Gender, Culture, and Corruption: Insights from an

Experimental Analysis*
Vivi Alatasa, Lisa Cameronb, Ananish Chaudhuric, Nisvan Erkalb,†, Lata Gangadharanb
a
World Bank, Jakarta 12190, Indonesia. (e-mail: valatas@worldbank.org)
b
Department of Economics, University of Melbourne, VIC 3010, Australia.
(e-mail: {lcameron, latag, n.erkal}@unimelb.edu.au)
c
Department of Economics, University of Auckland, Private Bag 92019, Auckland,
New Zealand. (e-mail: a.chaudhuri@auckland.ac.nz)

January 2008

Abstract
In recent years, a substantial body of work has explored the differences in the behavior of men
and women in a variety of economic transactions. We contribute to this literature by
investigating gender differences in behavior when confronted with a common bribery problem.
Our study departs from the previous literature on gender and corruption by using economic
experiments. Based on data collected in Australia (Melbourne), India (Delhi), Indonesia (Jakarta)
and Singapore, we show that while women in Australia are less tolerant of corruption than men
in Australia, there are no significant gender differences in the propensities to engage in and
punish corrupt behavior in India, Indonesia and Singapore. Hence, our findings suggest that the
gender differences reported in the previous studies may not be nearly as universal as stated and
may be more culture-specific. We also explore behavioral differences by gender across countries
and find that there are larger variations in women’s behavior towards corruption than in men’s
across the countries in our sample.

JEL Classification: C91, J16, K42, O12.

Keywords: Gender, Corruption, Experiments, Punishment, Multicultural Analysis

*
We are grateful to the World Bank, the Faculty of Economics and Commerce at the University of Melbourne, and
the University of Auckland for their financial assistance. Lynette de Silva, Syarifah Liza Munira, Daniel Piccinin,
Revy Sjahrial, Jonathan Thong, and Vicar Valencia have provided excellent research assistance.
†
Corresponding author.

Electronic copy available at: http://ssrn.com/abstract=1088811

1. Introduction

In recent years, a substantial body of work has explored the differences in the

behavior of men and women in various economic transactions. This paper contributes to this

literature by investigating gender differences in behavior when confronted with a common

bribery problem.

Due to the negative impact of corruption on economic development, eliminating

corruption is a major concern for many countries. Two recent empirical papers have

examined the relationship between gender and corruption. Dollar et al. (2001) use country-

level data for a sample of more than 100 countries and find that the greater the representation

of women in parliament, the lower the country’s level of perceived corruption. This finding is

consistent with the results of Swamy et al. (2001), who use both micro-level survey-data from

a range of countries and country-level data. They also find that on average women are less

tolerant of corruption than men.1

Our study departs from these two papers by using economic experiments, which

allows us to explore individuals’ attitudes towards corruption.2 One issue with drawing

conclusions on the basis of surveys is that actual behavior (especially when confronted with

non-trivial amounts of money) may be quite different from survey responses. Experiments

differ from surveys and perception indices in that the participants in the experiments receive

actual monetary payments, the amounts of which depend on the decisions they make during

1
Their micro-level data is based on surveys that ask respondents about the acceptability of various dishonest or
illegal behaviors. They find that a larger proportion of women than men believe that illegal or dishonest
behavior are never justifiable. These results are consistent with those of Glover et al. (1997) and Reiss and Mitra
(1998), who find that gender affects whether an individual regards certain workplace behavior as unacceptable.
2
In the experimental literature, behavioral differences between men and women have been studied using public
goods, ultimatum, dictator, and trust games. The results have been mixed, with some studies suggesting that
women are more socially oriented, others finding that men are more socially oriented, and still others finding
that there are no significant gender differences. See, for example, Andreoni and Vesterlund (2001), Bolton and
Katoc (1995), Brown-Kruse and Hummels (1993), Nowell and Tinker (1994), Cadsby and Maynes (1998),
Eckel and Grossman (1998 and 2000), and Solnick (2001). Croson and Gneezy (2005) provide an excellent
survey.

Electronic copy available at: http://ssrn.com/abstract=1088811

the experiments. Hence, we explore whether the gender differences reported in the previous

studies on corruption are also evident in an experimental setting.3

Gender differences may be the result of both biological and social differences, i.e.,

differences in social roles of men and women. An individual’s social role and presence in the

public domain may play an important role in that individual’s exposure to corruption. Hence,

if women and men differ in their social roles, one may also expect them to differ in their

attitudes towards corruption. Higher levels of exposure to corruption in daily life may

promote a tolerance and acceptance of corruption that is reflected in norms of behavior. In

addition, women may be more victimized by (and, hence, less tolerant of) corruption in

countries where their presence in the public domain is lower.4,5

To investigate whether there are consistent gender differences across countries, we

conducted experiments in four countries: Australia (Melbourne), India (Delhi), Indonesia

(Jakarta), and Singapore. Two of the countries in our sample are consistently ranked among

the least corrupt countries in the world (Australia and Singapore, with scores of 8.7 and 9.4

out of 10 respectively) and two of them are consistently ranked among the most corrupt

(India and Indonesia, with scores of 3.3 and 2.4 respectively).6

Our results show that the gender differences found in the previous studies, which are

largely based on data from Western countries, are also evident in the experimental data from

3
There is a growing literature that analyses corruption using experimental methodology. See Abbink (2005) for
a survey. However, except for Frank and Schulze (2000), none of these papers explore the relationship between
gender and corruption. Frank and Schulze (2000) analyze whether economists behave in a more self-interested
way than other people. They find that economics students are significantly more corrupt than others, with male
economists being the most corrupt and male non-economists the least.
4
Although all of the participants in our experiments were upper-level undergraduate or graduate students, their
expectations and attitudes would nevertheless be influenced by the differing roles of men and women in their
societies.
5
We discuss in Section 4 possible explanations for why gender differences may vary across cultures.
6
These rankings are based on the Corruption Perceptions Index (CPI), released annually by Transparency
International. The CPI ranks countries in terms of the degree to which corruption is perceived to exist among
politicians and public officials, based on the views of analysts and business people around the world. See
www.transparency.org/policy_research/surveys_indices for more information. Table A1 in the Appendix
contains a selective list of country rankings from the latest (2006) Corruption Perceptions Index.

2
Australia. That is, Australian men are more likely to engage in and more tolerant of

corruption than Australian women. However, we find no systematic gender differences in the

three Asian countries included in our study. Thus, gender differences in attitudes towards

corruption appear not to be as robust as suggested by the previous evidence and may be

culture-specific.

