeking so costly to growth? AEA Papers and Proceedings, 83(2),
Neter, J., Kutner, M., Nachtsheim, C., & Wasserman, W. (1996).
Applied linear statistical models (4th ed.). Chicago: Irwin.
Newman, D. (2003). Longitudinal modeling with randomly and
systematically missing data: A simulation of ad hoc, maximum likelihood,
and multiple imputation techniques. Organizational Research Methods,
Oldenburg, P. (1987). Middlemen in third-world corruption. World
Politics, 39(4), 508-535.
Robertson, C., & Watson, A. (2004) Corruption and change: The
impact of foreign direct investment. Strategic Management Journal,
Rodriguez, R, Uhlenbruck, K., & Eden, L. (2005). Government
corruption and the entry strategies of multinationals. Academy of
Management Review, 30(2), 383-396.
Rose-Ackerman, S. (1975). Economics of corruption. Journal of
Public Economics, 4(1), 187-203.
Royston, P. (2005). Multiple imputation of missing value update.
The Stata Journal, 5(2), 1-14.
Rubin, D. (1987). Multiple imputation for nonresponse in surveys.
New York: Wiley.
Salomon, R., & Shaver, J. (2005). Learning by exporting: New
insights from examining firm innovation. Journal of Economics &
Management Strategy, 14(2), 431-460.
Sartori, A. (2003). An estimator for some binary-outcome selection
models without exclusion restrictions. Political Analysis, 11(2),
Shaked, I. (1986). Are multinational corporations safer? Journal of
International Business Studies, 17(1), 83-106.
Shaver, M. (1998). Accounting for endogeneity when assessing
strategy performance: Does entry mode choice affect FDI survival?
Management Science, 44(4), 571-585.
Shleifer, A., & Vishny, R. (1993). Corruption. Quarterly
Journal of Economics, 108(3), 599-617.
Siegel, S., & Castellan, N. (1988). Nonparametric statistics
for the behavioral sciences (2nd ed.). New York: McGraw-Hill.
Spell, C., & Blum, T. (2005). Adoption of workplace substance
abuse prevention programs: Strategic choice and institutional
perspective. Academy of Management Journal, 48(6), 1125-1142.
Svensson, J. (2003). Who must pay bribes and how much? Evidence
from a cross section of firms. Quarterly Journal of Economics, 111,
Tirole, J. (1999). Incomplete contracts: Where do we stand?
Econometrica, 67(4), 741-781.
Transparency International. (2006). Bribe payers index 2006:
Uhlenbruck, K., Rodriguez, P., Doh, J., & Eden, L. (2006). The
impact of corruption on entry strategy: Evidence from telecommunication
projects in emerging economies. Organization Science, 17(3), 402-414.
UNCTAD, World Investment Report 2006. (2006). Transnational
corporations and the internationalization of R&D. New York: United
Vaaler, P., & McNamara, G. (2004). Crisis and competition in
expert organizational decision making: credit-rating agencies and their
response to turbulence in emerging economies. Organization Science,
Vaaler, P., & Schrage, B. (2009). Residual ownership, policy
stability and financial performance following strategic decisions by
privatizing telecoms. Journal of International Business Studies, 40,
Vernon, R. (1971). Sovereignty at Bay. The multinational spread of
US enterprises. New York: Basic Books.
Wei, S.-J. (2000). How taxing is corruption on international
investors? The Review of Economic Statistics, 82(1), 1-11.
Westphal, J. (1999). Collaboration in the boardroom behavioral and
performance consequences of CEO-board social ties. Academy of Management
Journal, 42(1), 7-24.
Zaheer, S. (1995). Overcoming the liability of foreignness. Academy
of Management Journal, 38(2), 341-363.
Zaheer, S., & Mosakowski, E. (1997). The dynamics of the
liability of foreignness: A global study of survival in financial
services. Strategic Management Journal, 18(6), 439-464.
