Escolar Documentos
Profissional Documentos
Cultura Documentos
A Non-Proxied Empirical
Investigation of Cultures
Effect on Corruption
DEKUWMINI MORNAH AND RAYMOND J. MACDERMOTT
ABSTRACT
C 2018 W. Michael Hoffman Center for Business Ethics at Bentley University. Published by
V
Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington
Road, Oxford OX4 2DQ, UK. DOI: 10.1111/basr.12142
270 BUSINESS AND SOCIETY REVIEW
INTRODUCTION
O
n October 31, 2014, Blaise Compaore, the President of
Burkina Faso was forced to resign through violent demon-
strations after he attempted to flout the constitution and
extend his 27-year rule of the country. Corruption and manipula-
tion of the judiciary has been cited as the reason for the violent
uprising that forced him out of power. The Arab Spring which
began in 2010 ousted leaders in Tunisia, Egypt, and Yemen with
major protests and civil uprisings occurring in more than a dozen
other countries. In all these uprisings, political corruption is widely
speculated to be the main cause as demonstrated by Wikileaks dip-
lomatic cables.1 While corruption—defined as the misuse of public
office for private gain—is difficult to measure, perceptions of cor-
ruption are measurable. Perception of corruption in Burkina Faso,
Tunisia, Egypt, and Yemen as reported by Transparency Interna-
tional’s corruption index from 2005 to 2015 averaged between 7.0
and 8.2 on a scale of 0–10; where 0 is least corrupt and 10 most
corrupt.2 However, other countries in similar or worse circumstan-
ces and with high political corruption were relatively stable during
this period. What makes some countries react more strongly to
political corruption than others and why is this relevant for busi-
ness? All things being equal, the different reactionary effects from
countries of similar political corruption levels suggest differences in
cultural orientations are probable causes (Ashour 2006; Olken
2005; Seleim and Bonti, 2009).
In this article, we seek to empirically examine the question: “Can
cultural and social orientations explain differences in national cor-
ruption levels?” Which aspects of a country’s culture will grow cor-
ruption and which aspects stifle corruption? Using 3-year
averages, we construct a panel data set covering upwards of 80
countries and 15 years with three measures of corruption, mea-
sures of culture from the World Values Survey and Hofstede’s six
dimensions of culture as well as other control variables. We adopt
recent innovations in econometrics to assess the effect of cultural
MORNAH AND MACDERMOTT 271
The WVS began in 1981 as the European Values Survey. This first
wave, as they are called, included just 20 countries.4 Since then,
there have been four additional waves of surveys.5 Over the years,
these five waves of surveys have covered 99 countries.6
The WVS questions cover ecology, economy, education, emo-
tions, family, gender and sexuality, government and politics,
health, happiness, leisure and friends, morality, religion, society
and nation, and work.7 To measure culture, many studies have
selected four questions to give guidance on Trust, Obedience,
Respect, and Control.8
276 BUSINESS AND SOCIETY REVIEW
Data
Data for our analysis were collated from a wide range of sources. A
detailed description of the variables employed and their sources
are provided in the Appendix.
Corruption (Corrupt): The principal measure of corruption used
is that of Transparency International CPI. This index reflects the
level of perceived corruption in a country. This measure is most
commonly used in empirical work largely because it combines
many different sources of data used in other measures of corrup-
tion. Other measures of corruption across countries considered in
this study include the international country risk guide (ICRG) cor-
ruption index by the PRS group, and Control of Corruption Index
(WB Corrupt) by the World Bank. The ICRG corruption index as
defined by the PRS group is a “measure of political corruption
280 BUSINESS AND SOCIETY REVIEW
Other Variables
Baseline Model
Estimation
Corruption
Corruption Perception 898 4.221529 2.140524 0.8 9.93
Index (CPI)
Corruption By ICRG 1,172 3.040908 1.363201 0 6
Corruption By 1,158 4.943226 2.897635 0 10
World Bank
Culture: WVS
Respect 176 66.77841 14.41039 14.4 93.6
Obedience 176 35.99574 16.91025 2.3 81.7
Trust 176 30.79545 15.81485 2.8 74.2
Control 170 68.15294 7.274834 44 84
Culture Hofstede
Power Distance 804 58.65672 21.73671 11 104
Individualism 804 44.32836 23.91541 6 91
Masculinity 804 48.8806 19.97173 5 110
Uncertainty Avoidance 804 67.43284 23.74222 8 112
Long Term Orientation 1,080 45.326 23.68049 0 100
Indulgence 1,080 45.6128 22.41679 0 100
Other Control Variables
Government 1,338 2147.103 3341.689 8 233,95
Expenditure
Per Capita
Government Size 1,357 0.1666093 0.090335 0 1
Gdp Per Capita 1,975 10021.64 16485.51 68 152,972
Natural Resource 2,027 8.338962 14.91753 0 179
Rent (Total)
Secondary School 1,691 66.50911 32.77056 3 157
Enrollment
Tertiary Enrollment 1,551 22.9121 21.68757 0 113
Ethnic Diversity 2,196 0.4385361 0.2583657 0 0.93
Property Rights 2,112 43.46591 23.77486 5 95
control for the lag of corruption (Table 6), Respect loses significance
in three of the six regressions.