We also investigate whether cross-country variation in behavior is similar for men and

women. We find greater variation in the behavior of women across the four countries we

study than in the behavior of men. Cross-country variation in attitudes to corruption may

reflect the differing levels of exposure to corruption in the different countries.7 Women may

react differently to this exposure than men since there may be a larger variation in the social

roles of women than in the social roles of men across countries.

The paper proceeds as follows. We discuss the experimental design in Section 2 and

present the results in Section 3. We then discuss the implications of our results, possible

explanations for why gender differences may vary across cultures, and avenues for future

research in Section 4.

2. Experimental Design and Procedure

Since different cultures may have different perceptions of corruption, we wanted to

capture in our experimental design behavior which would be viewed as corrupt in all of the

countries included in our study. One of the fundamental aspects of corruption is that the

parties who engage in it benefit from it at the expense of parties external to the corrupt

transaction. We wanted to examine the behavior of parties on both sides of the corruption

problem, those who are perpetuators of it as well as those who are victims of it. Hence, our

7
See Cameron et al. (2006) for a detailed discussion of how attitudes towards corruption vary across the four
countries considered in this study.

3
experiment is based on a game where two players can act corruptly to increase their own

payoff at the expense of a third player. The bribery that takes place between the first two

players harms the third player and is illegal. Hence, the third player, the victim, is allowed to

punish the first two players at a cost.8

More specifically, the experiment is based on a three-person, sequential-move game.

The first player in the game is called the firm and is given the option to initiate a corrupt act

by offering a bribe to a government official. The second player, whom we call the official,

can either reject or accept the bribe. If the bribe is accepted, both the firm and the official are

monetarily better off at the expense of the third player – the citizen. The citizen can, however,

respond to the corrupt act by choosing to punish both the firm and the official. The

punishment is costly to the citizen, but imposes a much larger monetary sanction on the firm

and the official.9

This set-up allows us to examine two types of behavior: (1) the incentive to engage in

a corrupt act from which one reaps benefits and (2) the incentive to incur a cost to punish a

corrupt act that decreases one’s payoff. This distinction enables us to examine whether

individuals behave differently depending on whether they directly benefit from a corrupt act.

Figure 1 contains an extensive-form representation of the game, where all of the

payoffs are denoted in experimental dollars. We constrain the amount of the bribe that the

firm can offer to B ∈ [4, 8]. It costs the firm two experimental dollars to offer a bribe and the

firm incurs this cost regardless of whether the bribe is accepted. This cost represents, for

8
Note that the World Values Survey also assesses the attitudes of people in different countries by asking their
views on how justifiable it is to accept a bribe. See www.worldvaluessurvey.org/.
9
We chose to use emotive terms such as “bribe” and “punishment” in the instructions since our aim was to
simulate a real-life corrupt transaction. Cooper and Kagel (2003) consider the role of loaded language in
signaling games and suggest that the use of a meaningful context might better capture behavior in field settings
than the use of neutral language. On the other hand, Abbink and Hennig-Schmidt (2002) find that the use of
words like “bribe” does not make a difference in the corruption game that they study.

4
example, the cost of finding the right official to bribe.10 If a bribe is offered, the official

decides whether to accept it. If the official decides to accept the bribe, the payoffs to the firm

and the official increase by 3B. The payoff to the citizen decreases by the amount of the

bribe, B. Hence, the net benefit to the firm from paying the bribe is 3B – 2. This may, for

example, represent the benefit the firm gets from avoiding a regulation. We assume that the

official’s payoff also increases by 3B even though the amount of bribe paid by the firm is B.

This is due to an assumption of difference in the marginal utilities of income. Since the

income earned in the public service is likely to be lower than that earned in private firms, the

same amount of money can be assumed to have a lower marginal utility value to the firm than

to the official.11

If a bribe has been offered and accepted, the citizen, who moves last after observing

the choices made by the firm and the official, is given a chance to punish the firm and the

official for the corrupt transaction. The citizen can choose an amount P in punishment. Such

punishment is costly for the citizen and reduces the citizen’s payoff by the amount of the

punishment, P. We assume punishment is costly to the citizen for two reasons. First, the cost

may represent the amount of tax the citizen has to pay for a legal system to exist. Second, it

may represent the costs of filing a police report, appearing in court, etc. Since in most cases,

these costs are much less than the amount of punishment actually imposed on the parties, we

assume that if the citizen chooses a punishment amount of P, the firm and the official suffer a

payoff reduction of 3P.

In the subgame perfect equilibrium of this game, regardless of the parameters chosen,

a payoff-maximizing citizen chooses not to punish. Knowing this, the official accepts the

10
To offer a bribe, firms usually have to incur some transaction costs. These costs are usually constant and have
to be incurred irrespective of the size of the bribe being offered.
11
The choice of multipliers has the additional advantage of helping us prevent negative total payoffs.

5
bribe and the firm offers the bribe. Moreover, the firm offers the maximum amount of bribe it

can since its payoff is increasing in the amount it offers.

We have deliberately chosen to conduct a one-shot game because in a one-shot game

the punishment has no economic benefit to the citizen. The decision to punish is not affected

by the anticipation of possible future economic gains. This implies that if we observe any

punishment by the citizens, we can infer that it is motivated by either negative reciprocity or

moral considerations. Hence, with a one-shot game, a comparison of the citizens’ willingness

to punish across different countries reveals the differences in the tolerance of corrupt acts in

those countries.

The one-shot nature of the game also helps us avoid issues associated with repeated

games, such as signaling, reputation formation and serial correlation in decisions. Each

subject in our database participated in the experiment only once and played only one role.12

The subjects playing the three roles were grouped anonymously in the experiment to avoid

conscious or unconscious signaling.

The experiments were run at the University of Melbourne, the Delhi School of

Economics, the University of Indonesia in Jakarta, and the National University of Singapore

using third year undergraduate or postgraduate students. In order to minimize the

experimenter effects, we made sure that one of the authors (the same one) was present in all

the countries where we ran the experiment.13

12
One standard response in cases such as these is to have random re-matching of subjects. Kandori (1992) states
that it is not clear whether random re-matchings do actually succeed in eliminating supergame effects. However,
Duffy and Ochs (2005) consider an experiment with an indefinitely repeated 2-player prisoner’s dilemma game
and find that contrary to Kandori’s theoretical conjecture, a cooperative norm does not emerge in the treatments
where players are matched randomly. In the current paper, we decided to adopt a conservative stance and have
players participate in pure one-shot games to avoid any repeated game effects.
13
See Roth et al. (1991) and Cardenas and Carpenter (2005) for a discussion of the methodological issues
arising in multi-site experiments.