(1) We do not include legal ways of affecting government officials
such as facilitating payments specified in the Foreign Corruption
Practices Act (1997) in the United States.
(2) We are indebted to a reviewer who asked us to consider the
difference between exposure and vulnerability.
(3) In this study, we mainly focus on home country government
officials. One potential exception is when we consider the level of
foreign ownership. From the foreign partner finn's standpoint the
government can be viewed as the host country government.
(4) At the same time, we recognize that firms with little
bargaining power may be more likely to voluntarily pay bribes,
particularly if they see the reciprocal private benefits as essential to
doing business in that country or industry. Voluntary bribes, in this
case, become "good faith" payments that help build long-term
relationships with government officials.
(5) These two variables, age and size, are separate and do not load
into one variable using confirmatory factor analysis. Thus, they may
capture distinct aspects of firm visibility respectively.
(6) The multiple imputation procedure (Rubin 1987) replaces each
missing value with a set of plausible values, instead of filling in a
single value for each missing value that represents the uncertainty
about the right value to impute. We imputed missing values, around 30%
of sample, with new values by multiple imputation (MI) procedures
(Newman 2003; Royston 2005). Extant research shows that the multiple
imputation estimator is not only more efficient with a smaller standard
error but also larger in the magnitude of the effect parameter compared
to other substitutions such as list wise deletion, linear interpolation
or a single imputed value (Allison 2002; Brownstone and Valletta 2001;
King et al. 1999; Newman 2003).
(7) To see if multiple imputation is effective, we compare the
equality of distribution function of the bribe money in the final sample
with that of the original sample, using the two sample
Kolmogorov-Smirnov test (Siegel and Castellan 1988; Westphal 1999). The
results show that the two samples are not different from each other
(8) Sartori (2003) shows that if two equations have the same
variables and the variables have substantially the same influence on
selection and second-stage dependent variable, then the Heckman
procedure faces a problem of having to estimate the effect of the
variables and functions of the same variables on the dependent variable.
The recommended correction is the exclusion restriction; that is, add
another more meaningful variable in the first stage selection equation
that is not included in the second stage equation.
(9) In order to determine whether the relationships between bribery
and the organizational and country characteristics were robust to the
type of country, we repeated our econometric analyses (not shown) by
splitting our firms in subsamples of more and less corrupt countries,
based on the CPI provided by Transparency International. The results
show that the more corrupt subsample has more consistent and significant
effects in the main relationships than the less corrupt subsample,
providing additional support for our hypotheses. Results are available
from the authors on request.
(10) We also used export cost based on the Doing Business Survey
data (2005) as an alternative measure of the export tax as a robustness
check and we do not find any qualitative differences of the results.
(11) For example, in Somalia, anybody can shoot a person, but only
a few (e.g. doctors and nurses) can save a person's life.
Especially given that who will gain is uncertain, it would be important
to bribe all stakeholders in such an environment.
(12) The WBES dataset does not have data on firm imports; as a
result, we could not investigate the relationship between import
intensity and bribe size.
Received: 25.08.2008 / Revised: 11.02.2009 / Accepted: 27.05.2009 /
Published online: 16.11.2010 [c] Gabler-Verlag 2010
Assoc. Prof. S.-H. Lee ([mail])
School of Management, University of Texas at Dallas, Richardson,
e-mail: lee. firstname.lastname@example.org
Asst. Prof. K. Oh
College of Business Administration, University of Missouri-St.
