Note: how to compute an index of corruption

Transcription

Note: how to compute an index of corruption
A New Cross-National Measure of Corruption
Laarni Escresa and Lucio Picci*
April 15, 2015
Abstract
We propose a new measure of cross-national corruption, the Public Administration
Corruption Index (PACI), based on the geographic distribution of public officials involved in
cross-border corruption cases. We consider to what extent differences between the PACI and
perception based measures are driven by systematic factors, and conclude that they are not. As
more data on cases of cross-border bribes become available, the PACI will provide an
increasingly valid cross-national measure of corruption.
Keywords: Corruption, Measures of corruption, Judicial statistics
J.E.L. Codes: H11, H50, D73, C18, C43, F53, F55
*
Laarni Escresa: School of Economics, University of the Philippines. Lucio Picci
(corresponding author): Department of Economics, University of Bologna, Strada Maggiore
45, I-40125 Bologna, Italy. Email: [email protected]. We would like to thank, for their
comments on a previous version of this paper, Rajeev Goel, Johann Graf Lambsdorff, Miriam
Golden, Jerg Gutmann, Robert Klitgaard, Nicholas McLean, Angela Reitmaier, Susan RoseAckerman, participants to a workshop at the Inter-American Development Bank in Washington,
D.C., to the European Political Science Association 4th Annual Conference in Edinburgh, and
to the 10th Annual Conference of the Italian Society of Law and Economics in Rome. Laarni
Escresa acknowledges research funding from the Deutsche Forschungsgemeinschaft (German
Research Foundation). All Website contents have been accessed on April 15, 2015.
1
Introduction
We present a new measure of corruption, the Public Administration Corruption Index
(PACI), constructed using data on cross-border corruption cases. Our proposal is motivated by the
need to find viable alternatives to currently available cross-national measures of corruption, which,
despite their shortcomings, have been and are being extensively used in academia,1 within policy
circles, and in the public debate at large.
There are two types of cross-national measures of corruption currently available. The most
widely used are perception-based, such as the Transparency International Corruption-Perception
Index (hence, TI-CPI; Transparency International 2012, Saisana and Saltelli 2012) or the World
Bank Control of Corruption Indicator (hence, WB-CCI; Kaufmann, Kraay and Mastruzzi 2009).
Their shortcomings are well-known. On the one hand, perceptions may be weakly correlated with
actual experiences of corruption (Razafindrakoto and Roubaud, 2010, Seligson, 2006, and Olken
2009). Also, what is being assessed often is not precisely defined, and the aggregation
methodology based on different sets of data, that is sometimes performed, complicates the
interpretation of results. A second set of cross-national measures are based on surveys assessing
first-hand experience of corruption.2 These are rather costly endeavours, and their results are
affected by respondents’ reticence in answering questions related to their participation in corrupt
activities.3
At the outset, our proposal to use judicial statistics to develop a cross-national measure of
corruption may appear unseemly. Even setting aside differences in legal definitions across
jurisdictions, the fraction of observed corrupt transactions, relative to their actual number, is
unknown, and varies widely across countries. These differences could be so important, as to even
imply a negative correlation between the actual and observed corrupt transactions. After all, where
1
On the use of perception-based measures to inquire into the nature, causes and
consequences of corruption, see Lambsdorff 2006 and 2007, Treisman 2007, and Rose-Ackerman
1999.
2
A well-known example is Transparency International “Global Corruption Barometer” (TI-
GCB). See http://www.transparency.org/research/gcb/overview.
3
See for instance Treisman 2007, Clausen et al. 2011, and Kraay and Murell 2013.
2
corruption is endemic, the judiciary may also be corrupt or vulnerable to threats (Van Aaken et al.
2010). 4
However, judicial statistics on cross-border corruption, which refer to corrupt transactions
between firms headquartered in a particular country (hence, the “headquarters country”) and public
officials elsewhere (the “foreign country”), permit to compute a valid cross-national corruption
index. By a valid index we mean one that is increasing in the probability that a transaction is
corrupt. The intuition behind our reasoning is quite simple, as the following example illustrates.
The instances of corruption involving US firms caught bribing public officials abroad is not
informative of the level of corruption in the US. However, the distribution of these cases, with
respect to the nationality of the foreign public officials involved, is informative of the relative level
of corruption abroad. For example, if out of all the cases concerning US firms, we observed that
half of them involved Chinese public officials, we would have some evidence that the level of
public sector corruption in China is relatively high. Obviously, such a conclusion should take into
account the intensity of bilateral transactions – the United States interact more often with China
than, say, Denmark. As a suitable proxy for the number of bilateral transactions which could be
amenable to corruption, we use exports from the headquarters to the foreign country. Going
beyond this simple intuition, measure uses information on the spatial distribution of cases enforced
in a given country to evaluate levels of corruption in all other countries. Moreover, it considers
cases arising not just in a single jurisdiction, but in all relevant ones.5
Our data derives from the reporting of cross-border corruption associated with the
criminalization of foreign bribery in a number of jurisdictions. With the Foreign Corrupt Practices
Act (FCPA) of 1977, the United States were the first country to declare foreign bribery a crime.
On 15 February 1999 the OECD Anti-Bribery Convention came into force, requiring signatory
countries to put in place adequate legislation to combat foreign bribery in their own jurisdictions.
4
Judicial statistics have been used to measure corruption in studies at the sub-national level,
where the assumption of spatially homogenous enforcement is plausible. See for instance, Glaeser
and Saks 2006, Goel and Nelson 2011, Fisman and Gatti 2002, Alt and Lassen 2012, for the US;
Chang et al. 2010, and Golden and Picci 2008.
5
To the best of our knowledge, we are the first to propose an index of corruption based on
cross-border occurrences of corruption. McLean 2012 contemplates the possibility of computing
such an index; see in particular his Figure 1
3
To date, 41 countries have signed the Convention, and hundreds of cases have been investigated
worldwide.6
The number of observed cases, however, is not sufficient to compute a reliable yearly
measure, and in what follows we bundle together corruption cases over a 15-year period, from
1998 until 2012. However, as more countries sign the OECD Anti-Bribery Convention, and as
more cases of cross-border corruption are exposed, we foresee the possibility of computing the
PACI for shorter periods of time. Increased data availability would also allow for the computation
of more granular measures, treating separately different sectors of the economy, or different public
administrations.
We describe the PACI in the next section. In Section 3 we introduce our dataset and we
compute the index. In Section 4 we show that the PACI is highly correlated with existing
perception-based measures of corruption, and we consider the determinants of the observed
differences. Section 5 discusses the available evidence regarding the assumptions that ensure the
validity of the index, which are described in detail in Appendix A. Section 6 concludes.
2. The Public Administration Corruption Index
To illustrate the PACI, we need first to distinguish cases of cross-border corruption,
(involving firms from the headquarters country i bribing public officials in the foreign country j),
depending on where they were enforced first, meaning, the country whose judiciary was the first
to take action. The vast majority of cross-border corruption occurrences were first enforced in the
headquarters´ country, a majority of which are developed countries. This is expected, as the FCPA
and the OECD Anti-Bribery Convention came into place precisely to counter the lax policing of
corruption in many of the foreign countries where multinational firms operate. It may also happen
that cases are first enforced in a third country, mainly because of the expansive interpretation of
US jurisdiction, which includes companies registered in the Security and Exchange Commission
(SEC) and, more generally, entities carrying business in the US.7
6
See
Carr
and
Outhwaite
2008
and
http://www.oecd.org/daf/anti-
bribery/oecdantibriberyconvention.htm.
7
After a case is first enforced in the headquarters country, or in a third-country jurisdiction,
it may also fall under the radar of the foreign country’s judiciary.
4
Having made this important distinction, we further describe the intuition behind our index
by means of a simple example, which we illustrate in Table 1. Our data (to be presented in detail
in Section 3) indicate that, between 1998 and 2012, there were 315 cases of alleged cross-border
corruption, first enforced in the US and involving firms headquartered in the same country. If the
public officials of all countries trading with the US were equally corrupt, we could expect those
315 occurrences to be distributed according to the number of bilateral transactions, which we proxy
with bilateral trade flows. Exports to China represented 9.29% of total US exports, implying an
expected number of 29.26 cases. However, the actual observed occurrences of corruption is 65, or
20.63% of the total. That is, we observe 2.22 times more cases of cross-border corruption involving
Chinese public officials, than we would expect if the total number of cases first enforced in the
US, and involving firms headquartered there, were distributed geographically according to the
countries’ shares in US exports (see the last column of Table 1). We interpret this as evidence that
the level of corruption among public officials in China is higher than the average of all US trading
partners.
[Table 1 about here]
Table 1 also shows the same comparison for Austria as the foreign country. It was the
destination of 0.32% of total US exports, and it had one reported case of bilateral corruption,
corresponding to 0.31% of the total. Based on these figures, the level of corruption involving
Austrian public officials appear to be close to average, with the ratio at 1.02.