6
 All the sessions were run as non-computerized experiments. At the beginning of each

session subjects were asked to come to a large lecture theatre. Each session consisted of at

least 30 subjects. These subjects, on entering the room, were randomly designated as either

firms, officials or citizens. Each group was located far apart from the others in a recognizable

cluster. Thus, each group could see the members of the other groups, but individual subjects

were unaware of which three specific subjects constituted a particular firm-official-citizen

trio.

At the beginning of each session, each subject received a copy of the instructions,

which were then read out loud to them. They were also given a number of examples

explaining how the payoffs would be calculated for specific bribe and punishment amounts.

Then, the subjects playing the role of a firm were asked to decide whether or not to offer a

bribe. If they chose to offer a bribe, they also had to choose an amount. After they made their

decisions, the record sheets with the bribe amounts were collected by the experimenter and

distributed to the corresponding officials. After the officials made their decisions, the record

sheets of both the firms and the government officials were given to the corresponding

citizens. Hence, the citizens learned whether a bribe was offered and whether it was accepted.

The game ended after the citizens decided whether to punish by choosing a punishment

amount. All the subjects were then asked to fill out a demographic survey, which included

questions on age, gender, income, education stream, employment history, and frequency of

exposure to corruption. Those in the role of the citizen were also asked to explain the

motivation for their decisions.

Each experiment lasted about an hour. At the end of each session, the decisions made

by all of the subjects were entered into a spreadsheet which generated their payoffs. The

payoffs were converted into cash using an appropriate conversion rate, taking into

7
consideration purchasing power parity across the countries where the experiment was

conducted.14 These conversion rates were public information. To guarantee parity in the

payoffs to the different type of players (firm, official and citizen), we used a different

conversion rate for each type.15

3. Results

With our experimental design, we are interested in exploring two issues. In Section

3.1, we start by investigating whether, controlling for culture (i.e., within each country),

women are less tolerant towards corruption than men. We then control for gender in Section

3.2 and investigate whether there are larger cross-country variations in the behavior of

women than in the behavior of men in the context of our game.

A total of 1326 subjects participated in the experiments. Of these, 596 (45%) were

men. The number of participants in Australia, India, Indonesia and Singapore were 642, 309,

180 and 195 respectively.16

14
The conversion rates in each country were based on 1) the standard hourly wage paid for a student research
assistant in each country and 2) a typical basket of goods bought by students in each country. This is similar to
the procedure used by other researchers who have conducted cross-cultural studies (e.g., Carpenter and
Cardenas, 2004 and Cardenas and Carpenter, 2005).
15
In Australia, the conversion rates were 3 experimental currency = 1 real currency for the firms, 2 experimental
currency = 1 real currency for the officials and 1.5 experimental currency = 1 real currency for the citizens.
Each subject made on average AU$20. This amount is approximately equivalent to US$15. In India subjects
were paid an average of US$11, in Singapore US$13, and in Indonesia US$9. Different conversion rates for
different player types are sometimes used in experiments if the payoffs are expected to be very different across
the subjects. Davis and Holt (1993) recommend that average payments in experiments should be high enough to
compensate all participants for the opportunity cost of their time (pp. 24-26). Having different conversion rates
for different types of players helped us achieve this goal since the equilibrium payoffs were highly asymmetric
across the different player types (Firm, Official, and Citizen) in the experiment.
16
In Australia, 107 men and 107 women made decisions as firms, 89 men and 95 women as officials, and 59
men and 99 women as citizens. In India, 49 men and 54 women made decisions as firms, 39 men and 58 women
as officials, and 44 men and 43 women as citizens. In Indonesia, 32 men and 28 women made decisions as
firms, 22 men and 26 women as officials, and 17 men and 20 women as citizens. In Singapore, 23 men and 42
women made decisions as firms, 26 men and 30 women as officials, and 23 men and 27 women as citizens. As
is clear from the design of the experiment, not all officials and citizens got the opportunity to make a decision,
which is the reason we did not have a complete gender balance across the different roles.

8
 We report results based on t-tests and multivariate regression analysis, where we

estimated binary probit models for the bribe, acceptance and punishment rates, and ordinary

least square models for the bribe and punishment amounts.17 The regression results control

for variables not accounted for in the t-tests, such as field of study (whether it is economics)

and the percentage of each Australian subject’s life that has been spent outside of Australia.18

Of the variables we collected information on in the surveys, these were the only ones that

were found to be consistently significant determinants of subject behavior. In the regressions

for the officials’ and citizens’ behavior, we also control for the bribe amount.

The reported results are based on two different treatments that were conducted. In the

Indian experiments and a subset of the Australian sessions, the citizens’ punishment range

was restricted to P ∈ [2, 8].19 We refer to this treatment as “Treatment 1.” In the other

countries and the remaining Australian sessions, the punishment range was extended to P ∈

[2, 12].20 This is “Treatment 2.” The t-tests below make comparisons within treatment and

the regression results include a control for treatment. The variation in treatment design

enabled us to examine the effectiveness of the punishment regime. We discuss the treatment

effects in detail in Cameron et al. (2006). Since the focus of the current paper is gender

differences and since gender differences do not vary across the treatments, we do not discuss

the treatments effects here.21

17
We also estimated ordered probit models for positive bribe and punishment amounts. These recognize that the
dependent variable is not continuous. The results were very similar to the reported results from the estimation of
ordinary least squares models.
18
The last variable controls for the high number of foreign students that study in Australian universities. The
majority of these students come from Asia. We find this variable to be insignificant in explaining behavior in
most of the regressions. This is possibly because those who choose to study in Australia are more westernized
than their counterparts and/or quickly absorb the social norms of the new environment.
19
Due to resource constraints, we do not have data for all treatments from all the countries.
20
These values were chosen to guarantee two things. First, we wanted to ensure that no one obtained a negative
payoff. Second, we wanted to make sure that the average earning was high enough to offset the participants’
opportunity cost of time (Davis and Holt, 1993).
21
Cameron et al. (2006) also present and discuss results from a third treatment. In both Treatments 1 and 2, the
bribe is welfare-enhancing in that the total payoff gains to the firm and the official exceed the payoff loss to the

9
3.1 Are women less tolerant of corruption than men?

As stated above, both Dollar et al. (2001) and Swamy et al. (2001) find that women

are less tolerant of corruption than men. Within the design of our experiment, this is

equivalent to asking whether female participants in the four countries where we ran our

experiment had a lower propensity to pay bribes, a lower propensity to accept bribes, and a

higher propensity to punish bribery than the male participants.

