Fig. 1A Predicted count of piracy events (95% CIs)
Fig. 2A Marginal effects of the number of ethnic groups on piracy. Note: The marginal effects are based on a negative binomial regression that includes an interaction between the (log) number of ethnic groups and bottom tier election. The value of the logarithm of the number of ethnic groups ranges between 0 and 3.22 in our data. We restrict the values of the ethnic groups at 2.75 in this graph since all marginal effects become insignificant beyond 2.75. For zero or partial bottom tier elections, the marginal effects of ethnic groups are significant at the 10% level of significance or less over the range shown in the graph. The marginal effects are never significant for (full) bottom tier election.
Table A1 Description and source of variables
Variable / Description / SourceCountry-level sample
Piracy / The number of piracy incidents / Daxecker and Prins (2015d), Maritime Piracy Event and Location Data Set
IRISK / A composite index of institutional quality with higher values reflecting better institution. It is an aggregate of ICRG indices of government stability, investment profile, corruption, military in politics, religion in politics, rule of law, democratic accountability and bureaucratic quality / ICRG (2015)
Corruption (ICRG) / An index ranging between 0 and 6 with higher scores representing lower level of corruption in government / ICRG (2015)
Rule of Law (ICRG) / An index ranging between 0 and 6 with higher scores indicating higher quality of political institutions, strong judiciary, and provisions for proper succession of power / ICRG (2015)
Gov. Effect / Government effectiveness indicator ranging between –2.5 and 2.5 with higher values representing more effective government. It “reflects perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies.” / Kaufmann et al. (2010); World Bank (2014b)
Corruption (WB) / An indicator ranging between –2.5 and 2.5 with higher values representing less corruption. It “reflects perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as "capture" of the state by elites and private interests.” / Kaufmann et al. (2010); World Bank (2014b)
Rule of Law (WB) / An indicator ranging between –2.5 and 2.5 with higher values reflecting better governance performance. It “reflects perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence.” / Kaufmann et al. (2010); World Bank (2014b)
Expenditure Ratio / Ratio of subnational (state and local) to total government expenditures (state, local, and central or general government expenditure depending on data availability) / International Monetary Fund (2015)
Subnational Revenue / Subnational revenues as a percentage of GDP / Fan et al. (2009)
Tiers / The number of levels of government / Fan et al. (2009)
Bottom Unit Size / The average size (in thousand square kilometers) of the bottom tiers, which is computed as the country’s surface area is divided by the number of bottom level units / Fan et al. (2009)
Autonomy / A dummy variable for autonomy of subnational governments, where regional units have authority over one or more public services. This dummy is one if autonomy exists. / Fan et al. (2009)
Bottom Tier Election / A variable that equals one if all executives at the bottom level of government are directly elected, 0.5 if some of the executives are elected, and zero if all of the executives are appointed. / Fan et al. (2009)
Second Lowest Tier Election / A variable that equals one if all executives at the second lowest level of government are directly elected, 0.5 if some of the executives are elected, and zero if the all of the executives are appointed. / Fan, Lin et al. (2009)
ln(GDP per capita) / Logarithm of GDP per capita in PPP constant 2011 international dollars / World Bank (2014a)
ln(POP) / Logarithm of total population / World Bank (2014a)
Terror / Number of terrorist incidents / National Consortium for the Study of Terrorism and Response to Terrorism (2013)
Conflict / An aggregate index of civil violence, civil war, ethnic violence, and ethnic war. Its values range between 0 and 10 with higher values indicative of a more severe conflict. / Marshall (2014)
ln(coastline) / Logarithm of coastline length in kilometers / Daxecker and Prins (2015c)
ln(regional trade) / Logarithm of regional trade volume (in US dollars) / Daxecker and Prins (2015c)
ln(fish value) / Logarithm of fisheries production value per capita (in US dollars) / Daxecker and Prins (2015c)
Traffic / A dummy variable reflecting a nearby shipping line carrying at least 5000 ships annually / Coggins (2012); Daxecker and Prins (2015c)
Regional sample*
Piracy / Number of piracy incidents / Authors’ calculation based on Daxecker and Prins (2015d), Maritime Piracy Event and Location Data Set
Institutional Quality / A latent variable based on the following measures: informal payments, tax days, days without electricity, security costs, access to land, security costs, access to land, access to finance, government predictability, and doing business rank / Gennaioli et al. (2013)
ln(GDP per capita) / Logarithm of regional GDP per capita in PPP constant 2005 international dollars / Gennaioli et al. (2013)
ln(POP) / Logarithm of the number of population in a region / Gennaioli et al. (2013)
Inverse Distance / The inverse of the (one plus) region’s distance to the closest coastline in thousands of kilometers / Gennaioli et al. (2013)
ln(Ethnic Groups) / The logarithm of the number of ethnic groups in a region in 1964 / Gennaioli et al. (2013)
Education / Average years of schooling in a region / Gennaioli et al. (2013)
Temperature / Average regional temperature during 1950-2000 / Gennaioli et al. (2013)
Number of Ports / Number of major ports in a region / Authors’ calculation based on information from National Geospatial-Intelligence Agency (
*See country-level sample for description of Expenditure Ratio, Subnational Revenue, Tiers, Bottom Unit Size, Autonomy, Bottom Tier Election, and Second Lowest Tier Election.