Just as we compare China and Austria from the point of view of the US, we may also
consider occurrences of corruption from the point of view of other countries. Consider Germany
as the headquarters and country of first enforcement. During the period under consideration, there
were four cases involving Chinese public officials out of a total of 56 documented cases. This
would also lead us to conclude that China’s public officials are more corrupt than the average trade
partners of Germany – see the last column of Table 1. In the case of Austria, we instead observe
slightly fewer cases than its share of German exports would lead us to expect.
The PACI is based not just on one or two “points of observation”, but it appropriately
aggregates cases involving firms from all countries. Note that this reasoning does not need any
assumption with respect to the probability of enforcement of cases in a given country. In fact, we
just noted that there were 315 cases of corruption of firms headquartered in the US, and only 56
in Germany (with the respective countries acting as first enforcers). We do not use this information
to determine the level of corruption in either country. It is the distribution of these cases (regardless
5
of how many there are) abroad which conveys information on the level of corruption in the
countries where the bribery of the foreign officials takes place.
The PACI
We indicate with cases _ obs _ HQij those cases of corruption by firms in i, of public
officials in j, and that are first enforced in the headquarters’ country i. The PACIz compares the
total number of those cases, with the expected number of corrupt transactions that would be
observed if their spatial distribution reflected bilateral trade shares between the headquarters
countries and z:
N
 cases _ obs _ HQ
i, z
PACI z =
 100 , with i≠z.
i=1
N
 E(cases _ obs _ HQ
i, z
(1)
)
i=1
The numerator,  cases _ obs _ HQi , z , is the total number of observed corrupt exchanges
between officials from country z and firms from all other headquarters countries, first enforced in
those other countries. The denominator is the total number of like cases which we would observe,
if cases of corruption were distributed according to the ratio of exports of country i to z ( X iz ),
with respect to the total amount of country i exports to the rest of the world:
N
N
i 1
i 1
 E(cases _ obs _ HQi, z ) = 
N
X iz
  cases _ obs _ HQi , j .
N
X
(2)
j=1
ij
j=1
The denominator may be interpreted as the total number of cross-border cases involving country
z public officials, and first enforced elsewhere, that we’d expect to observe if the level of
corruption of public officials were the same in all countries. If the actual and expected values are
equal, the PACI equals 100. The lowest value that the index may take is zero, which obtains when
no corrupt cases (first enforced in all headquarters countries) are observed in country z.
The composite PACI
The PACI zALL , or “composite PACI,” follows the same logic as the simple PACI z , but
considers cases that were first enforced not only in all headquarters countries, but also in other
6
countries, with the exception of z. We denote cases that are first enforced in a third country w as
cases _ obs _ OTHijw (where i,j ≠ w) and. The index is as follows:
N
D
 cases _ obs _ HQ
i, z
PACI zALL =
i=1
 100
w=1 i=1
D N
N
 E(cases _ obs _ HQ
i, z
i 1
N
+  cases _ obs _ OTH iw, z
) +  E(cases _ obs _ OTH
w=1 i 1
w
i, z
)
(3)
with i ≠,j, w ≠i,j, D is the number of jurisdictions that served as a third-country first enforcer, and
where for the denominator the following holds:
D
N
 E(cases _ obs _ OTH
w=1 i 1
D
w
z ,i
N
) = 
w=1 i 1
N
X i,z
 cases _ obs _ OTH
N
X
w
i, j
j=1
i, j
j=1
The interpretation of the PACI zALL is conceptually the same as that of the PACI z (equation 1), but
it considers all available cases of observed corruption, first enforced either in the country where
firms are headquartered, or in third country jurisdictions. This is the version of the index that we
compute, and that we present in Table 4 below.
Conditional probability of observing zero case
We are interested in expressing the varying precision with which the PACI measures
corruption, which is influenced by the total number of transactions, as represented by trade flows.
To fix ideas, consider the case of a foreign country that has no reported case of cross-border
corruption. We would interpret it as a signal that corruption is relatively low, but not necessarily
equal to zero. The strength of such a signal, in fact, depends on the number of cross-border
transactions, which we proxy with bilateral trade flows. To express this concept, we introduce the
“probability of observing zero cases, conditional on the probability of corruption being equal in
all foreign countries”:
N
Pr_zero _ caseszPACI = Pr(  cases _ obs _ HQi , z = 0 | Pr_corr _ FOz = c )
i 1
where Pr_ corr _ FO z is the underlying probability that a public officials accepts a bribe (see
Appendix A for a detailed definition) and c is a constant. The measure expresses the probability
that no case involving country z public officials (first enforced in the headquarters country) is
observed, when the true country z probability that a cross-border transaction involving its public
7
officials is corrupt is the same everywhere in the world. Pr_zero _ caseszPACI may be computed as
the product of individual probabilities:
Pr_ zero _ cases zPACI = i 1( Pr_ cases _ obs _ HQi , z = 0 | Pr_ corr _ FO z = c)
N
Each one of these probabilities, referring to a rare event, is described by a Poisson
distribution, and may be easily computed by making the condition Pr_corr _ FOz = c correspond
to an ideal situation where occurrences of corruption are distributed according to trade shares.
High values of Pr_zero _ caseszPACI indicate that the country in question has relatively few crossborder transactions – that is, that information is relatively scarce.
3. Computing the PACI
Using various sources, which we document in Appendix B, we collected information on
cases of cross-border corruption from 1977 until the end of 2012. However, we only use cases
from 1998 onward, corresponding to the 15-year post-Convention period. Of a total of 979 cases
for which we have detailed information8, 796 cases were first enforced either in the headquarters
country (569), in the US acting as a third country jurisdiction (177), or in other third country
jurisdictions (50). The remaining 183 cases that were first enforced in the foreign country were
not considered for the purpose of computing the PACI.
We coded each case according to the observed outcome, as “positive”, if the accused party
was either found to be guilty or, while not admitting guilt, accepted to pay a fine (as, in the case
8
As we document in Appendix B, the database that we use is the result of painstaking work,
and we believe that it resulted in the observation of the virtual totality of cases of cross-border
alleged corruption that were considered under some jurisdiction. A small number of cases could
not be considered for the computation of the PACI because of lack of information on the identity
of one of the two countries involved. We excluded all cases linked to the United Nations Oil-forFood Programme, allowing Iraq to sell oil in exchange for food, medicines and other humanitarian
goods, because of their peculiar origin. All the coding reflects information available on April 1,
2015. The raw data are available at http://www2.dse.unibo.it/picci/measure_corruption.html.
8
of a “consent to a cease-and-desist order” in the US);9 as “not positive”, if the case was eventually
dropped and no action was taken, or as “ongoing”, if no available evidence was found to determine
it as “positive” or “not positive.” We also recorded where it was enforced first, and both the
home-country of the corrupting agent (the headquarters country, or HQ), and the country where
public officials allegedly were corrupted (the foreign country, or FO). The term public official is
used in a broad sense, encompassing both bureaucrats and politicians.
Table 2 shows the number of cases by headquarters country. Of a total of 796 cases that
may be used for the purpose of computing the PACI, 444 are classified as positive, 272 as ongoing,
and the rest have either been dropped, or have resulted in an acquittal. Firms are headquartered in
40, mostly industrialized, countries. Among them, the US takes the lion’s share, reflecting both its
early adoption of the FCPA and the proactive stance taken by the Department of Justice and the
Securities and Exchange Commission after the ratification of the OECD Anti-Bribery Convention.
Germany, Britain and France follow in the list.
[Table 2 about here]
The set of countries in which public officials are at the receiving end of alleged bribes is
much wider, as Table 3 shows. At least one case is recorded in a total of 128 countries. China leads
the list, with 88 cases, followed by Nigeria.
[Table 3 about here]
ALL
To compute our index of public administration corruption, PACI z
, which we report in
Table 4, we consider all 796 cases, regardless of their outcome. The inclusion of all cases is based
on the relatively high evidentiary tests that trigger an investigation and a higher burden of proof
that would lead to a conviction as supported by the data and the literature. Thus, false negatives
are likely to be more numerous than false positives. However, to allow for a more agnostic view
on this issue, we also compute the index adopting more stringent standards. In particular, we
consider excluding cases that have been dropped or that have resulted in an acquittal, and also, we
compute the PACI excluding cases that involve administrations in charge of the health and the
9
It should be emphasized that coding the outcome of cases as positive does not imply
conviction or guilt on the part of the accused.
9
telecom sector, for reasons that we discuss in Section 5 below. The combination of these
alternatives results in the computation of four versions of the PACI, whose reciprocal Spearman
rank correlations range between 0.885 and 0.945, as indicated in the top-left part of Table 5, to be
discussed in its entirety below.
Table 4 also shows the probability of observing zero cases, conditional on the level of
corruption to be the same in all countries, Pr_zero _ caseszPACI . We use this quantity to rank those
countries that have no observed cases of corruption that were first enforced abroad. Compare
Canada and Finland, countries for which we report zero cases of corruption (first enforced abroad)
of their public officials. For Finland, the probability to observe zero cases of corruption, if in fact
the level of corruption were the same everywhere (if the number of cases were distributed
according to trade shares) is 1.3%. Being a middle-sized economy, it would be rather unlikely not
to observe any cases, as it happens, while the true level of corruption of its public officials were
equal to the world average. For Canada, Pr_zero _ caseszPACI = 0.0000: the signal provided by the
PACI, indicating that the level of corruption in Canada is low, in this case is very strong, reflecting
the fact that Canada, a bigger country, generates more information than Finland.