Table 1 presents the results of t-tests for differences in the means of the behavior of

the male and female participants in the three roles. Panel A of Table 1 pools the data and

shows that overall the male participants have a higher propensity to bribe than the female

participants (p = 0.04), but that there are no other statistically significant gender differences

in behavior. However, if we break the data down by individual countries (Panels B-E), we

observe that the difference in the bribe rates is driven by Australia. In Australia, 91.6% of

male participants offered bribes compared with 80.4% of female participants (p = 0.02). In

none of the other countries do we see any significant gender differences in the propensities to

offer bribes. Further, in Australia, the male subjects also had higher acceptance rates and

lower punishment rates than the female subjects. The bribe was accepted 92.1% of the time

when it was offered to a male participant in Australia while it was accepted 80% of the time

when it was offered to a female participant. This difference is statistically significant

according to a test of difference of means (p = 0.02). The Australian male participants in the

role of the citizen chose to punish 49.2% of the time while the Australian female participants

chose to punish 62.6% of the time. This difference is marginally significant at the 10% level.

citizen. In Treatment 3, the payoffs are altered so that the combined gains to the firm and the official are less
than the payoff loss to the citizen. Hence, the bribe is welfare-reducing. Since the gender differences are similar
across all three treatments, we chose not to discuss Treatment 3 in this paper for ease of exposition.

10
 In India, Indonesia and Singapore, we find no significant differences in the behavior

of the male and female participants in the three roles. The point estimates also do not vary

systematically by gender. For example, in India men bribe more often, but also punish more

often.

The regression results presented in Table 2 confirm the results from the t-tests. Panel

A pools all the data across all the countries. Overall, men offer bribes with a higher frequency

(significant at the 5% level) and punish corrupt acts by higher amounts (significant at the

10% level). In Panel B, the effect of gender is allowed to differ by country. For example, the

coefficient on the variable “Male-Australia” captures the difference between men and women

in Australia. The results show that in Australia men bribe approximately 8 percentage points

more often, accept bribes approximately 8 percentage points more often, and punish bribery

about 14 percentage points less often than women. However, if the Australian men do punish,

then they do so by a larger amount than the Australian women. In the other countries, there

are no significant gender differences in the bribe, acceptance, and punishment rates. The only

significant differences we find are in the bribe and punishment amounts. Specifically, the

Indian male subjects, when they bribe, offer larger bribes than the Indian female subjects, and

the Indonesian male subjects, when they punish, choose higher punishment amounts than the

Indonesian female subjects.

A possible criticism of our results is that the difference we observe in the behavior of

men and women in Australia may be the result of gender differences in other-regarding

preferences, such as inequity aversion, or in motivations for punishment, such as negative

reciprocity. To examine this issue further, we conducted a set of experiments with Australian

subjects using neutral language, where we replaced words such as “bribe” and “punishment”

11
with “transfer” and “forego money to reduce others’ payoff,” respectively.22 Moreover,

instead of designating different types of players as firms, officials and citizens, we referred to

them as players 1, 2 and 3.

Table 3 presents the results from these experiments. In comparison, both genders offer

and accept transfers more often in the neutral-language treatment than they offer and accept

bribes in the loaded-language treatment. They also punish less often. However, the gender

differences in behavior are much less in the neutral-language treatment than they are in the

loaded-language treatment. In the neutral-language treatment, women’s propensity to offer a

transfer is not significantly different from the men’s (100% of the time instead of 94%). Their

propensity to punish is not different either (30% in both cases). These results suggest that the

use of loaded language stimulates a reaction to corruption and that Australian women react

more strongly against a corrupt transaction than Australian men do.

The only exception is in the acceptance rates. Women still accept less often than men

(85% versus 100%) and the difference remains statistically significant (p = 0.08). It is not

clear why the acceptance rate decision would differ from the other two decisions. If women

are more risk averse or more concerned about fairness than men, this would also lead them to

“bribe” less often in the neutral-language treatment, which they do not do. The difference in

the decision to accept is driven by the behavior of only 4 (out of 26) women. In fact, if we

group the decision to offer and accept a transfer together, we find that the probability of

engaging in a transaction to increase one’s own payoff at the expense of another player is

very similar across the genders (92% for women versus 97% for men, p = 0.34) Doing the

same exercise with the loaded-language data reveals that the difference is large and

statistically significant (75% for women versus 87% for men, p = 0.03). Hence, we conclude

22
The neutral-language experiments were run with Treatment 2 only, where a larger range of punishments were
allowed.

12
that the neutral-language treatment supports our contention that the gender differences we

observe in the loaded-language experiments reflect different reactions to the corrupt

context.23,24

3.2 Does the cross-country variation in behavior differ by gender?

Our finding in the previous section is that the differences between men and women do

not necessarily lead to statistically significant behavioral differences in terms of corruption.

Another way to think of the impact of social roles is to observe how it affects the behavior of

one gender across countries. To do this, we start by discussing the variations in the behavior

of men. Table 4A, Panels (i)-(iv) compare the means of behavior across the Australian,

Indian, Indonesian and Singaporean male subjects. These pair-wise country comparisons

show that there are no significant differences in the propensities to bribe, the bribe amounts,

and the propensities to accept. Hence, in terms of the propensities to engage in corrupt

behavior, the male subjects in all four countries display similar tendencies.

It is only when we consider the propensities to punish corrupt behavior that we see

some significant differences in the behavior of male subjects in the four countries.

Specifically, the Indonesian male subjects have the highest rate of punishment followed by

the Australian male subjects (76.5% and 50% respectively). This difference is significant at

the 10% level. The Singaporean male subjects punished in 39.1% of the cases. Although their

rate of punishment is not statistically significantly different from that of the Australian male

23
As further evidence, we also checked to see whether there were significant gender differences in the reasons
the citizens gave for their decisions to punish in the post-experimental survey we asked them to fill out. If it is
the case that Australian women differ from Australian men in terms of their other-regarding preferences or
motivations for punishment, one would expect them to cite reasons of fairness or negative reciprocity more
frequently while explaining their decisions to punish. However, we find this not to be the case. On the contrary,
the Australian women cite punishing for moral reasons more often than the Australian men (39% of the female
citizens who had the chance to punish versus 25% of the male citizens who had the chance to punish). The
difference is statistically significant with a p-value of 0.07. Hence, our view that it is the Australian women’s
lower tolerance of corruption that causes the gender differences in behavior was further strengthened.
24
Note that it is possible that men and women react differently to the framed context and that what we observe
is not the real difference in their tolerance of corruption. Although it is not clear why this would be the case, we
cannot rule this out as a possible explanation of the gender difference we observe.

