Working Paper Series

2 downloads 337 Views 263KB Size Report
Keywords: maritime piracy, economic development ... 1 University of Paderborn, Center for International Economics. Germa
CENTER FOR INTERNATIONAL ECONOMICS Working Paper Series

Working Paper No. 2014-02

Maritime Piracy: Socio-Economic, Political, and Institutional Determinants Thomas Gries, Margarete Redlin January 2014

Maritime Piracy: Socio-Economic, Political, and Institutional Determinants T. Gries,1 M. Redlin2

Abstract Over the last twenty years piracy has become an increasing threat. Yet there are only very few econometric studies that examine under which conditions this phenomenon arises. As the number of maritime piracy and armed robbery incidents is characterized as count data and exhibits overdispersion, we apply random-effects negative binomial regressions for a panel dataset covering the period 1991-2010. Our results indicate that poor socio-economic, political, and institutional conditions in the host country increase the likelihood of piracy attacks.

Keywords: maritime piracy, economic development JEL Classification: C25, F51, P16, O10

                                                             1

University of Paderborn, Center for International Economics. Germany. Tel.: +49-(0)5251-60-2113, fax: +49-

(0)5251-60-3540, e-mail: [email protected]. 2

University of Paderborn, Center for International Economics. Germany. Tel.: +49-(0)5251-60-3823, fax: +49-

(0)5251-60-3540, e-mail: [email protected].

1   

1. Introduction Maritime piracy is a serious problem with far-reaching consequences. In the last decade (20032012) a total of 3,436 piracy acts were registered, most of them in the region of East Africa including Somalia, the Red Sea, and the Gulf of Aden, as well as in South East Asia (see Figure 1). The negative impacts of piracy comprise not only casualties and ransom payments, but also damage to the region and the global economy (e.g., the transport sector (Martínez-Zarzoso and Bensassi, 2013; Bendall, 2010; Fu et al., 2010) and tourism (Bowden et al., 2010)). Accordingly, ”… the World Bank has now put an annual price on piracy, during its surge between 2005 and 2011, of $18 billion.” The Economist (May 19, 2013). However, the roots of maritime piracy are still largely unexplored. In this contest, Geise (2007) argues that (trans-) regional countermeasures will fail to generate the desired effects as long as the underlying regional problems are not addressed effectively. Most studies take a qualitative approach to identifying the sources of piracy, very few use econometric methods. While Maija et al. (2009) look at ship characteristics (flag of registry, type of vessel) and Hastings (2009) at the effects of failed states on the type of hijackings, Cariou and Wolff (2012) address economic and freedom indicators. However, none of the studies consider count data characteristics, which leads to insignificant results for important variables. Resolving this methodological problem, we identify an even more comprehensive set of determinants of maritime piracy.

2   

Figure 1: Regional piracy attacks (1991-2012) 500 450 400 Total

350

SE Asia

300

Far East

250

Indian SC

200

Americas

150

Africa

100

Rest of the World

50 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

0

Source: International Maritime Bureau

2. Determinants of Maritime Piracy Following rational choice theory (Cornish and Clarke, 1986; Leeson, 2009) costs and benefits are evaluated. Even if an estimated third of pirate financiers spends profits for militias, political influence, or religious extremists (World Bank, 2013) piracy seems mostly motivated by profits, mainly from ransom earnings. Accordingly, the notion that precarious economic conditions create piracy is widespread (see Cariou and Wulff, 2012). Scholvin (2009) suggests that for coastal dwellers, the lack of opportunities to earn a living through fishery is a crucial motivation for participating in piracy. Furthermore, institutional characteristics such as legal and jurisdictional weakness (Murphy, 2013), poor domestic governance, civil conflicts and disorder (Hastings, 2009; Murphy, 2013), corruption and inefficient bureaucratic quality (Murphy, 2013; Chalk, 2009), political institutions (Cariou and Wulff, 2012; Murphy, 2013) as well as insufficient effectiveness of international law and 3   

institutions (Roach, 2010) seem to create a breeding ground for organized piracy. Finally, favorable external (e.g., heavy shipping traffic), demographic (e.g., country size or population density (Hastings, 2009)), and geographic characteristics (e.g., geographic location and coastline (Hasting, 2009)) may also promote piracy.