[Table 4 about here]
Pr_zero _ caseszPACI provides a useful indication of the precision of the PACI also for
countries with at least one case of corruption on record. For smaller countries, the signal that the
PACI provides is rather noisy, because of the rareness of the observed corruption events. For
example, for Tunisia the probability that we observed zero cases, in a situation where the level of
corruption were the same everywhere in the world, would be equal to 0.31. In fact, we actually
observe a total of 2 cases, close to twice as many as we would expect if cases were distributed
according to trade shares, with a resulting PACI of about 170 – and a ranking measure for Tunisia
which is similar to those indicated by the WB-CCI and the TI-CPI (see the last two columns of
Table 4).
4. Comparing the PACI with the leading perception-based measures
The last two columns of Table 4 show the difference in rank between the PACI and the
WB-CCI and TI-CPI. A positive value indicates that a given country is considered to be relatively
10
more corrupt based on the PACI. In most cases, differences are rather modest, but for a few
countries results diverge considerably.
Table 5 summarizes these information by showing the Spearman rank correlations between
the different versions of the PACI and the perception-based alternatives, for the year 2005. The
rank correlations of our preferred index (indicated as PACI1 –the same of Table 4) with those
measures are rather high, and typically above 0.7.10
[Table 5 about here]
While the ranking of the PACI is similar to those of the two most popular perception-based
indexes, the scales are very different, as Figure 1 clarifies. Panel A shows a scatter diagram of our
index together with the WB-CCI for the year 2005, while Panel B shows the same indexes, with
PACI log-transformed. The Pearson correlation of log(PACI) and WB-CCI in 2005 is quite high
(equal to -0.841).
[Figure 1 about here]
We would like to examine whether the observed differences between the PACI and
perception-based measures of corruption are systematic and, to the extent that they are, what we
may learn from the nature of their determinants. We consider the residual of a linear regression
between the log of the PACI and the WB-CCI (as in Figure 1.B) or the TI-CPI, and denote them as
“measurement residuals” mesresWB-CCI and mesresTI-CPI, respectively. Negative (positive) values
indicate that a given country appears to be less (more) corrupt according to the PACI, compared
to the perception-based index used.
We first consider whether these differences are correlated with Pr_zero _ caseszPACI . If the
precision of the PACI deteriorated significantly for smaller countries, while such a phenomenon
did not occur (or it occurred to a lesser degree) for the perception-based measures, then we’d
expect the absolute value of the measurement residuals to be positively correlated with
Pr_zero _ caseszPACI . This is not the case: the correlation between the absolute values of mesresWBCCI
10
and mesresTI-CPI, and Pr_zero _ caseszPACI , is insignificant, and equal to -0.0524 and -0.1001
Note that more corruption corresponds to lower values of the WB-CCI and the TI-CPI
indexes, but to higher values for the PACI and the TI-GCB.
11
respectively. We take this result as prima facie evidence that any problem which the PACI may
have due to the lack of precision in measuring corruption for small countries, would also be shared
by the two perception-based alternatives.
To proceed further, we select a set of variables which may be linked to biases in one or the
other measure.11 We consider variables of an economic and demographic nature: per capita GDP
based on purchasing-power-parity (gdp_cap), population (pop), and the ratio of public expenditure
over GDP (r_g/gdp). We consider variables that express the ease with which publicly relevant
information is generated and debated and, more generally, the democratic characteristics of a
country. We include emp_rights, an index of empowerment rights; free_press and free_speech,
measuring respectively the freedom of the press and guarantees to free speech; a measure of voice
and accountability, voice_acc; an index of democratization, democ; a measure of “checks and
balances”, checks; and one of political stability, stability. To test whether the observed residuals
follow some recognizable geographic pattern, we also consider a set of geographic dummies. Table
6 shows pairwise correlations between mesresWB-CCI, mesresTI-CPI, and these variables. We also
present results excluding the ten biggest (in absolute value) measurement residuals,, roughly
corresponding to ten percent of the available observations.
[Table 6 here]
Two of the geographic dummies are significantly correlated with both types of
measurement residuals. Countries in the American continent on average appear to be less corrupt
when we look at the PACI, compared to either one of the perception-based measures, while the
opposite holds for countries in Africa and in the Middle East. We observe further that populous
countries fare better according to the PACI. Countries that are more democratic and have stronger
checks and balances appear to be less corrupt on average when the PACI is used. The significance
of some of the other variables depends on which measurement residual we consider, and on
whether we include outliers. In interpreting the results, note that the squared estimated correlation
coefficient represents the fraction of the variance of the measurement residuals that is explained
by a given variable (the R2 of the bivariate Ordinary Least Squares regression). We see that such
11
A detailed description of the variables that we use is in Appendix B.
12
fraction is always rather small, even when the estimated correlation coefficient is statistically
significant.
To go beyond simple bivariate correlations, we resort to the multivariate regression of
Table 7, where the dependent variable is either mesresWB-CCI or mesresTI-CPI , and the regressors
consist of all the explanatory variables described above.12 As in the bivariate analysis, we also
report results obtained when excluding the ten observations of the dependent variable having the
greatest absolute value.
[Table 7 here]
Only few of the explanatory variables considered are statistically significant, and jointly
all the regressors only explain between 30 and 35 percent of the total variability of the
measurement residuals. We still find a significant negative effect for the dummy variable
pertaining to the American continent, in three out of four cases. The effect of the dummy variable
for Africa and the Middle East, which we detected in the bivariate analysis, is now insignificant.
Countries where the index of democratization (democ) is higher appear to be less corrupt according
to the PACI. The same applies to countries having a high share of public expenditure over GDP
(R_g/gdp). We also detect, in two out of four cases, a significant effect of population. Overall,
differences in the PACI with respect to the two leading perception-based cross-national measures
of corruption appear to be rather idiosyncratic, at least with respect to the set of factors that we
considered.
5. The validity of the PACI
In this section we discuss the assumptions necessary for the PACI to be a valid measure
of corruption. While in Appendix A we present these assumptions formally, to show how they
imply index validity, here we focus on the one hand on their overall meaning and, most
importantly, we discuss the extent to which they may hold in practice.
12
We exclude the dummy for Europe and Central Asia, to avoid the dummy variable trap.
The estimated coefficients of the other dummies should then be interpreted as the estimated effect
relative to that reference group of countries.
13
The first assumption states that the probability of observing a corrupt transaction involving
firms from country i and public officials in country j, that are first enforced in country i, does not
depend on the identity of country j. It implies that the judiciary, when deciding which cases to
pursue, does not “discriminate” based on the foreign officials´ country of origin. The possibility
of assessing corruption in the foreign country by looking at the geographical distribution of cases
first enforced elsewhere, hinges on this key assumption. Whether this assumption holds is an
empirical question. McLean 2012 assesses the relevance of foreign policy considerations and of
opportunities for enforcement cooperation in determining the geographic distribution of FCPA
cases, and finds that bilateral frameworks for securities regulatory and enforcement cooperation
appear to be associated with higher levels of FCPA enforcement. The magnitude of the effect is
however rather modest. McLean does not find any effect of other candidate explanatory variables
that he considers. Choi and Davis 2012 also find at most modest effects that would indicate
departure from our Assumption 1.
The assumption could also be violated in case the probability of detecting corruption cases
depended on the conditions surrounding freedom of expression and information in the foreign
country. However, our data indicate that most cases are first enforced and based on evidence
gathered in the headquarters countries. Further, supposing that the relevant perception-based
measure is unbiased, then we would expect variables that capture ease of expression and
circulation of information in the foreign country to positively affect the measurement residuals
introduced in the previous sections (implying that, in those cases, the PACI would signal a higher
level of corruption compared to its perception-based counterpart). We considered three variables
expressing various dimension of the ease of expression and circulation of information (Free_press,
Free_speech and Voice_acc - see Appendix B for a description) and the results of both bivariate
and multivariate analysis (Table 6 and 7) indicate statistical significance only in very few cases,
and at most only very modest effects.13
Assumptions 2 and 3 in the appendix are rather technical, and describe how the probability
of offering a bribe, or accepting a bribe when offered, may depend on the level of corruption in
the other country. Of more interest to us is assumption 4, establishing that the number of cross-
13
The probability of detection and taking action may also be sector-specific. For example,
corruption in arms trade is likely to be more difficult to detect, as national security concerns may
constrain the actions available to the judiciary (see also Rose-Ackerman 1999). To address such
issues, the PACI could be computed separately for different sectors of the economy.
14
border transactions is proportional to bilateral trade flows. An alternative proxy of cross-border
transaction would be Foreign Direct Investments (FDI) (as in Mclean 2012 and Choi and Davis
2012). However, many transactions are not reflected in FDI flows, nor stocks, and FDI eventually
enables trade flows between the countries involved. For these reasons, we believe that our choice
of proxy is more appropriate.