13
subjects (p = 0.46), it is significantly less than that of the Indonesian male subjects (p = 0.02).

The Indian male subjects have the lowest punishment rate of all (27.3%). This is significantly

less than the punishment rate of the Australian male subjects (p = 0.06).

The regression results presented in Table 2, Panel C confirm the results from the t-

tests.25 We test for equality of coefficients across the four countries for each gender. As

shown in the table, the tests indicate that we are unable to reject the hypothesis that male

behavior in each of the countries is the same, except in the case of punishment rates (p =

0.08). In the case of punishment rates, the regression results show that, once we control for

the field of study (i.e., whether it is economics), the percentage of each Australian subject’s

life that has been spent outside of Australia, and treatment effects, the punishment behavior

of the male subjects in Australia is not significantly different from that in any of the other

countries. However, since the male subjects in Indonesia have significantly higher rates of

punishment than those in India and Singapore, we get the result that the coefficients in this

case are not equal to each other.26

In contrast, the t-tests reported in Table 4B and regression results reported in Table 2,

Panel C reveal differences in female behavior across the four countries in all categories of

comparison. Testing for equality of regression coefficients, we find that female behavior

varies across the four countries in the case of bribe rates, bribe amounts, and punishment

rates. All of these differences are significant at the 5% level. In the case of acceptance rates

and punishment amounts, we are only narrowly unable to reject a hypothesis of equality of

25
These results are the same as those presented in Table 2, Panel B. However, they are configured (by
interacting both the male and female dummies with the country dummies) to enable an easier interpretation of
within-gender cross-country differences.
26
The pairwise regression tests give p-values of 0.058 and 0.028 respectively. The high rate of punishment we
observe among the Indonesian male subjects is an unexpected outcome given the high level of corruption in this
country. One possible explanation for this outcome is the recent institutional changes that have occurred in
Indonesia. Since the introduction of democracy in Indonesia in 1998 and the relaxation of media restrictions,
corruption has received a lot more negative media attention. This may have resulted in a hardening of attitudes
against corruption. See Cameron et al. (2006) for a more detailed discussion of this point.

14
coefficients at the 10% level (with p-values of 0.12 and 0.11 respectively). Moreover,

unreported pairwise tests of the regression coefficients show that the acceptance rate in

Singapore is significantly higher than that in each of the other three countries.

The magnitude of the cross-country variation in female behavior is quite large. For

instance, the regression results show that the female bribe rate in Australia is 16.6 percentage

points lower than that in Indonesia and 17.2 percentage points lower than that in Singapore (p

= 0.02 and p = 0.007 respectively). Similarly, the female acceptance rate in Singapore is 19.7

percentage points higher than that in Australia, 15.2 percentage points higher than that in

India, and 12.9 percentage points higher than that in Indonesia (with p = 0.016, p = 0.089,

and p = 0.089 respectively).27

In summary, we find less cross-country variation in the behavior of men than in the

behavior of women. When we compare the behavior of the male subjects, we find significant

differences only in the propensity to punish corrupt behavior. In contrast, when we compare

the behavior of the female subjects, we find significant differences in both the propensity to

engage in corrupt behavior (the bribe rate and amount) and the propensity to punish corrupt

behavior. Overall, the Australian female subjects seem to have the lowest tolerance of corrupt

behavior.

4. Discussion

Our goal in this paper was to examine gender differences in behavior when

confronted with a common bribery problem. We explored two issues. First, we investigated

whether women are less likely to offer bribes and more likely to punish corrupt behavior. We

27
As explained in Cameron et al. (2006), one possible explanation for the relatively higher tolerance of
corruption we find in Singapore is the top-down policy approach that has been adopted in this country. Such an
approach could have had the effect of eradicating corruption at a faster rate than it takes to fundamentally
change society’s social norms.

15
find this to be the case in only one of the four countries studied – Australia. We do not find

significant gender differences in India, Indonesia or Singapore.

The results for the only Western country in our study are similar to those found in the

existing literature. In both Dollar et al (2001) and Swamy et al. (2001), the Western countries

make up a large part of their sample.28,29 Our findings suggest that the gender differences

found in these previous studies may be culture-specific. This is important because the gender

differences found in the previous studies on corruption have prompted policy makers in many

developing countries to recommend higher rates of female participation in the political and

economic institutions. Our results indicate that, although there may be other valid reasons for

advocating policy measures that promote female political involvement, some caution needs to

be taken in asserting that increased female participation will lower corruption in all

countries.30,31

Further work is needed to understand the reasons for the variations in gender

differences in attitudes towards corruption across countries and to establish in which

countries gender differences exist. It is possible that countries with different cultural

backgrounds display gender differences to different degrees. For example, Gneezy et al.

(2006) find that the gender differences in attitudes towards competition that are observed in

28
Swamy et al. (2001) present some results disaggregated to the country level. Interestingly, scrutiny of these
results reveals that there are no gender differences in tolerance of corruption in the three Asian nations in their
sample (China, India and South Korea). This is also true of Nigeria, the only African nation in their sample
other than South Africa.
29
Most of the previous experimental studies that have examined behavioral gender differences have been based
on data from the Western nations, with the majority being from the U.S.
30
See Duflo (2005) for a discussion of the various reasons for reserving positions for groups that are perceived
as being disadvantaged.
31
In fact, the World Values Survey (WVS, available at www.worldvaluessurvey.org/), which asks respondents
whether someone accepting a bribe is acceptable, yields results consistent with ours. Specifically, the WVS also
shows that while the Australian women are significantly less tolerant of corruption than the Australian men
(88% of the women stated accepting a bribe is never acceptable versus 83% of the men, p < 0.01), there are no
statistically significant gender differences in India and Singapore. However, according to the WVS, the
Indonesian women are significantly less tolerant of corruption than the Indonesian men (86% versus 79%, p <
0.01). The WVS was also conducted in Vietnam, the Philippines, Bangladesh, and China, where the results
again yield no statistically significant gender differences. These figures are all for the most recent survey run in
each country.

16
the Western countries are reversed in matrilineal societies. Their results provide insights into

how the existing societal structure is crucially linked to the observed gender differences in

competitiveness. In the context of corruption, one possible explanation for the different

gender effects that are observed in our data is the differing social roles of women across

cultures. In relatively more patriarchal societies where women do not play as active a role in

the public domain, women’s views on social issues may be influenced to a greater extent by

men’s views. Hence, in such societies, one would expect to see less of a gender difference in

behavior towards corruption in comparison to societies where women feel more comfortable

in voicing their own opinions.32

The second issue we investigated is whether cross-country variation in behavior is

similar for men and women. The behavior of the male subjects was shown to be quite similar

in all four countries. In contrast, there are important differences in the behavior of the female

subjects across the four countries. One possible explanation for these results is that there are

greater variations in women’s social roles across countries than in men’s. Understanding why

the cross-country variation in attitudes towards corruption differs by gender is another

important agenda for future research.