Louis, St. Louis, USA
Prof. L. Eden
Department of Management, Mays Business School, Texas A&M
University, College Station, USA
Table 1: Country list and sample size (3,119)
Region Countries Sample size
Transition Armenia 105
Europe Azerbaijan 83
Czech Rep 92
East Asia China 75
South Asia Pakistan 90
Latin Bolivia 69
America Colombia 67
Costa Rica 65
El Salvador 77
OECD United Kingdom 74
Region Countries Sample size
Transition Lithuania 48
Europe Moldova 70
East Asia Singapore 89
South Asia India 164
Latin Nicaragua 78
America Panama 78
Trinidad & 98
OECD Italy 51
United States 76
Table 2: Descriptive statistics and Pearson correlation coefficients
Mean SD 1 2 3 4
1. Bribe size 3.25 1.31 1.00
2. GDP (Log) 10.58 0.83 -0.08 1.00
3. GDP growth 5.16 2.78 -0.03 0.11 1.00
4. Political right 3.49 1.58 0.09 -0.18 0.39 1.00
5. Trans-Europe 0.61 0.49 0.00 -0.33 0.30 0.18
6. East Asia 0.06 0.24 0.03 0.19 0.10 0.07
7. South Asia 0.07 0.25 0.05 0.22 -0.09 0.07
8. Latin America 0.21 0.41 -0.01 -0.07 -0.30 -0.15
9. OECD 0.04 0.20 -0.05 0.43 -0.13 -0.29
10. Manufacturing 0.38 0.48 -0.08 0.06 -0.06 -0.06
11. Service 0.40 0.49 0.07 -0.01 -0.01 -0.02
12. Agriculture 0.14 0.35 0.00 -0.10 0.03 0.09
13. Construction 0.08 0.28 0.02 0.05 0.08 0.04
14. Competition 2.44 0.71 0.05 -0.06 0.19 0.08
15. Firm size 1.74 0.70 -0.16 0.15 -0.08 -0.02
16. Firm age 16.89 20.15 -0.13 0.15 -0.15 -0.11
17. Export 2.13 1.24 0.08 -0.07 -0.01 0.03
18. Export tax 28.38 73.27 0.01 0.00 -0.03 -0.03
19. Foreign 44.87 38.99 -0.19 0.04 -0.07 -0.05
20. Export 30.90 34.26 -0.10 -0.05 -0.09 -0.01
21. State ownership 35.30 40.52 -0.20 -0.08 -0.03 -0.05
22. Pervasiveness 0.14 1.30 0.13 0.10 -0.04 0.01
23. Arbitrariness -0.36 0.84 -0.22 0.05 0.01 -0.08
5 6 7 8 9 10
1. Bribe size
2. GDP (Log)
3. GDP growth
4. Political right
5. Trans-Europe 1.00
6. East Asia -0.33 1.00
7. South Asia -0.34 -0.07 1.00
8. Latin America -0.66 -0.14 -0.14 1.00
9. OECD -0.27 -0.06 -0.06 -0.11 1.00
10. Manufacturing -0.11 0.02 0.17 0.02 -0.01 1.00
11. Service 0.08 0.07 -0.11 -0.10 0.05 -0.63
12. Agriculture -0.01 -0.09 -0.08 0.16 -0.07 -0.32
13. Construction 0.07 -0.04 -0.01 -0.06 0.02 -0.23
14. Competition 0.46 -0.07 0.07 -0.45 -0.20 -0.10
15. Firm size -0.21 0.04 0.12 0.13 0.06 0.23
16. Firm age -0.23 0.00 0.04 0.16 0.17 0.12
17. Export 0.04 -0.02 0.01 -0.01 -0.07 0.13
18. Export tax -0.08 -0.06 -0.05 -0.01 0.35 0.03
19. Foreign -0.08 -0.01 -0.03 0.09 0.06 0.07
20. Export -0.10 0.03 0.03 0.08 -0.01 0.13
21. State ownership -0.05 -0.05 -0.04 0.11 0.01 0.09
22. Pervasiveness -0.14 0.13 0.07 -0.01 0.09 -0.01
23. Arbitrariness 0.07 -0.03 -0.06 -0.07 0.07 0.01
11 12 13 14 15 16
1. Bribe size
2. GDP (Log)
3. GDP growth
4. Political right
6. East Asia
7. South Asia
8. Latin America
11. Service 1.00
12. Agriculture -0.33 1.00
13. Construction -0.25 -0.12 1.00