Differences in the scope of the public sector across countries would also be a motive of
concern. Consider the case of a pharmaceutical firm headquartered in country A successfully
bribing employees in hospitals in country B and C in order to sell its products. Assume that country
B has a public health system, so that the bribery qualifies as a cross-border occurrence of
corruption of a foreign public official, whereas in country C hospitals are private, so that the act
qualifies as private corruption and, as such, is not included in our dataset. At first sight, neglecting
this difference may lead to underestimating the level of corruption in country B with respect to C,
since ceteris paribus we would observe more corruption in B than in C because public officials in
B have a wider set of responsibilities. Apparently, in this case trade flows would not be a good
proxy of the cross-national exchanges involving public officials in a given country – for each dollar
of imports, there would be more interactions with public officials in B than in C. To account for
this issue, we have computed the PACI disregarding cases involving procurement in the health and
the telecommunication sector (as opposed to transactions involving regulatory bodies in those
sectors, since they are invariably public). Arguably, these are the main sectors where we can
observe variation across countries with respect to the extent of government activity in the
economy.14 The results of Table 5 indicate that these different choices deliver similar results.15
14
Alternatively, we could correct the number of cases for the relative size of the public sector
in the different countries. However, the concept of public sector is in itself blurred, particularly in
some countries. At a more general level, we stress that the decision on how to account for the
different reach of public sectors worldwide depends on which position the researcher takes on
conceptual issues that in themselves are debatable.
15
If our interest is in measuring the magnitude of public sector corruption, instead of its
frequency, then cross-country variations in the scope of activities of the public sector should not
be a cause of worry. The scope of government is also arguably endogenous over the long-run,
since corrupt elites have an interest in maintaining and possibly expanding their reach.
15
6. Discussion
The PACI reflects a narrow definition of corruption: the propensity of public officials to
accept bribes from foreign firms.16 In naming our index a general “public administration” measure
of corruption, we implicitly assumed that what is observed by means of cross-country corruption
statistics is also informative of the level of corruption in the public administration as a whole.
Obviously, this assumption may be put to question, and we know that levels of corruption may
vary sensibly across public administrations within a given country. We argue that adopting narrow
definitions of phenomena of interest is in fact very desirable, since it allows for the testability of
the assumptions on which rests any extensive interpretation of the resulting measures.
For this reason, the PACI represents a welcome departure from most indices of governance
currently available, for which “sometimes it is not clear what precisely is being measured,
rendering questionable the validity of at least some of the proxies” (Klitgaard and Light 1998).
Such a state of affairs may reflect a situation “experienced in many ‘new areas’ of the social
sciences: an explosion of measures, with little progress toward theoretical clarity or practical
utility” (Klitgaard et al. 2005, p.414). We believe that in the future, as open data become more
widely available, there will be a greater abundance of measures of governance which are based on
hard data and not on perceptions, and that are as narrowly and precisely defined as ours (see the
discussion in Picci 2011, pp 117-119). We are also convinced that such a development, in turn,
will facilitate progresses towards greater theoretical clarity.
Data availability poses certain constraints to the usefulness of the PACI, and it is the reason
why we illustrated it for a 15-year period. When comparing this characteristic of the PACI with,
for example, the TI-CPI, we should note that Transparency International has warned against
comparisons of that index across time, due to year-by-year changes in the methodology and
country coverage. So, to some extent, the advantage of having a yearly measure is only apparent.17
16
A Bribe Payer’s Corruption Index (BPCI) may also be computed along the same lines as
the PACI, but using cases first enforced in the foreign country. However, limited data is an obstacle
towards its computation.
17
Recent changes in the way the TI-CPI is computed should assure that “[it] will better
capture changes in perception of corruption in the public sector of [a given] country over time.
However, due to the update in the methodology, 2011 CPI scores are not comparable with CPI
2012 scores” (Transparency International 2012).
16
However, as more countries join the OECD Anti-Bribery Convention, more cases are likely to be
reported, adding precision to the PACI and allowing to compute it for shorter intervals of time.
Also, more data would permit to compute the PACI separately for different sectors, and the
resulting different measures could eventually be aggregated into a general index.
We showed that, under a set of assumptions, the PACI is valid, in the sense that it takes a
higher value for countries for which the probability that a transaction is corrupt is higher. Validity
of the index is sufficient to deliver a correct ranking of countries. In a previous version of this
paper, we showed that under a further assumption, the PACI represents the probability that a
transaction is corrupt, relative to a world average. So that, for example, if the PACI equals 200 for
a given country, it would imply that the probability of corruption for that country is twice a world
average. We leave it fort the future to analyse to what extent such interpretation of the PACI would
be warranted in practice. A better understanding of this issue would be particularly welcome,
considering that current cross-national measures of corruption are wanting in this respect. For
example, based on the 2014 TI-CPI, Germany and Turkey have scores of 79 and 45, and ranks 12th
and 64th the list of countries respectively. The differences imply that the (perceived) level of
corruption in Turkey is considerably higher than in Germany, but they are not amenable to any
interpretation regarding how much more corruption there is in the latter than the former.
Under this light we also interpret our finding that our index is highly correlated with the
main perception-based measures of corruption, but it has a different scale. We showed that the
scales become comparable once we take the log of the PACI. If indeed the PACI approximated the
probability that a transaction is corrupt with respect to a world average, this would imply that the
exponential of the WB-CCI and of the TI-CPI are interpretable as being roughly proportional to
levels of corruption. Further work in this direction, that is, may also help clarify the scales of
already existing perception-based indicators.
We foresee other venues for future research. First, the availability of the PACI allows to
reassess our understanding of the causes and consequences of corruption. More data would allow
to compute the PACI for separate sectors of the economy. It would also be interesting to study the
differences among different versions of the same indexes, computed by focusing on different
jurisdictions – along the lines of the example of Table 1. They all measure the same concept of
corruption, so that we expect them to provide similar results. Any differences among them would
be explained by sampling error, but also, possibly, by the violation of one or more of the
maintained assumptions, the more so, the bigger are the differences. Tests for the validity on the
maintained assumptions could be developed by leveraging on the magnitude of such observed
distances.
17
Last, the intuition behind the use of judicial statistics that we have discussed in this paper
may be applied to other domains. The essential ingredient needed is to have data, generated in a
set of jurisdictions, conveying information on crimes committed in a given jurisdiction by actors
residing in a different one. We speculate that some types of financial crimes may possibly lend
themselves to a treatment according to the methodology which we have introduced in this paper.
18
Appendix A. Assumptions and validity of the PACI.
For a corrupt transaction to occur, both parties must be willing to engage in it. Firms
headquartered in country i may decide to offer bribes to public officials in the foreign country j,
with a probability that depends on characteristics of the foreign country. In particular, the
probability of offering a bribe may be higher if the perceived level of corruption in the foreign
country is high, since high levels of corruption would imply a lower risk of being caught, and a
higher social acceptability of bribery. The probability that a public official in the foreign country
accepts a bribe when offered one, may also depend on characteristics of the headquarters country.
For example, if the latter is known to be very proactive in enforcing cross-border corruption, public
officials may be deterred: the discovery of a corrupt act in the headquarters country may be
followed by an enforcement action in their home country, making them liable.
We define pr _ corr _ FO j , as the probability that a public official in country j accepts a
bribe, if it perceives that there is no risk of being caught following a case first enforced in the
headquarters country. The advantage of this concept is that, logically, it does not depend on
characteristics of the headquarters country, as it purely reflects the propensity of country j’s public
officials to accept bribe. We identify this probability as the level of corruption of public officials
in the foreign country. We define a measure of corruption to be valid, if it is monotonically
increasing in the level of corruption.
Definition: The PACI is valid iff: PACI z /  pr _ corr _ FOz  0 .
The expected number of corruption cases observed and enforced first in the headquarters
country i, involving public officials in the foreign country j, is determined as follows:
cases _ obs _ HQi , j = pr _ obs _ HQi , j  corr _ exchi , j
(Eq. A1)
where corr _ exchi , j is the number of occurrences of corruption involving firms headquartered in
country i and public officials in country j (which later we will equate to its expected value), and
pr _ obs _ HQi , j is the probability that the corrupt exchange is observed, and enforced first in the
headquarters country.
The expected number of corrupt exchanges is:
19
corr _ exchi , j = pr _ bribe _ HQi , j  pr _ bribe _ FOi , j  transactionsi , j
(Eq. A2)
where pr _ bribe _ HQi , j is the probability that a firm headquartered in country i proposes a bribe
to public officials in foreign country j, while pr _ bribe _ FOi , j is the probability that the public
official in j accepts the bribe offered by the firm headquartered in i, and transactionsi,j is the total
number of business transactions involving firms in country i and public officials in foreign country
j.
This formulation simply states that in order for a transaction to be corrupt, both parties have
to agree upon it. We rule out the possibility of extortion, so that the probability that a transaction
is corrupt is equal to a product of probabilities. Please note that we are not assuming statistical
independence between events, and in what follows we will explicitly consider the possibility that
the probabilities of offering and of accepting a bribe are interdependent.