Table A2 Summary statistics: country-level sample
Variable / Obs / Mean / Std. Dev. / Min / Max / N Countries / PeriodPiracy / 2046 / 3.05 / 10.82 / 0.00 / 136.00 / 93 / 1993–2014
IRISK / 1716 / 35.27 / 8.25 / 5.50 / 56.29 / 78 / 1993–2014
Corruption (ICRG) / 1716 / 2.62 / 1.11 / 0.00 / 6.00 / 78 / 1993–2014
Rule of Law (ICRG) / 1716 / 3.39 / 1.32 / 0.00 / 6.00 / 78 / 1993–2014
Gov. Effect / 1393 / -0.21 / 0.91 / -2.45 / 2.36 / 93 / 1996–2013
Corruption (WB) / 1393 / -0.26 / 0.88 / -2.06 / 2.55 / 93 / 1996–2013
Rule of Law (WB) / 1395 / -0.32 / 0.91 / -2.67 / 2.00 / 93 / 1996–2013
Expenditure Ratio / 492 / 0.26 / 0.18 / 0.00 / 0.86 / 40 / 1993–2014
Subnational Revenue / 638 / 4.40 / 3.80 / 0.28 / 14.37 / 29
Tiers / 880 / 4.18 / 0.97 / 2.00 / 6.00 / 40
Bottom Unit Size / 748 / 0.72 / 1.67 / 0.00 / 9.32 / 34
Autonomy / 726 / 0.33 / 0.47 / 0.00 / 1.00 / 33
Bottom Tier Election / 704 / 0.73 / 0.43 / 0.00 / 1.00 / 32
Second Lowest Tier Election / 682 / 0.60 / 0.47 / 0.00 / 1.00 / 31
ln(GDP per capita) / 1862 / 8.79 / 1.19 / 4.96 / 11.62 / 90 / 1993–2013
ln(POP) / 1953 / 16.21 / 1.96 / 11.19 / 21.03 / 93 / 1993–2013
Terror / 1760 / 13.08 / 54.37 / 0.00 / 1389.00 / 88 / 1993–2012
Conflict / 1827 / 0.76 / 1.71 / 0.00 / 10.00 / 87 / 1993–2013
ln(coastline) / 2046 / 7.20 / 1.53 / 3.61 / 10.91 / 93 / 1993–2014
ln(regional trade) / 1674 / 28.07 / 1.29 / 25.98 / 30.22 / 93 / 1993–2010
ln(fish value) / 1472 / 2.53 / 1.60 / 0.00 / 8.14 / 92 / 1993–2008
Traffic / 2046 / 0.59 / 0.49 / 0.00 / 1.00 / 93 / 1993–2014
Note:Expenditure Ratio, Subnational Revenue, Tiers, Bottom Unit Size, Autonomy, Bottom TierElection, and Second Lowest Tier Election do not vary over time.
Table A3 Summary statistics: regional sample
Piracy / 1598 / 2.72 / 21.01 / 0 / 397 / 110
Institutional Quality / 507 / 0.00 / 0.12 / -0.38 / 0.31 / 81
Expenditure Ratio / 882 / 0.33 / 0.18 / 0.00 / 0.84 / 49
Subnational Revenue / 894 / 6.79 / 5.15 / 0.28 / 23.42 / 55
Tiers / 1078 / 3.91 / 0.87 / 2.00 / 6.00 / 70
Bottom Unit Size / 987 / 0.98 / 2.15 / 0.00 / 13.21 / 64
Autonomy / 908 / 0.43 / 0.50 / 0.00 / 1.00 / 60
Bottom Tier Election / 914 / 0.82 / 0.37 / 0.00 / 1.00 / 57
Second Lowest Tier Election / 857 / 0.62 / 0.46 / 0.00 / 1.00 / 53
ln(GDP per capita) / 1537 / 8.76 / 1.31 / 1.76 / 11.87 / 107
ln(POP) / 1569 / 13.78 / 1.61 / 9.14 / 19.00 / 110
Inverse Distance / 1569 / 0.85 / 0.15 / 0.33 / 1.00 / 110
ln(Ethnic Groups) / 1568 / 0.99 / 0.73 / 0.00 / 3.22 / 110
Education / 1519 / 7.10 / 3.14 / 0.22 / 13.21 / 107
Temperature / 1568 / 15.17 / 8.42 / -12.73 / 29.15 / 110
Number of Ports / 1570 / 1.56 / 4.75 / 0.00 / 94.00 / 107
Table A4 Panel Negative Binomial Regressions: Incidence rate ratios
Model 1 / Model 2 / Model 3 / Model 4 / Model 5 / Model 6 / Model 7IRISK / 0.967** / 1.357*** / 0.929** / 1.002 / 0.944*** / 0.973 / 1.039
IRISK Sq. / 0.997***
Expenditure Ratio / 0.045** / 0.000***
Subnational Revenue / 1.041
Tiers / 0.816 / 0.399 / 1.525 / 1.194
Bottom Unit Size / 6.195*** / 1.199 / 9.181*** / 2.983***
Autonomy / 0.098***
Bottom Tier Election / 0.183***
Second Lowest Tier Election / 0.230***
NT / 1159 / 1159 / 283 / 178 / 368 / 416 / 384
Note: Significance levels: *** is <.01, ** is <.05, and * is <.10. Only main variables of interest are presented to save space. See Table 4 for full specifications.