Corruption as a Predictor of Invasive Species Trade Risk

Evan Brenton-Rule, Faculty of Law / School of Biological Sciences, Dr Rafael Barbieri, School of Biological Sciences & Professor Phil Lester, School of Biological Sciences, Victoria University of Wellington, New Zealand. [email protected] [email protected] [email protected]

The opinions expressed and arguments employed herein are solely those of the authors and do not necessarily reflect the official views of the OECD or of its member countries. This document and any map included herein are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. This paper was submitted as part of a competitive call for papers on integrity, anti-corruption and trade in the context of the 2016 OECD Integrity Forum.

Abstract Invasive species have enormous economic and environmental impacts. International trade is the leading pathway for the introduction of non-native species. Current risk assessment paradigms for invasive species utilise detailed knowledge of pathway strength, species’ biology, as well as recipient region ecosystems and environments to make risk predictions. These analyses are often specific to narrow pathways and taxonomic levels. Various developmental and socio-economic factors such as corruption, have been demonstrated as having an impact on regions’ economic growth, invasibility and environmental standards, but no studies have used such indicators as predictors of the risk of nonnative species in trade. This is surprising as indicia of governance are likely to impact upon the implementation and efficacy of biosecurity measures. Here we examined a non-native species interception database for commodities imported to New Zealand across 10 years of interception records – over 47,000 non-native species interceptions – as a proxy for country-specific invader pressure. We test the hypothesis that governance and development indicators – including corruption – can be used in risk analysis as a predictor of non-native species incursion risk associated with international trade. Corruption was able to estimate invasive species’ interception rates across all three commodity pathways analysed. Counter-intuitively an increase in corruption is correlated with a decrease in non-native species’ detections. This is surprising as we had hypothesised commodities originating in corrupt countries carry greater non-native species risk as, anecdotally, biosecurity protocols are often poorly implemented or ignored. We posit several explanations for this result, including the adaptive approach of New Zealand’s Ministry for Primary Industries when granting licenses to export from certain regions. This approach aims to circumvent issues, such as high levels of corruption, that may impact upon the quality of biosecurity processing of commodities destined for New Zealand.

Keywords corruption, invasive species, risk

1

1. Introduction Globalisation has resulted in increasing interconnectivity of human communities, as well as heightening the impact of those communities on the environment. Biological invasions are a significant global challenge. Invasive species may be defined as non-native species that arrive in a new area, establish a population and increase in distribution and density to the detriment of the recipient environment and its ecological communities. Species that become invasive have enormous economic impacts, as well as posing a major threat to biodiversity (Vilà et al. 2010). An early study in the field estimated the costs of invasive species to the United States as US$120 billion annually (Pimentel et al. 2005). Similarly, management costs in Europe are billions per year (Scalera 2010). Forest pests alone are estimated to cost the United States US$2.5 billion per year (Aukema et al. 2011). Many vectors of human disease, such as Aedes mosquitos, are invasive species and impose enormous health costs. Environmental impacts are similarly deleterious. Invasive species have been implicated in a reduction in ecosystem services (van Wilgen et al. 2008; Pejchar & Mooney 2009) and agricultural productivity (DiTomaso 2000), changes in ecosystem structure (Vitousek & D’Antonio 1997) and the extinction of species (Fritts & Rodda 1998). Well-known invasives include the Red Imported Fire Ant (Solenopsis invicta), the ship rat (Rattus rattus) and the Asian giant hornet (Vespa mandarinia) which decimates bee hives and has been responsible for several deaths in France.

Global trade is the primary driver of the introduction of non-native species (Westphal et al. 2008; Hulme 2009). Introductions may be intentional or unintentional. Intentional introductions occur when species are deliberately introduced to an area, for instance rabbits and deer into New Zealand. Unintentional introductions typically occur when non-native species arrive in a new area associated with imported commodities. Examples include, insects on fresh fruit and vegetables or pathogens associated with imported live plants (Liebhold et al. 2012). Other unintentional introductions include hitchhiker species associated with methods of commodity conveyance; for example attached to the hull, or in the ballast water of, vessels (Roman & Darling 2007).