Assumption 1
The probability that a corrupt transaction involving firms from country i and public
officials in country j is observed and first enforced in country i, does not depend on the identity of
the foreign country j:
pr _ obs _ HQi , j = pr _ obs _ HQi
This assumption is of key importance because it is the basis for using the geographic
distribution of the cases involving country i firms to infer levels of corruption in all other countries.
Assumption 2
The probability that for a given cross-border transaction, a firm headquartered in country
i offers a bribe to public officials in country j is as follows:
pr _ bribe _ HQi , j  pr _ corr _ HQi   ( pr _ corr _ FO j )
where  (.) is a continuous and differentiable monotonically increasing function, and  (0)  1 .
This Assumption describes how the probability that a firm proposes a bribe increases with
the level of corruption in the foreign country. We define pr _ corr _ HQi to be the probability that
a firm offers a bribe abroad if it perceives that there is no risk of being caught following a case
20
initiated in the foreign country. If such a risk is present, it serves as a disincentive for firms in
country i to propose bribe. The second multiplicative factor,  (.) is meant to capture such a
deterrence effect. It equals one when there is no perceived risk that a corrupt transaction will be
caught following a case first enforced in the foreign country, and increases with the level of
corruption in the foreign country.18
Assumption 3
The probability that a public official in country j accepts a bribe when offered one by a
firm headquartered in country i is as follows:
pr _ bribe _ FOi , j  pr _ corr _ FO j   ( pr _ corr _ HQi )
where  (.) is a continuous and differentiable function, and  (0)  1
Simmetrically with respect to Assumption 2, this assumption expresseses the probability
that a public official accepts a bribe, and how it may depend on the level of corruption in the
headquarters’ country. We define pr _ corr _ FO j as the probability that a public official accepts
a bribe, if it perceives that there is no risk of being caught following a case initiated in the foreign
country and “spilling over” into the domestic jurisdiction. We consider this as the “underlying
probability of corruption of public officials in the foreign country j.” Assumption 3 admits the
possibility that, all else being equal, public officials in a given country may be more wary of
accepting a bribe from a firm headquartered in a country which is very proactive in fighting crossborder corruption. It affirms that such a deterrence effect is functionally the same for all foreign
country: the function  (.) does not depend on j.
Assumption 4:
18
In fact, the argument of the function  (.) should more appropriately be taken as
expectations of levels of corruption in the foreign country. This, in particular, would lead to the
possibility of expectations being partly self-fulfilling: in a country with a (initially, possibly
undeserved) reputation for corruption, foreign firms would be prone to offer bribes more often,
leading to more corrupt transactions. Considering pr _ corr _ HQi instead of expectations is
justifiable if we assume that the latter depend monotonically from that underlying probability.
21
Bilateral transactions are proportional to the value of exports from country i to country j,
xij , according to a constant factor k: transactionsij = k  xij
This assumption makes the number of cross-border transaction depend on an observable
variable, bilateral trade.
Given Assumptions 1, 2, 3 and 4, equation (Eq. A1) becomes:
cases _ obs _ HQi , j = pr _ obs _ HQi  pr _ corr _ HQi   ( pr _ corr _ FO j ) 
 pr _ corr _ FO j   ( pr _ corr _ HQi )  k  xij
(Eq. A2)
Substituting this expression into the definition of the PACIz (Eq. 1), it is straightforward to
prove:
Proposition 1: the PACIz is valid: PACI z / pr _ corr _ FOz  0 .
Assumptions 1-4 guarantee that the PACI is valid, in the sense that higher levels of the
probability of corruption of foreign public officials, results in a higher value for the index
22
Appendix B. Data sources.
Corruption cases. We gathered all reported cases of cross-border corruption since the
adoption of the OECD Anti-Foreign Bribery Convention in 1998 to. The main sources of our data
are Trace International Compendium´s which maintains a database international anti-bribery
enforcement (http://www.traceinternational.org/compendium), US DOJ and SEC documents,
OECD (various years). We also consulted other database and publications, such as Shearman and
Sterling 2013, Transparency International 2009 and 2013, and Cheung et al. 2012: we also
considered various news sources and, among them, the Wall Street Journal Risk and Compliance
Journal (http://www.wsj.com/news/risk-compliance-journal), and also corruption blogs, such as
the “FCPA Blog” http://www.fcpablog.com/. Cases reported in multiple sources were laboriously
consolidated to avoid double counting. The reference period for each case is the year when the
bribe was allegedly paid, but in some instances this date had to be presumed from the available
data.
Exports: Barbieri and Keshk 2012. Values for the last three years under consideration
(2010-2012) have been set equal to the 2009 figure.
Gdp_cap: Gross domestic product based, per capita. Expressed in in PPP dollar per person.
Source: IMF-World Economic Outlook, October 2014
(variable name: PPPPC)
(http://www.imf.org/external/Pubs/ft/weo/2015/01/).
The secondary source of the following variables is the “Standard data” of the Quality of
Government Institute dataset (Teorell et al. 2013), which assembles various sources. We use the
May 2014 release of the dataset. For each variable we provide a brief description, which is taken
verbatim from that dataset’s codebook, and succinctly indicate the original source, which the
codebook allows to fully identify. “Variable name” indicates how the variable is denoted in the
codebook.
Pop: population. Source Heston, Summers and Haten 2012. Name of variable in dataset:
pwt_pop).
R_g/gdp: the share of public expenditure over GDP (Source Heston, Summers and Haten
2012. Original name of the variable: pwt_gsg)
Emp_right: empowerment rights index (Source: Cingranelli and Richard 2010. Name of
the variable in dataset: ciri_empinx_new). It is an “additive index constructed from the Foreign
23
Movement, Domestic Movement, Freedom of Speech, Freedom of Assembly & Association,
Workers’ Rights, Electoral Self-Determination, and Freedom of Religion indicators. It ranges from
0 (no government respect for these seven rights) to 14 (full government respect for these seven
rights).”
Free_press: freedom of the press index (Source: Freedom House. Name of the variable in
dataset: fh_fotpc3). “The press freedom index is computed by adding four component ratings:
Laws and regulations, Political pressures and controls, Economic Influences and Repressive
actions. The scale ranges from 0 (most free) to 100 (least free).”
Free_speech: freedom of speech (Source: Cingranelli and Richard 2010. See Teorell et al.
for details. Name of the variable in dataset: ciri_speech). “This variable indicates the extent to
which freedoms of speech and press are affected by government censorship, including ownership
of media outlets”.
Voice_acc: Voice and Accountability (The World Bank. Name of variable in dataset:
wbgi_vae). “Voice and Accountability” includes a number of indicators measuring various aspects
of the political process, civil liberties and political rights. These indicators measure the extent to
which citizens of a country are able to participate in the selection of governments. This category
also includes indicators measuring the independence of the media […].”
Checks: a measure of “checks and balances” (Source: Database of Political Intitutions
Name of variable in dataset: dpi_checks). “Equals 1 if the Legislative Index of Political
Competitiveness (dpi_lipc) or the Executive Index of Political Competitiveness (dpi_eipc) is less
than six. In countries where dpi_lipc and dpi_eipc are greater than or equal to six, dpi_checks is
incremented by one if there is a chief executive, by a further one if the chief executive is
competitively elected (dpi_eipc greater than six), and by a further one if the opposition controls
the legislature […]”.
Democ: index of democratization (Source: Vanhanen 2011. Name of variable in dataset:
van_index). “This index combines two basic dimensions of democracy – competition and
participation – measured as the percentage of votes not cast for the largest party (Competition)
times the percentage of the population who actually voted in the election (Participation). This
product is divided by 100 to form an index that in principle could vary from 0 (no democracy) to
100 (full democracy). (Empirically, however, the largest value is 49).”
Stability: political stability (The World Bank. Name of variable in dataset: wbgi_pse).
“Political Stability” combines several indicators which measure perceptions of the
likelihood that the government in power will be destabilized or overthrown by possibly
unconstitutional and/or violent means, including domestic violence and terrorism.
24
25
Tables and Figures
Table 1. An illustrative example.
HQ
FO
(1)
(2)
(3)
Ratio
Cases,
Cases, as % of total
Exports HQ→ FO
Col. (2) /
first pursued
n. cases first
as % of tot
Col. (3)
in HQ
pursued in HQ
exports of HQ
USA
China
65
20.63%
9.29%
2.22
USA
Austria
1
0.32%
0.31%
1.02
Germany
China
4
7.14%
3.96%
1.96
Germany
Austria
3
5.36%
5.62%
0.95
Notes.
Cases reported are those first pursued in the headquarters country (HQ), 1998-2012. FO: Foreign country.
The total number of cases first pursued in the US is 315, and in Germany is 56. Data sources are reported
in Appendix B.