32
See, for example, Chan (2000), Bessell (2005) and Ganguly-Scrase (2000) for discussions of the limited roles
of women in the public domain in Singapore, Indonesia and India respectively. Australia, in contrast, has
historically had a pioneering role in the advancement of women’s rights (Sawer, 1994). See also Nelson and
Chowdhury (1994) for a discussion of the variation in the women’s attitudes towards participation and activism
in societal affairs across different cultures.

17
References

Abbink. K. 2005. “Laboratory Experiments on Corruption” in The Handbook of Corruption,

ed. by S. Rose-Ackerman, Edward Elgar Publishers, Cheltenham, UK and
Northampton, US.

Abbink, K. and H. Hennig-Schmidt. 2002. “Neutral versus Loaded Instructions in a Bribery

Experiment,” University of Nottingham, CeDEx Working Paper 2002-13.

Andreoni, J. and L. Vesterlund. 2001. “Which is the Fair Sex? Gender Differences in
Altruism,” Quarterly Journal of Economics, 116, 293-312.

Antecol, H. 2000. "An Examination of Cross-Country Differences in the Gender Gap in

Labor Force Participation Rates," Labour Economics, 7(4), 409-426.

Bessell, S. 2005. “Indonesia,” in Y. Galligan and M. Tremblay (eds.), Sharing Power:

Women, Parliament, Democracy, Ashgate, Burlington, VT.

Bolton, G. and E. Katoc. 1995. “An Experimental Test for Gender Differences in Beneficent
Behavior,” Economics Letters, 48, 287-92.

Brown-Kruse, J. and D. Hummels. 1993. “Gender Effects in Laboratory Public Goods

Contribution: Do Individuals Put Their Money Where Their Mouth Is?” Journal of
Economic Behavior and Organization, 22, 255-67.

Cadsby, C. and E. Maynes. 1998. “Gender and Free Riding in a Threshold Public Goods
Game: Experimental Evidence,” Journal of Economic Behavior and Organization, 34,
603-20.

Cameron, L., A. Chaudhuri, N. Erkal, and L. Gangadharan. 2006. “Propensities to Engage in

and Punish Corrupt Behavior: Experimental Evidence from Australia, India,
Indonesia and Singapore,” University of Melbourne, Department of Economics,
mimeo.

Cardenas, J. C. and J. Carpenter. 2005. “Experiments and Economic Development: Lessons

from Field Labs in the Developing World,” at
http://community.middlebury.edu/~jcarpent/papers.html.

Carpenter, J. and J. C. Cardenas. 2004. “An Inter-Cultural Examination of Cooperation in the

Commons,” at http://community.middlebury.edu/~jcarpent/papers.html.

Chan, J. 2000. “The Status of Women in a Patriarchal State: The Case of Singapore,” in L.
Edwards and M. Roces (eds.), Women in Asia: Tradition, Modernization, and
Globalization, Allen & Unwin, St. Leonards, N.S.W.

Cooper D. and J. Kagel. 2003. "The Impact of Meaningful Context on Strategic Play in
Signalling Games,” Journal of Economic Behavior and Organization, 50(3), 311-337.

18
Croson, R and N. Buchan. 1999. “Gender and Culture: International Experimental Evidence
from Trust Games,” American Economic Review Papers and Proceedings, 89(2),
386-391.

Croson, R. and U. Gneezy. 2005. “Gender Differences in Preferences,” Wharton School,

University of Pennsylvania, mimeo.

Davis, D. and C. Holt. 1993. Experimental Economics. Princeton University Press, Princeton,
NJ.

Dollar, D., R. Fisman and R. Gatti. 2001. “Are Women Really the "Fairer" Sex? Corruption
and Women in Government,” Journal of Economic Behavior and Organization,
46(4), 423-29.

Duffy, J. and J. Ochs. 2005. “Cooperative Behaviour and the Frequency of Social
Interaction,” at http://www.pitt.edu/~jduffy/papers/cooperation0405.pdf.

Duflo, E. 2005. “Why Political Reservations?,” Journal of the European Economic

Association, 3, 2-3, 668-678.

Eckel, C. C. and P. J. Grossman. 1998. “Are Women less Selfish than Men? Evidence from
Dictator Experiments,” Economic Journal, 108(448), 726-735.

Eckel, C. C. and P. Grossman. 1996. “The Relative Price of Fairness: Gender Differences in a
Punishment Game,” Journal of Economic Behavior and Organization, 30, 143-58.

Frank, B. and G. G. Schulze. 2000. “Does Economics make Citizens Corrupt” Journal of
Economic Behavior and Organization, 43, 101-113.

Ganguly-Scrase, R. 2000. “Diversity and the Status of Women: The Indian Experience,” in L.
Edwards and M. Roces (eds.), Women in Asia: Tradition, Modernization, and
Globalization, Allen & Unwin, St. Leonards, N.S.W.

Glover, S. H., M. A. Bumpus, J. E. Logan, and J. R. Ciesla. 1997. “Reexamining the

Influence of Individual Values on Ethical Decision-Making,” Journal of Business
Ethics, 16(12/13), 1319-1329.

Gneezy, U., K. L. Leonard and J. A. List. 2006. “Gender Differences in Competition: The
Role of Socialization,” University of Chicago, mimeo.

Kandori, M. 1992. "Social Norms and Community Enforcement," Review of Economic

Studies, 59, 63-80.

Nelson B. and N. Chowdhury (eds.). 1994. Women and Politics Worldwide, Yale University
Press, New Haven.

Nowell, C. and S. Tinker. 1994. “The Influence of Gender on the Provision of a Public
Good,” Journal of Economic Behavior and Organization, 25, 25-36.

19
Reiss, M. C. and K. Mitra. 1998. “The Effects of Individual Difference Factors on
Acceptability of Ethical and Unethical Workplace Behaviors,” Journal of Business
Ethics, 17(14), 1581-1593.

Roth, A., V. Prasnikar, M. Okuno-Fujiwara and S. Zamir. 1991. “Bargaining and Market
Behaviour in Jerusalem, Ljubljana, Pittsburgh and Tokyo: An Experimental Study,”
American Economic Review, 81, 1068-1095.