2

High levels of variability exist in the quality of governance exhibited by state participants in the global trading system. Many trading states suffer from poor governance, including high levels of corruption (Transparency International 2015). Further, at an individual level, people in corrupt countries have been shown as more likely to lie than individuals in non-corrupt countries (Gächter & Schulz 2016). In this article we adopt the definition of corruption used by Kaufmann et al. (2010); corruption may be defined as: ‘capturing 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’. In many countries corruption is a significant impediment to economic and social development, as well the implementation of effective environmental policies. From an economic perspective, corruption has been shown to reduce growth and foreign investment, as well as retarding development more generally (Mauro 1995; Wei 2000; Fisman & Svensson 2007). Environmentally, the effects of corruption are similarly damaging. It has been linked with deforestation, a reduced stringency of environmental regulations and negative conservation outcomes (Jepson et al. 2001; Fredriksson & Svensson 2003; Smith et al. 2003). Even flagship species, such as African elephants, show declines in corrupt countries (Smith et al. 2003). This is despite high levels of external scrutiny from project funders and other interest groups. Even where there is interest by locals in environmental projects, they often lack the ability to challenge corrupt officials (Whitten et al. 2001).

The most efficient and effective method of preventing non-native species incursion is through preborder risk assessment and management (Keller et al. 2007; Springborn et al. 2011; Kumschick & Richardson 2013). Risk assessment protocols have been shown to produce a net economic benefit, even accounting for type II errors (false positives) (Keller et al. 2007). Financial benefits are positive in both the animal and plant trade (Springborn et al. 2011; Schmidt et al. 2012). Species and pathway specific risk assessment tools and methodologies have been developed that allow for transparent and robust predictions of risk posed by potential species introductions (Pheloung et al. 1999). Some of these tools have been empirically validated as effective in different ecoregions globally (Gordon et al. 2008). However, despite their utility, risk assessments are generally only informative for individual 3

species, or narrow pathways and taxonomic groups (e.g. Pheloung et al. 1999; Lester 2005). Inputs are usually based on species’ biological characteristics – for instance, breeding cycles, temperature requirements, natural enemies. Or in the case of pathways, risk characteristics – such as species associated with certain commodities. If commodities for export are high risk, the destination country will often rely on the implementation of pre-export sanitary and phytosanitary measures by the exporting state. Such a process is particularly important for small destination states such as Pacific islands with limited biosecurity capacity.

Governance may have a considerable impact on the implementation and efficacy of export biosecurity measures. Internationally there are a range of trade institutions and agreements intended to govern the movement of non-native species in trade. The World Trade Organization’s Agreement on the Application of Sanitary and Phytosanitary Measures sets out specific issues of human, animal and plant health and expounds detailed rules for coordinating policies. The International Plant Protection Convention (IPPC) 1952 is the WTO reference body for plant health. As of February, 2016 the IPPC has adopted 36 different international standards for phytosanitary regulations. WTO members are expected to import and export products with reference to these standards. For instance, New Zealand’s Ministry for Primary Industries requires that a completed phytosanitary certificate issued by the exporting country’s national plant protection organisation accompany all consignments of fresh fruit/vegetables imported to New Zealand. For this certificate to be issued, various activities must be undertaken such as: 

A visual inspection;



Molecular testing for non-visually detectable pest species;



Treatment against specific risk species;



Confirmation the commodity has been sourced from a pest free area;



The commodity has been harvested at a particular stage in its lifecycle. E.g. when it is physiologically immature so seeds will not germinate; and/or



Specific transit requirements.

4

The implementation of a well-developed and technical regulatory regime such as this requires a high level of governmental institutional capacity. Whilst this implementation may be feasible for countries with effective government and good regulatory systems, countries lacking this capacity may implement regulatory requirements with lesser efficacy. Exporters or regulatory bodies may also spuriously claim compliance to facilitate rapid export. Further, poorly paid or trained biosecurity agents may apply management protocols poorly, or not implement them at all. Prior research has investigated a variety of recipient region indicators as predictors of invasibility, including development and socio-economic factors (Hulme 2009; Gallardo & Aldridge 2013). Given the importance of trade to the introduction of invasive species, it is surprising that no studies to our knowledge have investigated how donor region governance acts to modulate non-native species trade risk. Here we assessed the influence of governance and development – including corruption – on the rate of non-native species incursions associated with international trade across three commodity pathways – vegetable material, timber and vehicles.

2. Methods Interception data New Zealand’s Ministry for Primary Industries provided our interception data. We used data available from 2002-2011 within our three pathways (see below) – a total of 47,328 interception records.

Indicator data Corruption data were sourced from the World Bank’s World Governance Indicators (WGI) developed by Kaufman et al. (2010) These data can be found at the website: www.govindicators.org. The WGI consist of six composite indicators of broad dimensions of governance covering over 200 countries (Kaufmann et al. 2010). These indicators are based on several hundred variables obtained from 31 different data sources, capturing governance perceptions as reported by survey respondents, nongovernmental organisations, commercial business information providers, and public sector organisations worldwide (Kaufmann et al. 2010). According to Kaufman et al. (2010), these 5

perceptions of governance are organised into six clusters corresponding to the six broad dimensions listed by Kaufmann et al. (2010). For each of these clusters a statistical methodology known as an Unobserved Components Model is used to: (1) standardise the data from diverse sources into comparable units; (2) construct an aggregate indicator of governance as a weighted average of the underlying source variables; and (3) construct margins of error that reflect the unavoidable imprecision in measuring governance.