26
Table 2. Total number of cases by headquarters country. 1998-2012
Country of firm’s headquarter
United States
Germany
United Kingdom
France
Switzerland
Italy
Spain
Australia
Canada
Japan
Netherlands
Sweden
Korea
Portugal
Norway
China
Argentina
Austria
Brazil
Finland
Bermuda
Chile
Denmark
Israel
Belgium
Hungary
Angola
Bangladesh
Czech Republic
Ghana
Ireland
India
Luxembourg
New Zealand
Poland
Russia
Slovak Republic
Turkey
British Virgin Islands
South Africa
Total
Total Cases
331
86
54
47
46
27
22
19
19
17
15
15
14
13
9
8
6
6
6
5
4
3
3
3
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
796
Positive Cases
227
50
25
27
40
5
0
5
4
11
8
1
14
0
5
7
1
2
0
4
0
0
0
2
0
1
0
1
0
1
0
1
1
0
0
0
0
0
1
0
444
Ongoing Cases
82
36
24
18
6
16
8
14
8
5
5
12
0
7
3
1
2
4
5
1
0
2
3
1
2
1
0
0
0
0
1
0
0
1
1
0
1
1
0
1
272
Notes. Cases are those first enforced in the headquarters country or in any third-country jurisdiction. The
“headquarters country” is where the firm which allegedly corrupted public officials abroad is
headquartered. Positive Cases refers to cases that were found guilty, settled (see the coding). Ongoing cases
are those that are still ongoing.
27
Table 3. Total number of cases by country where alleged corruption takes place
Foreign country
China
Nigeria
India
Russia
Indonesia
Libya
Brazil
Kazakhstan
Angola
Egypt
Argentina
Philippines
Greece
Mexico
Thailand
Saudi Arabia
United States
Vietnam
Venezuela
Algeria
Poland
Turkey
United Arab Emirates
Iraq
Ghana
Iran
Korea
Malaysia
Romania
Bangladesh
Bulgaria
Czech Republic
Hungary
Liberia
Serbia
Congo, Republic of
Costa Rica
Croatia
Italy
Kenya
Panama
Uganda
Uzbekistan
South Africa
Austria
Peru
Pakistan
Syrian Arab Republic
Bahrain
France
Equatorial Guinea
Haiti
Cambodia
Mali
Qatar
Slovenia
Tanzania
Azerbaijan
Côte d'Ivoire
Ecuador
Gabon
Georgia
Hong Kong
Honduras
Kuwait
Morocco
Mauritania
Total Cases
88
42
29
28
24
24
22
22
19
17
16
15
14
14
14
12
12
12
11
10
10
10
9
9
8
8
8
8
8
7
7
7
7
7
7
6
6
6
6
6
6
6
6
6
5
5
5
5
4
4
4
4
4
4
4
4
4
3
3
3
3
3
3
3
3
3
3
28
Positive Cases Ongoing Cases
49
34
32
6
12
14
15
10
18
5
6
16
7
11
8
9
9
8
14
3
9
7
9
5
9
4
10
4
11
3
7
4
12
0
11
1
7
3
3
6
6
2
6
1
5
4
3
6
2
3
4
2
3
5
5
1
3
3
5
2
4
1
2
4
3
3
4
1
4
2
4
2
3
3
3
2
2
3
4
2
2
3
3
2
4
2
2
3
2
2
1
3
4
1
3
2
4
0
4
0
2
1
2
1
2
2
3
0
2
2
2
2
4
0
3
0
2
1
3
0
2
0
1
1
3
0
3
0
1
2
1
1
2
1
Malawi
Oman
Rwanda
Taiwan
Ukraine
Zimbabwe
Albania
Colombia
Germany
Spain
Guinea
Lithuania
Latvia
Mozambique
Nepal
Sudan
Singapore
Slovak Republic
Senegal
Somalia
Turkmenistan
Tunisia
Yemen
Zambia
Afghanistan
Bosnia and Herzegovina
Belgium
Burkina Faso
Benin
Brunei Darussalam
Bolivia
Bahamas, The
Belarus
Cameroon
Cuba
Djibouti
Dominican Republic
Eritrea
Jamaica
Jordan
Japan
North Korea
Luxembourg
Madagascar
Macedonia
Myanmar
Mongolia
Moldova
Niger
Netherlands
Norway
Nairu
French Polynesia
Portugal
São Tomé and Príncipe
El Salvador
Turks and Caicos Islands
Chad
Trinidad and Tobago
United Kingdom
Uruguay
Total
3
3
3
3
3
3
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
796
1
1
2
2
1
0
1
1
1
1
1
1
1
1
2
2
1
0
1
0
2
0
1
0
0
1
1
1
1
1
1
1
0
0
1
1
0
0
0
1
0
0
1
1
1
1
1
0
1
1
0
0
1
0
0
0
0
1
0
1
0
444
1
2
1
1
1
2
1
1
0
1
1
1
1
0
0
0
1
2
1
0
0
2
1
1
1
0
0
0
0
0
0
0
0
1
0
0
1
1
1
0
1
1
0
0
0
0
0
1
0
0
1
1
0
1
0
1
1
0
0
0
0
272
Notes. Cases are those first enforced in the headquarters country or in any third-country jurisdiction. The
“Foreign country” is the country where the act of (alleged) corruption took place.
29
Table 4. Public Administration Corruption Index (PACI), 1998-2013
Country
Canada
Switzerland
Australia
Sweden
Ireland
Denmark
Finland
United Kingdom
Japan
Netherlands
Germany
Belgium
Spain
France
Singapore
Portugal
Norway
USA
Italy
Taiwan
Mexico
Korea
Austria
Dominican
Colombia
Republic
Luxembourg
Slovakia
Bahamas
El Salvador
Malaysia
South Africa
Jamaica
Turkey
Czech Republic
Poland
China
Trinidad and
Morocco
Tobago
Ukraine
Hungary
Tunisia
Saudi Arabia
Belarus
Jordan
Kuwait
Uruguay
Brazil
Lithuania
Cuba
India
Thailand
Qatar
Slovenia
Romania
Venezuela
Russia
Ecuador
Peru
Honduras
Brunei
Latvia
Darussalam
Costa Rica
Pakistan
Greece
Iran
PACI zALL
Pr_zero _ caseszPACI
0
0
0
0
0
0
0
2.58773
3.017315
3.520985
3.723345
3.786931
8.576681
10.58484
14.45795
19.10472
19.3984
21.54232
22.29935
25.40663
29.16911
46.02413
47.04836
50.16856
65.02713
80.35295
90.9612
99.06574
99.22466
106.3388
107.8829
119.4216
119.5222
125.0943
126.9792
144.8671
148.6733
149.0878
154.7996
168.4742
173.0795
190.4156
192.3195
198.2679
202.4064
204.3132
213.2271
217.4639
236.8499
241.05
255.6769
264.63
268.9919
276.0276
281.1082
310.2815
315.465
322.8425
339.1537
351.3141
367.9257
378.6169
380.5453
388.5903
396.8289
.0000
.0000
.0000
.0001
.0002
.0018
.0128
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0053
.0058
.0000
.0000
.0000
.0000
.0000
.0000
.1362
.0462
.2881
.1109
.3644
.3650
.0005
.0038
.4328
.0002
.0037
.0004
.0000
.5104
.1337
.1440
.0156
.3148
.0018
.5945
.6038
.2271
.6129
.0000
.3986
.6555
.0000
.0041
.2205
.2260
.0551
.0199
.0001
.3863
.2125
.4128
.7522
.5806
.2050
.2687
.0272
.1331
Country Rank Rank difference, WB CCI Rank difference, TI CPI
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
30
-10
-2
-4
-1
-9
4
6
-2
-10
2
-1
-5
-5
-2
12
-4
11
3
-22
-4
-35
-5
14
-52
-27
13
-2
-97
-38
-8
3
-33
-12
4
-5
-43
-10
-19
-43
14
-7
-8
-54
9
20
25
-6
9
13
-14
0
29
31
-1
-45
-35
-29
0
-28
-63
27
30
-44
31
-5
-11
-3
-4
0
-12
4
6
-2
-9
1
-2
-4
-6
-1
12
-4
11
4
-11
-4
-31
-7
15
-42
-17
15
-11
-94
-10
2
-4
-16
-20
-2
-21
-26
-7
-22
-51
9
9
-17
-46
17
11
21
0
15
4
-22
5
29
31
-11
-47
-42
-40
9
-28
-64
20
22
-50
27
-10
Oman
Afghanistan
Bosnia and
Bolivia
Herzegovina
Benin
Philippines
Algeria
Croatia
Yemen
Cameroon
Argentina
Macedonia
Bulgaria
Viet Nam
Sudan
Indonesia
Panama
Bahrain
Senegal
Syrian Arab
Côte d'Ivoire
Republic
Albania
Burkina Faso
Gabon
Madagascar
Egypt
Iraq
Azerbaijan
Serbia
Kenya
Bangladesh
Haiti
Niger
Georgia
Mozambique
Burma
Turkmenistan
Zambia
Nigeria
Ghana
Chad
Mongolia
Tanzania
Guinea
Djibouti
Kazakstan
Congo
Equatorial Guinea
Mauritania
Liberia
Libyan Arab
Zimbabwe
Jamahiriya
Uzbekistan
Cambodia
Nepal
Sao Tome & P.