3. Empirical Design Piracy is the independent variable drawn from the Piracy and Armed Robbery against Ships reports (International Maritime Bureau), as country-year observation of piracy incidents. We suggest three sets of explanatory variables. Economic conditions: The stage of development is indicated by real GDP per capita (Penn World Table (PWT)). Low income levels may reduce the opportunity costs of piracy and increase its likelihood. High growth rates of real GDP per capita would do the opposite. Trade openness as the ratio of trade to GDP (PWT) could foster development3 and reduce piracy, however pure shipping traffic would provide more opportunities for piracy in countries with high trade volumes. Oppression, civil and political liberties: Political tension and a political, social, and business culture in which not only the lawful use of force is present may breed criminal structures. We expect piracy to be more likely in countries with unstable and immature political conditions,4 as indicated by the civil liberties and political rights indices (Freedom House). To indicate political violence we use the physical integrity rights index (Cingranelli and Richards, 2010) and expect inhuman state behavior to drive criminal activity.

                                                             3

See, e.g., Yannikaya (2003).

4

See Cariou and Wulff (2012).

4   

Institutional quality: Functioning government institutions and a sound internal order are the prerequisites for a well functioning social and economic system. E.g., high bureaucratic quality and the rule of law create less favorable conditions for criminal action, while internal conflict and a corrupt government could facilitate access to arms and increase the chance of circumventing criminal prosecution. Our measures are risk ratings for internal conflict, law and order, corruption, and bureaucracy quality (International Country Risk Guide). After rescaling, higher values indicate higher risks and presume more piracy. Geographic, environmental, and demographic controls: Both demographic and geographic factors such as population (PWT) and coast access measured by the length of a country’s coastline (The World Factbook) should have a direct scaling effect on piracy. Fishery (FishJStat) is an income substitute for piracy, so we assume that decreasing catches encourage a switch towards illegal sources of income.

4. Results Our empirical analysis is based on a battery of estimations for a sample of 149 countries for the period 1991-2010. Since maritime piracy requires access to the sea we exclude countries without a sea coastline. We include year dummies and consider the influence of a lagged dependent variable to account for serial correlation, heterogeneity, and a bias from the omission of variables. All explanatory variables enter the model in lagged form to avoid the problems associated with reverse causation. The dependent variable (number of piracy incidents) is characterized as countdata and exhibits overdispersion, so we apply random-effects negative binomial regression.5                                                              5

In cases of overdispersion the standard Poisson estimator may underestimate the standard errors, causing

misleading inferences about the regression parameters. We prefer the random-effects estimator over the fixed-effects estimator since the latter may mask the influence of slowly changing or constant variables. However, alternative

5   

Table 1: Negative binomial estimation results GDP p.c.t-1

(1)

(2)

(3)

(4)

(5)

(6)

-0.491*** (0.084)

-0.485*** (0.084) 0.011*** (0.001)

-0.444*** (0.083)

-0.554*** (0.089)

-0.545*** (0.088)

-0.468*** (0.101)

-0.526 (0.623) 0.005*** (0.002) 0.320*** (0.041)

0.006*** (0.002) 0.295*** (0.049)

0.006*** (0.002) 0.297*** (0.048)

0.003 (0.002) 0.288*** (0.050)

-0.063** (0.029) -0.152*** (0.050)

-0.084*** (0.030)

Piracyt-1 GDP p.c. growtht-1 Opennesst-1 Populationt-1 Coast length Fishery p.c. (% change)t-1 Physical integrity rightst-1 Civil libertiest-1

0.006*** (0.002) 0.272*** (0.049) 0.122* (0.068) -0.120 (0.230)

0.005** (0.002) 0.300*** (0.043)

Political rightst-1 Internal conflictt-1 Law and ordert-1 Corruptiont-1

-0.157*** (0.032) 0.056** (0.027) 0.199*** (0.067) -0.218*** (0.061) -0.034 (0.081) YES 2.13 -1813.31 241.54*** 784.56*** 113 1986

Bureaucracy qualityt-1 Year dummies YES YES YES YES YES Mean VIF 1.83 1.79 1.81 2.04 2.02 Log likelihood -2143.96 -2100.05 -2145.64 -2002.62 -1995.16 255.11*** 369.57*** 247.94*** 253.08*** 272.30*** Wald χ2 Likelihood ratio 869.62*** 750.02*** 970.72*** 837.40*** 846.95*** Countries 149 149 149 144 144 Observations 2938 2819 2956 2508 2508 * p