All indicators used in our risk assessment were re-scaled using the function rescale() implemented in the R package scales (R Core Team 2015; Wickham 2015) to have a specific minimum (0) and maximum (1). For example, the governance indicator ‘regulatory quality’ originally ranges from -2.5 (low regulatory quality) to 2.5 (high regulatory quality). Therefore, the rescaled data corresponds to (0) low regulatory quality and (1) high regulatory quality.

Trade data Data on international trade values were sourced from Statistics New Zealand. The NZ.Stat web tool allowed for the generation of import and export tables of raw trade data over the period investigated. Import values for overseas merchandise trade are given on a cost, insurance freight basis.

Statistics New Zealand organises its import and export data based on the New Zealand Harmonised System Classification 2012 (NZHSC). The NZHSC is a document based on the World Customs Organisation’s Harmonised Commodity Description and Coding System (HS). The international HS follows a hierarchical structure, comprising 21 sections, 98 chapters, 1,231 headings and 5,212 subheadings. This structure is further broken down into approximately 14,500 statistical keys for New Zealand purposes.

6

Pathway selection – Vegetable material, Timber and Vehicles We selected the three pathways analysed based on evidence that these pathways are particularly high risk. Studies have found that vegetable material such as flower cuttings and nursery stock are important vectors of non-native species (Kraus et al. 1999; Work et al. 2005; Liebhold et al. 2012).

Although rates of non-native species interceptions on raw timber are typically low (Smith et al. 2007), introductions of raw timber and wood items still pose significant risks given the ability of fungi and insects to decimate forests (Loo 2008; Kovacs et al. 2010; Roy et al. 2014).

The import and export of vehicles is a major commodity trade worldwide. Serious pests have been traced to vehicle shipments. During 2003 a male Asian gypsy moth (Lymantri dispar) was trapped as part of a New Zealand monitoring program for this pest (Biosecurity New Zealand 2005). At the time is was estimated that were it to establish it would cost up to NZ$80 million per year to control (Biosecurity New Zealand 2005). A trace back on the likely source of entry to New Zealand concluded that the gypsy moth probably originated in Japan arriving in New Zealand as an egg mass on a used vehicle shipment (Biosecurity New Zealand 2005). A further gypsy moth egg mass was intercepted during a border check on an imported Japanese car at Auckland Port in 2012.

Pathway catergorisation The interception database provided by New Zealand’s Ministry for Primary Industries (MPI) uses a classification system for imports different from that of the NZHSC. In order to maintain consistency among databases we simplified our interception dataset into pathways directly comparable with NZHSC categories. This was done for two reasons: (1) The NZHSC is based on an accepted international standard, meaning our methods can be more easily reproduced; and (2) Using NZHSC categories allowed the creation of an interception index.

As noted previously, pathways chosen were: Vegetable material, Timber and Vehicles (Section 2, 9 and 17 of the NZHSC respectively). In order to create these simplified pathways we combined several 7

of the MPI defined pathways on the basis that the MPI defined pathway fits within the broad NZHSC categorisation (Table 1).

Table 1 | Data used from Ministry for Primary Industries and Statistics New Zealand databases when creating pathway and interception index Pathway

New Zealand Harmonised System Classification

Ministry for Primary

chapters combined (Trade $)

Industries pathways combined (Interception data)

Vegetable

-

Material NZHSC Section 2

-

Chapter 6: Trees and other plants; live; bulbs, roots -

Cut flowers/foliage;

and the like; cut flowers and ornamental foliage;

Fresh produce;

-

Chapter 7: Vegetables and certain roots and tubers; -

Medicine/herbal;

edible;

Nursery stock;

-

Chapter 8: Fruit and nuts, edible; peel of citrus -

Plant material; and

fruit or melons;

Seeds/grain.

-

-

Chapter 10: Cereals;

-

Chapter 12: Oil seeds and oleaginous fruit; Total interceptions: miscellaneous grains, seeds and fruit, industrial or 44,207 medicinal plants; straw and fodder; and

-

Chapter 14: Vegetable plaiting materials; vegetable products not elsewhere specified or included.

Timber

-

Chapter 44: Wood and articles of wood; wood -

Bamboo, cane, rattan

charcoal; and

items;

NZHSC Section -

Chapter 46: Manufactures of straw, esparto or -

Basketware;

9

other

Timber;

plaiting

wickerwork.

materials;

basketware

and -

Wood items; and

-

Willow.