Uganda
Mali
Malawi
460.9047
472.1338
472.9085
475.1096
479.1134
496.6861
533.9379
539.2692
545.6701
576.2632
586.4659
672.15
696.8326
714.1484
775.2856
846.4598
884.8087
900.8724
924.2173
955.4548
1061.689
1144.235
1171.676
1176.169
1219.936
1253.273
1268.119
1280.879
1356.355
1481.514
1549.458
1651.818
1812.204
1819.622
1900.602
1929.848
1947.478
1961.419
2101.201
2191.095
2207.392
2302.75
2307.337
2410.733
2425.423
2475.69
3088.969
3102.567
3242.89
3426.804
3550.355
4122.554
4290.974
4645.204
5354.408
5905.44
5911.488
6340.174
8598.199
.5215
.8091
.8094
.8101
.8116
.0488
.1536
.3286
.6931
.8407
.0653
.8617
.3662
.1863
.7726
.0586
.5075
.6414
.8054
.5925
.7538
.8396
.9181
.7748
.9212
.2575
.4917
.7911
.5968
.6669
.6365
.7849
.9463
.8480
.9001
.9495
.9024
.9030
.1354
.6941
.9557
.9575
.8408
.9203
.9596
.4112
.8234
.8790
.91164
.8152
.5086
.9298
.8695
.9174
.9633
.9832
.9034
.9388
.9657
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
113
114
115
116
117
118
119
120
121
122
123
124
125
29
-52
14
-20
-29
-6
4
31
-18
-33
7
6
34
-11
-38
-15
20
52
38
2
-26
-1
39
15
47
19
-28
-9
31
-6
-18
-20
14
38
27
-20
-14
10
-5
46
-9
32
28
5
29
8
8
-8
55
10
19
5
9
9
43
27
29
58
40
45
-27
-6
-24
-1
-21
-4
18
-9
-33
-1
-5
35
-9
-32
-26
32
57
23
27
-30
-13
34
21
10
34
-17
-17
30
-16
-24
-21
-1
-4
21
-18
-15
17
-10
54
-15
40
35
-6
41
27
12
-9
52
11
26
33
13
16
25
41
28
54
47
Notes. Index computed using all cases regardless of outcome, and administration. Countries for which
PACI
PACIz = 0 have been ranked according to the negative of Pr_zero _ casesz
. Countries for which PACIz
PACI
= 0 and Pr_zero _ casesz
< 0.015 and countries for which PACIz > 10000 have been excluded from the
list. WB-CC: World Bank Corruption Control Index. 2005. TI-CPI: Transparency International Corruption
Perception Index. 2005
31
Table 5. Spearman Rank Correlations between different indexes of corruption.
PACI1
PACI1
1
PACI2
0.938
PACI2
PACI3
PACI4
TI-CPI
WB-CC
TI-GCB
1
(123)
PACI3
PACI4
TI-CPI
WB-CC
TI-GCB
0.940
0.886
1
(123)
(123)
0.885
0.945
0.941
(123)
(123)
(123)
-0.779
-0.736
-0.751
-0.729
(123)
(123)
(123)
(123)
-0.768
-0.711
-0.729
-0.692
0.954
(123)
(123)
(123)
(123)
(123)
0.755
0.731
0.769
0.747
-0.800
-0.767
(56)
(56)
(56)
(56)
(56)
(56)
1
1
1
1
Notes.
PACI: Public Administration Corruption Index. 1998-2012. The subscript indicates:
1: All cases, all administrations (the same values shown in Table 4; our preferred index).
2: All cases, with the exclusion of health and telecom administration.
3: Only “positive” and “ongoing” cases, all administrations.
4: Only “positive” and “ongoing” cases, with the exclusion of health and telecom administrations.
WB-CC: World Bank Corruption Control Index, 2005. TI-CPI: Transparency International Corruption
Perception Index, 2005. TI-GCB: Percentage of persons who answered “yes” to the question: “In the past
12 months, have you or anyone living in your household paid a bribe in any form?.” Transparency
International Global Corruption Barometer, 2005. Number of observations are between parentheses. All
the estimated coefficients are significant at less than 1‰.
32
Table 6. Correlation coefficients between measurement residuals and selected variables.
All observations
mesresWB-CCI
mesresTI-CPI
Dummy
Dummy Asia
Dummy
Dummy
EU Asia
Pacific
America
Africa ME
Gdp_cap
Pop
R_g/gdp
-0.0604
-0.1405
-0.2507***
0.3621***
-0.0090
-0.2122**
-0.1182
(0.5071)
(0.1211)
(0.0052)
(0.0000)
(0.9242)
(0.0185)
(0.1946)
-0.1297
-0.0810
-0.2854***
0.4109***
0.0510
-0.2234**
-0.1010
(0.1527)
(0.3732)
(0.0014)
(0.0000)
(0.5916)
(0.0130)
(0.2683)
Democ
Checks
Stability
Free_press
mesresWB-CCI
Free_speech Voice_acc Emp_right
0.0368
-0.0739
-0.0604
-0.0677
-0.2687***
-0.2180**
0.1518*
(0.6876)
(0.4163)
(0.5067)
(0.4572)
(0.0033)
(0.0177)
(0.0938)
-
mesresTI-CPI
0.1579*
-0.1221
-0.1620
-0.1415
-0.3310***
0.2807***
0.1113
(0.0823)
(0.1786)
(0.0735) *
(0.1184)
(0.0003)
(0.0021)
(0.2205)
Excluding the ten biggest outliers
mesresWB-CCI
mesresTI-CPI
Dummy
Dummy Asia
Dummy
Dummy
EU Asia
Pacific
America
Africa ME
mesresTI-CPI
Pop
R_g/gdp
-0.0400
-0.1297
-0.1928**
0.2860***
0.0478
-0.2234**
-0.0933
(0.6737)
(0.1708)
(0.0408)
(0.0021)
(0.6315)
(0.0174)
(0.3278)
-0.1200
-0.0758
-0.2216**
0.3476***
0.1124
-0.2319**
-0.0772
(0.2056)
(0.4247)
(0.0184)
(0.0002)
(0.2585)
(0.0134)
(0.4186)
Free_press Free_speech Voice_acc Emp_right
mesresWB-CCI
Gdp_cap
Democ
Checks
Stability
-0.0221
-0.0673
0.0019
-0.0391
-0.1882*
-0.2234**
0.2073**
(0.8171)
(0.4787)
(0.9839)
(0.6904)
(0.0511)
(0.0195)
(0.0276)
0.1159
-0.1104
-0.1161
-0.1153
-0.2682***
-0.2818***
0.1619*
(0.2238)
(0.2443)
(0.2208)
(0.2241)
(0.050)
(0.0030)
(0.0866)
Notes.
Number of observations: Panel A: between 124 and 117; Panel B: between 104 and 113. P-values between
parentheses. * p-value < 0.1; ** p-value < 0.05; *** p-value < 0.001. For a description of the data, see the
Appendix B.
33
Table 7: Multivariate analysis of the determinants of the measurement residuals
Dep. var: mesresWB-CCI
Regressors
All observations
Excluding
Dep. var. mesresTI-CPI
All observations
10 outliers
Asia & Pacific
America & Caribbean
Africa & Middle East
Per capita GDP. PPP
Pop
R_g/gdp
Free_press
Free_speech
Voice_acc
Emp_right
Democ
Checks
Stability
Observations
R-squared
Excluding
10 outliers
-0.4102
(0.3213)
-0.9691**
(0.4018)
0.3450
(0.3502)
11.7404
(10.0708)
-0.0063
(0.0006)
-0.0102***
(0.0035)
-0.0036
(0.0120)
-0.1566
(0.2883)
0.4953
(0.3785)
0.0734
(0.0777)
-0.0440**
(0.0194)
-0.0436
(0.0457)
0.1302
(0.1705)
-0.2714
(0.2886)
-0.6412*
(0.3569)
0.4022
(0.3281)
8.8195
(9.5360)
0.0052
(0.0006)
-0.0097***
(0.0027)
-0.0008
(0.0116)
-0.0559
(0.280)
0.467
(0.361)
0.0114
(0.0663)
-0.0234
(0.0176)
-0.0642
(0.0420)
0.1582
(0.1491)
-0.0992
(0.3243)
-0.8116**
(0.3778)
0.4826
(0.3624)
21.7674*
(11.189)
-0.0008*
(0.0004)
-0.0075*
(0.0041)
0.0127
(0.0113)
-0.0909
(0.266)
0.1817
(0.3578)
0.1048
(0.0790)
-0.03245*
(0.0178)
-0.0332
(0.0404)
0.1605
(0.1592)
0.1070
(0.2972)
-0.4874
(0.3353)
0.4655
(0.3453)
11.8142
(8.4577)
-0.0010**
(0.0004)
-0.0010**
(0.0004)
0.0068
(0.0109)
-0.0357
(0.24089)
0.2087
(0.3176)
0.0539
(0.0682)
-0.0247
(0.0164)
-0.0228
(0.0346)
0.1560
(0.1467)
105
0.328
95
0.295
105
0.354
96
0.337
Notes. OLS estimates. Robust standard errors are between parentheses. * * p-value < 0.1; ** p-value <
0.05; *** p-value < 0.001. For a description of the variables, see the note at the bottom of Table 6, and
Appendix B.