Sawer, M. 1994. “ Locked out or Locked In? Women and Politics in Australia,” in B. Nelson
and N. Chowdhury (eds.), Women and Politics Worldwide, Yale University Press,
New Haven.

Solnick, S., 2001. “Gender Differences in the Ultimatum Game,” Economic Inquiry, 39, 189-
200.

Swamy, A., S. Knack., Y. Lee and O. Azfar. 2001. “Gender and Corruption,” Journal of
Development Economics, 64(1), 25-55.

20
 Figure 1: The Game Tree

FIRM

Do not offer bribe Offer bribe (B)

F: 60 OFFICIAL
O: 30
C: 40 Reject bribe Accept bribe

F: 60 - 2
O: 30 CITIZEN
C: 40
Do not punish Punish (P)

F: 60 + 3B - 2 F: 60 - 2 + 3B - 3P
O: 30 + 3B O: 30 + 3B - 3P
C: 40 - B C: 40 - B - P

21
 Table 1: Gender Differences

A. All Countries, Treatments 1 and 2

Male Female p-value
% firms bribing 90.52 83.98 0.04
Bribe Amount (if >0) 7.59 7.55 0.63
% officials accepting 88.64 84.21 0.21
% citizens punishing 44.06 51.85 0.16
Punishment Amount (if >0) 6.05 5.37 0.24

B. Australia, Treatments 1 and 2

Male Female p-value
% firms bribing 91.59 80.37 0.02
Bribe Amount (if >0) 7.63 7.72 0.42
% officials accepting 92.13 80.00 0.02
% citizens punishing 49.15 62.63 0.10
Punishment Amount (if >0) 6.48 5.34 0.12

C. India, Treatment 1
Male Female p-value
% firms bribing 95.92 92.59 0.48
Bribe Amount (if >0) 7.57 7.18 0.10
% officials accepting 89.74 89.66 0.99
% citizens punishing 27.27 20.93 0.50
Punishment Amount (if >0) 3.25 4.33 0.30

D. Indonesia, Treatment 2
Male Female p-value
% firms bribing 78.13 82.14 0.70
Bribe Amount (if >0) 7.40 7.61 0.47
% officials accepting 77.27 76.92 0.98
% citizens punishing 76.47 70.00 0.67
Punishment Amount (if >0) 7.00 4.29 0.12

E. Singapore, Treatment 2
Male Female p-value
% firms bribing 91.30 83.33 0.38
Bribe Amount (if >0) 7.67 7.60 0.77
% officials accepting 84.62 93.33 0.30
% citizens punishing 39.13 48.15 0.53
Punishment Amount (if >0) 7.00 7.38 0.82

22
 Table 2: Multivariate Regression Results

A. Pooled Regression Results

Bribe (0/1) Bribe Amount (>0) Accept (0/1) Punish (0/1) Punishment Amount (>0)

1 2 3 4 5 6 7 8 9 10
M. Effect♣ p-value Coeff p-value M. Effect♣ p-value M. Effect♣ p-value Coeff p-value
India 0.059 0.32 -0.456 0.03 * 0.012 0.86 -0.277 0.01 ⌂ -2.154 0.05 *
Indonesia 0.073 0.08 # -0.254 0.23 0.025 0.68 0.045 0.72 -1.068 0.30
Singapore 0.105 0.00 ⌂ -0.096 0.64 0.100 0.06 # -0.224 0.04 * 0.665 0.53
Male 0.063 0.04 * 0.089 0.35 0.035 0.31 -0.062 0.29 1.008 0.08 #
Econ major 0.026 0.42 0.200 0.05 * 0.082 0.03 * -0.159 0.01 ⌂ -0.380 0.58
% life out of Australia 0.148 0.01 ⌂ -0.119 0.55 0.092 0.14 -0.060 0.56 -0.730 0.42
Treatment 1 0.148 0.00 ⌂ 0.031 0.82 0.090 0.08 # -0.105 0.20 -0.741 0.32
Bribe amount -0.007 0.71 -0.035 0.27 0.191 0.55
const 7.641 0.00 ⌂ 4.797 0.05 *

R-squared 0.102 0.012 0.056 0.102 0.046

N 440 383 384 332 161
♣
We report marginal effects for the probits. *, #, and ⌂ denote statistical significance at the 5%, 10%, and 1% level, respectively.

23
B. Pooled Data, Gender-Country Interaction (Australian Female Subjects are the reference category.)
Bribe (0/1) Bribe Amount (>0) Accept (0/1) Punish (0/1) Punishment Amount (>0)

1 2 3 4 5 6 7 8 9 10
M. Effect♣ p-value Coeff p-value M. Effect♣ p-value M. Effect♣ p-value Coeff p-value
India 0.074 0.26 -0.725 0.00 ⌂ 0.036 0.61 -0.367 0.00 ⌂ -0.95 0.50
Indonesia 0.105 0.02 * -0.179 0.49 0.051 0.44 -0.047 0.76 -1.85 0.12
Singapore 0.110 0.01 ⌂ -0.181 0.45 0.135 0.02 * -0.237 0.07 # 1.33 0.28
Male-Aust 0.083 0.02 * -0.044 0.74 0.084 0.06 # -0.143 0.08 # 1.34 0.09 #
Male-India 0.048 0.44 0.472 0.01 ⌂ -0.024 0.73 0.073 0.54 -0.95 0.54
Male-Indonesia -0.030 0.68 -0.203 0.43 -0.008 0.92 0.070 0.69 2.74 0.04 *
Male-Singapore 0.060 0.33 0.110 0.66 -0.121 0.30 -0.101 0.48 0.497 0.74
Econ major 0.027 0.39 0.198 0.05 * 0.083 0.03 * -0.160 0.01 ⌂ -0.364 0.59
% life out of Australia 0.152 0.01 ⌂ -0.135 0.50 0.078 0.21 -0.063 0.54 -0.793 0.38
Treatment 1 0.145 0.00 ⌂ 0.040 0.76 0.077 0.13 -0.101 0.22 -0.777 0.29
Bribe amount -0.007 0.72 -0.035 0.28 0.192 0.55
const 7.719 0.00 ⌂ 4.728 0.06 #

R-squared 0.110 0.022 0.069 0.108 0.055

N 440 383 384 332 161
♣
We report marginal effects for the probits. *, #, and ⌂ denote statistical significance at the 5%, 10%, and 1% level, respectively.