14. Competition 0.07 -0.02 0.07 1.00
15. Firm size -0.25 0.06 -0.03 -0.16 1.00
16. Firm age -0.13 0.05 -0.05 -0.16 0.34 1.00
17. Export -0.13 0.02 -0.03 0.01 0.04 -0.02
18. Export tax -0.01 0.00 -0.03 -0.07 -0.01 0.11
19. Foreign -0.06 0.01 -0.01 -0.17 0.27 0.09
20. Export -0.09 -0.03 -0.03 -0.16 0.23 0.04
21. State ownership -0.13 0.06 0.00 -0.20 0.33 0.29
22. Pervasiveness 0.04 -0.02 -0.03 -0.04 0.05 0.02
23. Arbitrariness -0.01 0.03 -0.03 0.05 0.03 0.04
17 18 19 20 21 22
1. Bribe size
2. GDP (Log)
3. GDP growth
4. Political right
6. East Asia
7. South Asia
8. Latin America
15. Firm size
16. Firm age
17. Export 1.00
18. Export tax 0.06 1.00
19. Foreign 0.03 0.01 1.00
20. Export 0.03 0.00 0.35 1.00
21. State ownership 0.01 0.01 0.47 0.16 1.00
22. Pervasiveness -0.03 0.03 -0.01 0.05 -0.03 1.00
23. Arbitrariness -0.04 0.05 -0.04 -0.03 0.12 0.03
Observations N=3,119. Correlations above 10.0211 are significant at
the 5% level (2-tailed t-test)
Table 3: Heckman selection model: The incidence of bribery and the
size of bribes (probit and maximum likelihood analyses)
Variables The incidence of bribery
Constant 0.258 (0.290)
GDP (Log) -0.153 ([dagger]) (0.025)
GDP growth 0.010 (0.008)
Political rights 0.123 ([dagger]) (0.014)
Property rights protection 0.178 ([dagger]) (0.015)
Manufacturing 0.105 * (0.061)
Service 0.046 (0.062)
Construction 0.228 ** (0.094)
Competition 0.440 ([dagger]) (0.028)
Firm Size -0.103 ([dagger]) (0.031)
Firm Age -0.004 ([dagger]) (0.001)
Export promotion 0.076 ([dagger]) (0.017)
Export tax -0.002 ([dagger]) (0.000)
Export orientation 0.001 (0.001)
Foreign ownership -0.003 ([dagger]) (0.001)
State ownership -0.002 ([dagger]) (0.001)
Pervasiveness 0.133 ([dagger]) (0.018)
Arbitrariness -0.314 ([dagger]) (0.024)
Mills lambda -0.283 (0.290)
Observation N 5215
Wald [chi square] (36) 1366.42
Prob. > [chi square] 0.0000
Variables The size of bribery
Constant 4.788 ([dagger]) (0.460)
GDP (Log) -0.082 ** (0.036)
GDP growth -0.027 *** (0.009)
Political rights 0.037 ** (0.019)
Property rights protection
Manufacturing -0.122 * (0.071)
Service 0.010 (0.070)
Construction 0.004 (0.098)
Competition -0.046 (0.051)
Trans-Europe -0.168 (0.149)
East Asia -0.027 (0.158)
South Asia 0.090 (0.158)
Latin America -0.105 (0.142)
Firm Size -0.115 *** (0.038)
Firm Age -0.003 ** (0.001)
Export promotion 0.078 ([dagger]) (0.019)
Export tax 0.001 (0.000)
Export orientation -0.002 *** (0.001)
Foreign ownership -0.004 ([dagger]) (0.001)
State ownership -0.002 ([dagger]) (0.001)
Pervasiveness 0.122 ([dagger]) (0.019)
Arbitrariness -0.279 (0.035)
Observation N 3119
Wald [chi square] (36)
Prob. > [chi square]
* p < 0.001; ** p < 0.05; *** p < 0.01; ([dagger]) p < 0.10 (2-tailed)
The values shown in each block are the unstandardized regression
coefficients B. Standard errors are in parentheses. Dependent