34
Figure 1. Comparison between PACI and WB Corruption Control Index
A. PACI vs. WB-CCI
8000
MW
6000
ML
UG
ST
NP
PACI
4000
KH
UZ
ZW
LR LY
GQ CG
MR
0
2000
KZ DJ
GN
TZ
MN GH
TDNG
TM
MM GE
NEZMMZ
HT
BD
KE
RS
AZAL MG
IQ
GA EG
BF
CI
SNSY
BH
PA
ID VN
SD
BG
MK
AR
CMPH
YE
DZ
HR
BJ
AF
BO
BA
IR
GR
PKVE
CR
BN QASI OM
HN
EC
PEBR
RU
CU LV
LT
JOUY
BY INRO
TN
HU
UA
CNSA
MA
TT
PL
CZZAKW
TR
JMTH
SK
COSV
DO
KR
MX
ITMY
TW PT
2
4
US
ES JP
BE
FR
IE
6
WB CC, year 2005
LU
AT
NO
DECA
NL
UK
AU
CH
SESG
DK
FI
8
10
10
B. Log(PACI) vs. WB-CCI
MW
ST
UG NP ML
LR
MR
GQ
CGLY
KZ
DJ
TZ
MN GH
TD NGGN
ZM
TM
MM
MZ
NE
GE
HT
BD
KE
AZ
IQ
EGRS BF MG
AL GA
CI
SY
SN
BH
PA
IDVN
SD
BG
MK
AR
CM
HR OM
BJ YE
AF
BO PH DZ BA
IR
GR
PK
CR
LV
BN
HN
PE
EC
VE RU
SI
QA
TH
IN RO
CU
LT
BR
UY
KW
JO
BY
SA
HU
UA
MA TN
TT
CN
CZZA
TR PL
JM
MY
SV
SK
CO
DO
KR
MX
TW
IT
PT
6
4
LU
AT
US
NO
2
FR
ES
JP
BE
IE
DENL
CA
UK
AU
SE
CH
SG
FI
DK
0
Log PACI
8
KH
UZ
ZW
-2
-1
0
WB CC, year 2005
1
2
Notes.
In Panel B. when the PACI equals to zero (countries: FI. DK. IE. SE. AU. CA), it has been set equal to
arbitrary small numbers for the purpose of computing the log.
35
References
Alt, James E., and David Dreyer Lassen. 2012. "Enforcement and Public Corruption: Evidence
from the American States." Journal of Law, Economics, and Organization.
Barbieri, Katherine and Omar Keshk. 2012. Correlates of War Project Trade Data Set Codebook,
Version 3.0. Online: http://correlatesofwar.org.
Carr, Indira M. and Opi Outhwaite. 2008. “The OECD Anti-Bribery Convention Ten Years Later.”
Manchester Journal of International Law 5(1): 3-55.
Chang, Eric; Golden, Miriam Golden and Seth Hill. 2010. “Legislative Malfeasance and Political
Accountability.” World Politics 62(2): 177-220
Cheung, Yan Leung, P. Raghavendra Rau, and Aris Stouraitis. 2012. “How much do firms pay as
bribes and what benefits do they get? Evidence from corruption cases worldwide.”
National Bureau of Economic Research Working Paper No. 17981: 1-38.
Choi, Stephen and Kevin E. Davis. 2012. “Foreign Affairs Enforcement of the Foreign Corrupt
Pratices Act.” Unpublished paper.
Clausen, Bianca, Aart Kraay, and Peter Murrell. 2011. “Does Respondent Reticence Affect the
Results of Corruption Surveys? Evidence from the World Bank Enterprise Survey for
Nigeria,” in International Handbook on the Economics of Corruption, edited by Susan
Rose-Ackerman, 428-450. Vol.2. Cheltenam UK: Edward Elgar Publishing House.
Fisman, Raymond and Roberta Gatti. 2002. “Decentralization and corruption: Evidence from US
federal transfer programs.” Public Choice 113(1): 25-35.
Glaeser, Edward L., and Raven E. Saks. 2006. "Corruption in America." Journal of Public
Economics 90(6): 1053-1072.
Goel, Rajeev and Michael Nelson. 2011. “Measures of corruption and determinants of US
corruption.” Economics of Governance 12:155-176.
Golden, Miriam and Lucio Picci. 2008. “Pork Barrel Politics in Postwar Italy, 1953-1994.”
American Journal of Political Science 52(2): 268-289.
Kaufmann, Daniel, Aart Kraay, and Massimo, Mastruzzi. 2009. “Governance matters VIII:
aggregate and individual governance indicators, 1996-2008.” World Bank Policy Research
Working Paper 4978.
Klitgaard, Robert and Paul C. Light. 1998. “Economic Growth and Social Indicators: An
Exploratory Analysis.” Economic Development and Cultural Change, 46(3): 455-489
Klitgaard, Robert and Johannes Fedderke and Kamil Abramamov. 2005. “Choosing and Using
Performance Criteria” in High Performance Government. Structure, Leadership,
36
Incentives, edited by Robert Klitgaard & Paul C. Light. Santa Monica, CA: Rand
Corporation
Kraay, Aart and Peter Murrell. 2013. “Misunderestimating Corruption.” World Bank Policy
Research Working Paper 6488.
Lambsdorff, Johann Graf. 2006. “Consequences and Causes of Corruption - What do We Know
from a Cross-Section of Countries? in International Handbook on the Economics of
Corruption, edited by Susan Rose-Ackerman, 3-52 Cheltenam UK: Edward Elgar
Publishing House.
Lambsdorff, Johann Graf. 2007. The Institutional Economics of Corruption and Reform: Theory
and Policy. Cambridge, UK: Cambridge University Press.
McLean, Nicholas M. 2012. “Cross-National Patterns in FCPA Enforcement.” The Yale Law
Journal 121(7):1970-2012.
OECD, Various years. “Country reports on the implementation of the OECD Anti-Bribery
Convention.”http://www.oecd.org/daf/antibribery/countryreportsontheimplementationoft
heoecdanti-briberyconvention.htm.
Olken, Benjamin A. 2009. "Corruption perceptions vs. corruption reality." Journal of Public
Economics 93(7): 950-964.
Picci, Lucio. 2011. Reputation-based Governance. Stanford, CA: Stanford University Press.
Putnam, Tonya L. 2009. “Courts without borders: The Domestic Sources of US Extraterritorial
Regulation.” International Organization 63(3):459-90.
Razafindrakoto, Mireille, and François Roubaud. 2010. "Are international databases on corruption
reliable? A comparison of expert opinion surveys and household surveys in sub-Saharan
Africa." World development 38(8): 1057-1069.
Reinikka, Ritva and Jakob Svensson. 2006. "Using micro-surveys to measure and explain
corruption." World Development 34(2): 359-370.
Rose-Ackerman, Susan. 1999. Corruption and Government: Causes, Consequences and Reform.
Cambridge, UK: Cambridge University Press.
Saisana, Michaela and Andrea, Saltelli. 2012. “Corruption Perceptions Index 2012 Statistical
Assessment.” JRC Scientific and Policy Reports, European Commission.
Seligson, Mitchell. 2006. "The measurement and impact of corruption victimization: Survey
evidence from Latin America." World Development 34(2): 381-404.
Sherif, Muzafer and Hadley Cantril. 1945. “The psychology of 'attitudes': Part I.” Psychological
Review, 52(6): 295-319
37
Shearman & Sterling Llp. 2013. “FCPA Digest of Cases and Review Releases Relating to Bribes
to Foreign Officials under the Foreign Corrupt Practices Act of 1977.” Shearman and
Sterling LLP, New York. http://www.shearman.com/files/Publication/287c1af0-f9cb4c11-805d-91c409975b41/Presentation/PublicationAttachment/83d9dc0b-b80c-4ca4877b-9efbba0952e7/FCPA-Digest-Jan2013_010213.pdf.
Teorell, Jan, Nicholas Charron, Stefan Dahlberg, Sören Holmberg, Bo Rothstein, Petrus Sundin &
Richard Svensson. 2013. The Quality of Government Dataset, version 20 Dec13.
University of Gothenburg: The Quality of Government Institute, http://www.qog.pol.gu.se.
Transparency International. 2009. “OECD Anti-Bribery Convention Progress Report 2009:
Enforcement of the OECD Convention on Combating Bribery of Foreign Public Officials
in
International
Business
Transactions.”
http://www.transparency-
usa.org/news/documents/FinalOECDProgressReport2009.pdf.
Transparency International. 2012.
“Corruption Perceptions Index 2012: An updated
methodology.” http://cpi.transparency.org/cpi2012/in_detail/.
Transparency International. 2013. “Exporting Corruption: Progress Report 2013: Assessing
enforcement
of
the
OECD
Convention
on
Combating
Bribery.
http://www.transparency.org/whatwedo/pub/exporting_corruption_progress_report_2013
_assessing_enforcement_of_the_oecd.
Treisman, Daniel. 2007. “What Have We Learned About the Causes of Corruption From Ten
Years of Cross-National Empirical Research?” Annual Review of Political Science 10: 211244
Van Aaken, Anne, Lars Feld, and Stefan Voigt. 2010. "Do independent prosecutors deter political
corruption? An empirical evaluation across seventy-eight countries." American Law and
Economics Review 12(1): 204-244.
38