24
C. Pooled Data, Gender-Country Interaction (Australian Male Subjects are the reference category.)
Bribe (0/1) Bribe Amount (>0) Accept (0/1) Punish (0/1) Punishment Amount
(>0)

1 2 3 4 5 6 7 8 9 10
M. Effect♣ p-value Coeff p-value M. Effect♣ p-value M. Effect♣ p-value Coeff p-value
Female-Australia (α1) -0.117 0.02 * 0.044 0.74 -0.112 0.06 # 0.145 0.08 # -1.34 0.09 #
Female-India (α2) -0.011 0.89 -0.681 0.00 ⌂ -0.067 0.46 -0.237 0.06 # -2.30 0.13
Female-Indonesia (α3) 0.049 0.38 -0.135 0.60 -0.044 0.64 0.098 0.53 -3.19 0.02 *
Female-Singapore (α4) 0.055 0.28 -0.137 0.55 0.085 0.24 -0.102 0.47 -0.019 0.99
Male-India (β1) 0.040 0.59 -0.209 0.36 -0.036 0.69 -0.172 0.17 -3.244 0.02 *
Male-Indonesia (β2) 0.029 0.62 -0.338 0.18 -0.054 0.59 0.166 0.31 -0.452 0.74
Male-Singapore (β3) 0.089 0.08 # -0.027 0.92 0.016 0.85 -0.196 0.17 -0.507 0.74
Econ major 0.027 0.39 0.198 0.05 * 0.083 0.03 * -0.160 0.01 ⌂ -0.364 0.59
% life out of Australia 0.152 0.01 ⌂ -0.135 0.50 0.078 0.21 -0.063 0.54 -0.793 0.38
Treatment 1 0.145 0.00 ⌂ 0.040 0.76 0.077 0.13 -0.101 0.22 -0.777 0.29
Bribe amount -0.007 0.72 -0.035 0.28 0.192 0.55
Const 7.719 0.00 ⌂ 4.728 0.06 #

Tests:
Female: (α1= α2= α3= α4) 0.04 * 0.02 * 0.12 0.01 ⌂ 0.11
Male: (β1= β2= β3) 0.35 0.48 0.86 0.08 # 0.14

R-squared 0.110 0.022 0.069 0.108 0.055

N 440 383 384 332 161
♣
We report marginal effects for the probits. *, #, and ⌂ denote statistical significance at the 5%, 10%, and 1% level, respectively.

25
 Table 3: Neutral versus Loaded Language (Australia, Treatment 2)

Loaded Language Neutral Language

Male Female p-value Male Female p-value

% firms offering a bribe (transfer) 87.3 71.2 0.032 94.4 100 0.22
Bribe (transfer) amount (if >0) 7.67 7.64 0.85 7.71 7.37 0.23
% officials accepting 85.7 78.3 0.38 100 84.4 0.08
% citizens punishing 50.0 68.6 0.11 30.4 30.0 0.98
Punishment amount (if >0) 7.08 5.57 0.22 6.71 5.66 0.71
% participating in a corrupt act 86.7 74.6 0.03 97.2 92.5 0.34

26
 Table 4A: Differences Between Males Across Countries

Australia India
p-value
(i) (Treatment 1) (Treatment 1)
% of firms bribing 96.15 95.92 0.95
Bribe amount (if >0) 7.60 7.57 0.89
% of officials accepting 96.30 89.74 0.21
% of citizens punishing 48.48 27.27 0.06
Punishment amount (if >0) 6.00 3.25 0.01
Australia Indonesia
p-value
(ii) (Treatment 2) (Treatment 2)
% of firms bribing 87.27 78.13 0.27
Bribe amount (if >0) 7.67 7.40 0.22
% of officials accepting 85.71 77.27 0.42
% of citizens punishing 50.00 76.47 0.09
Punishment amount (if >0) 7.08 7.00 0.97
Australia Singapore
p-value
(iii) (Treatment 2) (Treatment 2)
% of firms bribing 87.27 91.30 0.62
Bribe amount (if >0) 7.67 7.67 1.00
% of officials accepting 85.71 84.62 0.91
% of citizens punishing 50.00 39.13 0.46
Punishment amount (if >0) 7.08 7.00 0.97
Indonesia Singapore
p-value
(iv) (Treatment 2) (Treatment 2)
% of firms bribing 78.13 91.30 0.20
Bribe amount (if >0) 7.40 7.67 0.38
% of officials accepting 77.27 84.62 0.53
% of citizens punishing 76.47 39.13 0.02
Punishment amount (if >0) 7.00 7.00 1.00

27
 Table 4B: Differences Between Females Across Countries

Australia India
p-value
(i) (Treatment 1) (Treatment 1)
% of firms bribing 95.12 92.59 0.62
Bribe amount (if >0) 7.82 7.18 0.01
% of officials accepting 82.86 89.66 0.35
% of citizens punishing 56.25 20.93 0.00
Punishment amount (if >0) 5.04 4.33 0.47
Australia Indonesia
p-value
(ii) (Treatment 2) (Treatment 2)
% of firms bribing 71.21 82.14 0.27
Bribe amount (if >0) 7.64 7.61 0.88
% of officials accepting 78.33 76.92 0.89
% of citizens punishing 68.63 70.00 0.91
Punishment amount (if >0) 5.57 4.29 0.28
Australia Singapore
p-value
(iii) (Treatment 2) (Treatment 2)
% of firms bribing 71.21 83.33 0.15
Bribe amount (if >0) 7.64 7.60 0.83
% of officials accepting 78.33 93.33 0.07
% of citizens punishing 68.63 48.15 0.08
Punishment amount (if >0) 5.57 7.38 0.13
Indonesia Singapore
p-value
(iv) (Treatment 2) (Treatment 2)
% of firms bribing 82.14 83.33 0.90
Bribe amount (if >0) 7.61 7.60 0.97
% of officials accepting 76.92 93.33 0.08
% of citizens punishing 70.00 48.15 0.14
Punishment amount (if >0) 4.29 7.38 0.04

28
 APPENDIX

Table A1: The 2006 Corruptions Perceptions Index

RANK COUNTRY SCORE

Finland
1. Iceland 9.6
New Zealand
4. Denmark 9.5
5. Singapore 9.4
…
Australia
9. 8.7
Netherlands
…
Austria
11. Luxembourg 8.6
United Kingdom
…
Belgium
20. Chile 7.3
USA
…
45 Italy 4.9
54 Greece 4.4
…
Brazil
China
Egypt
Ghana
70. India 3.3
Mexico
Peru
Saudi Arabia
Senegal
…
Azerbaijan
Burundi
Central African Republic
Ethiopia
130. 2.4
Indonesia
Papua New Guinea
Togo
Zimbabwe
…
163. Haiti 1.8

Source: Transparency International

29