variable for Probit is the incidence of bribery. Dependent variable
for Maximum Likelihood is the amount of bribery money. For 'industry'
control, omitted category (reference) is agriculture. For 'region'
control, omitted category (reference) is OECD
Table 4: Hierarchical regression analysis predicting bribe size
Variables Hypo. Model 1
Foreign ownership H1
Export orientation H2
State ownership H3
Inverse Mills ratio -1.463 * (0.144)
GDP (Log) -0.014 (0.052)
GDP growth -0.029 ([dagger]) (0.015)
Political rights -0.027 (0.018)
Trans-Europe -0.383 (0.293)
East Asia -0.123 (0.305)
South Asia 0.009 (0.299)
Latin America -0.302 (0.307)
Manufacturing -0.188 *** (0.082)
Services 0.005 (0.093)
Construction -0.115 (0.115)
Competition -0.270 * (0.053)
Firm size -0.117 * (0.034)
Firm age 0.000 (0.001)
Export promotion 0.030 (0.026)
Export tax 0.002 * (0.000)
Constant 5.495 * (0.649)
Change in F stat. 32.79 *
Variables Hypo. Model 2
Foreign ownership H1 -0.003 * (0.001)
Export orientation H2 -0.002 ** (0.001)
State ownership H3 -0.001 *** (0.001)
Inverse Mills ratio -1.268 * (0.152)
GDP (Log) -0.041 (0.048)
GDP growth -0.030 *** (0.014)
Political rights -0.017 (0.018)
Trans-Europe -0.360 (0.288)
East Asia -0.107 (0.300)
South Asia 0.015 (0.294)
Latin America -0.279 (0.302)
Manufacturing -0.163 *** (0.076)
Services 0.015 (0.088)
Construction -0.080 (0.112)
Competition -0.267 * (0.051)
Firm size -0.055 (0.035)
Firm age 0.000 (0.001)
Export promotion 0.040 (0.026)
Export tax 0.002 * (0.000)
Constant 5.715 * (0.615)
Change in F stat. 1.02 ***
Variables Hypo. Model 3
Foreign ownership H1 -0.004 * (0.001)
Export orientation H2 -0.002 ** (0.001)
State ownership H3 -0.002 ** (0.001)
Pervasiveness H4 0.122 * (0.021)
Arbitrariness H5 -0.279 * (0.037)
Inverse Mills ratio -0.283 ([dagger]) (0.161)
GDP (Log) -0.082 ([dagger]) (0.046)
GDP growth -0.027 *** (0.012)
Political rights 0.037 *** (0.018)
Trans-Europe -0.168 (0.260)
East Asia -0.027 (0.267)
South Asia 0.090 (0.266)
Latin America -0.105 (0.274)
Manufacturing -0.122 ([dagger]) (0.072)
Services 0.010 (0.084)
Construction 0.004 (0.112)
Competition -0.046 (0.050)
Firm size -0.115 ** (0.036)
Firm age -0.003 *** (0.001)
Export promotion 0.078 ** (0.024)
Export tax 0.001 *** (0.000)
Constant 4.788 * (0.604)
Change in F stat. 2.04 **
* p < 0.001; ** p < 0.01; *** p < 0.05; ([dagger]) p < 0.10
(Significance tests are two-tailed for control variables and
one-tailed for hypothesized effects)
The values shown in each block are unstandardized regression
coefficients. Robust White Standard errors are in parenthesis
(clustering identical country groups). Dependent variable is the
amount of bribery money. For 'industry' control, omitted category
(reference) is agriculture. For 'region' control, omitted category
(reference) is OECD