Hofstede Results
CPI 1
ICRG 0.7611 1
WB 0.9122 0.6773 1
MORNAH AND MACDERMOTT
Note: The number(s) in parenthesis indicate the waves. For instance, Zimbabwe
took part only in Wave 4 while the South Korea has been in all five waves.
Robust Standard errors in parentheses *P < 0.10. **P < 0.05. ***P < 0.01.
BUSINESS AND SOCIETY REVIEW
TABLE 6 Model 2: WVS Culture and Corruption OLS and GMM with Lagged Dependent Variable
CPI ICRG WB
Robust Standard errors in parentheses *P < 0.10. **P < 0.05. ***P < 0.01.
289
TABLE 7 Model 1: Hofstede Culture and Corruption OLS and GMM: No Other Control Variables
290
Included
CPI ICRG WB
Robust Standard errors in parentheses *P < 0.10. **P < 0.05. ***P < 0.01.
TABLE 8 MODEL 2: Hofstede Culture and Corruption OLS and GMM with Lag Dependent Variable
CPI ICRG WB
Robust Standard errors in parentheses *P < 0.10. **P < 0.05. ***P < 0.01.
292 BUSINESS AND SOCIETY REVIEW
We set out to find out whether culture has any effect on corruption
across countries using the WVS and Hofstede’s dimensions of cul-
ture and what the implications could be for business and policy.
We find that indeed, some aspects of culture have significant and
robust effects on corruption. With the WVS, by and large, Trust in
society, Control over one’s destiny and Respect for others decrease
corruption while Obedience increases corruption. Under the Hof-
stede dimensions, in most cases, societies characterized by high
Power Distance and Masculinity are more corrupt. Individualism,
Long Term Orientation, and Indulgence, in most cases, have a neg-
ative effect on corruption. The effect of Uncertainty Avoidance is
unclear.
MORNAH AND MACDERMOTT 293
NOTES
7. Waves 1 through 5 had 381, 744, 236, 245, and 259 questions, respec-
tively. The four key questions utilized in the current study were included in all
five waves.
8. For each metric, a higher score is interpreted as a greater indication of
that quality for the culture. For instance, a country that scores high on the Tol-
erance metric is believed to be more tolerant (of both the people and their ideas)
than others.
9. The other qualities were independence, hard work, feeling of responsi-
bility, imagination, thrift/saving money determination/perseverance, reli-
gious faith, and unselfishness.
10. We multiply the score by 10 to make it comparable to the other three
culture measures.
REFERENCES
APPENDIX
Trust [2] Percentage of people who While there may be concern that
responded “Most people can in a more trusting society,
be trusted” when asked individuals are likely to be
“Generally speaking, would more easily cheated, we would
you say that most people can expect less corruption in more
be trusted or that you need trusting cultures. This metric
to be very careful in dealing is a self-assessment of their
with people?” society and likely the strongest
indicator of corruption. If cor-
ruption were an issue, the
respondents would likely indi-
cate a lack of trust.
Obedience [1] Percentage of people who In more obedient societies, indi-
included obedience as an viduals are more likely to sub-
especially important quality mit to another’s authority
that children can be encour- which would encourage
aged to learn at home. corruption.
Respect [2] Percentage of people who We would expect that in societies
included tolerance and where respect for others is a
respect for other people as priority, corruption would be
an especially important qual- less common. Here, the ratio-
ity that children can be nale comes from the authority
encouraged to learn at home. figure who would be less likely
to take advantage of his/her
fellow citizens given the respect
s/he has for them.
Control [2] Average score of respondents to In societies where people feel
the question “Some people they have little control, we
feel they have completely free would expect more corruption
choice and control over their as these individuals would
lives, while other people feel likely feel powerless in the face
that what they do have no of authority. Thus, with more
real effect on what happens to control we would expect less
them where 1 means ‘none at corruption.
all’ and 10 means ‘a great
deal’ to indicate how much
freedom of choice and control
you feel you have over the
way your life turns out.” This
has been re-scaled to an
index of 0–100.
MORNAH AND MACDERMOTT 299
TABLE A2 (continued)
Corruption
Corruption Perception Index (CPI) The Corruption Perception Index Transparency International
(CPI) index by Transparency Inter- Corruption perception indi-
national measures the level of ces (various years)
perceived corruption in a country.
International Country Risk Guide (ICRG) Measures perception of corruption The PRS Group
World Bank Control of Corruption (WB) This is an index indicating the World Bank Database (World
degree to which countries are Development Indicators)
MORNAH AND MACDERMOTT
fighting corruption.
Culture measures
WVS (Trust, Obedience, Respect, Control) As defined in Appendix Table A1 World Values Survey (Inglehart
Hofstede (All six dimensions) As defined in Appendix Table A2 et al. 2014)
Hofstede (2001)
Other variables
GDP Per Capita
Education (Tertiary) WDI
Ethnic Diversity WDI
Genetic Distance Alesina
Natural Resource Endowment WDI
Property Rights The PRS Group
Trade Openness WDI
301