8

Total interceptions: 1,281

Vehicles

NZHSC Section

-

Chapter 87: Vehicles; other than railway or -

Car parts;

tramway rolling stock, and parts and accessories -

Tyres;

thereof.

-

Used vehicle; and

-

New vehicle.

17

Total interceptions: 1,840

Detection Index Raw interception values do not provide a relative measure of risk. Two countries’ commodities imported to New Zealand over time may result in a similar number of interceptions. However, the volumes of trade from the countries may be vastly different. For example: Exports by Country A and B to New Zealand in 2005 may both result in 1,000 interceptions of non-native species. Country A exports $NZ10 million worth of the commodity, whilst Country B exports $NZ5 million. The relative risk of non-native interceptions is different.

In order to determine the relative risk of commodity imports from different trade partners we developed a ‘detection index’. We did this by using the following formula:

(Total interceptions in category / Total trade ($NZ) in category) * 10,000,000.

For example for the pathway of vegetable material coming from Australia in 2008: 9

777 (vegetable material interceptions) / $NZ 232,304,266 (value of trade in selected NZHSC categories) * 10,000,000 = 33.4

In order for a country to be included in our analyses there must have been trade: (1) Of value > $NZ 1 million; (2) In one of the three composite pathways (Vegetable Material, Timber, Vehicle); and (3) Over the period analysed (2002-2011). The > $NZ 1 million limitation was to ensure there has been a sufficient volume of trade to provide a robust interception index.

Statistical Analyses All data analyses were performed in R v.3.2.1 (R Core Team 2015). Firstly, we used a linear mixed effect model (LMM) implemented in the package lme4 (Bates et al. 2015) to assess the likelihood of countries exhibiting a range of governance and development indicators to donate exotic species to New Zealand. For each commodity pathway, the detection index was the response variable and each of the governance and development indices and time (2002-2011) were set as predictors in the model structure. Country was set as a random effect. Secondly, we performed a parametric bootstrapping procedure for each of the three fitted models (n=1000 simulations) to provide a robust model estimation and account for any potential influential data and outliers in our dataset. Model coefficients (b) were bootstrapped using the function bootMER() and confidence intervals using the function boot.ci(), both implemented in the package lme4 (Bates et al. 2015). Then, we carried out a stepwise model selection to reduce model complexity. Each predictor in the general model was assessed using the function Anova() implemented in the package car (Fox & Weisberg 2010) and removed from the model structure if p > 0.100. The model fit of the general model was then compared with the restricted model using the generic function anova(). Because the global p-values provided in the stepwise model selection is larger than the simulation based parameter, we adopted a conservative approach and retained predictors exhibiting p-values < 0.100.

3. Results and discussion

10

Corruption is a significant impediment to economic and social development. It has been demonstrated to reduce foreign investment (Wei 2000) and retard development and growth (Mauro 1995; Fisman & Svensson 2007). So too does corruption impact the environment; having been linked with deforestation (Jepson et al. 2001), less efficacious environmental regulations (Fredriksson & Svensson 2003) and negative conservation outcomes for individual species (Smith et al. 2003). Recent evidence suggests that individuals have a greater likelihood of lying if they live in a country with high levels of corruption and fraud (Gächter & Schulz 2016).

The quantum of interceptions of non-native species even in a relatively small country such as New Zealand are significant (Fig. 1). Between 2002 and 2011 there have been 58,262 non-native species interceptions at the New Zealand border. As biosecurity protocols are not failsafe, these figures only represent a portion of non-native species incursions associated with trade.

Fig. 1 | Top 25 countries ranked by non-native species interceptions associated with commodities being imported to New Zealand over 2002-2011 – Note the ordinate scale is logarithmic. This table demonstrates the high volume of interceptions from commodities over time. It summarises interception data across all trade pathways i.e. not just vegetable material, timber and vehicles.

Risk assessment protocols are effective at predicting the likely invasibility of specific species (Pheloung et al. 1999; Gordon et al. 2008). Such protocols are economically efficient, even taking into 11

account type II errors (Keller et al. 2007). However, risk assessment protocols typically do not take countries’ levels of governance into account. This is important from the perspective of invasive species regulation as the quality of implementation of sanitary and phytosanitary regulations by institutions has a large effect on risk. For example, states may impose pre-shipping sanitary and phytosanitary requirements on consignments destined for import to their territory. Requirements may include: 

A visual inspection;



Molecular testing for non-visually detectable pest species;



Treatment against specific risk species;



Confirmation the commodity has been sourced from a pest free area; or



The commodity has been harvested at a particular stage in its lifecycle. E.g. when it is physiologically immature so seeds will not germinate

Such a regime is highly technical in nature and potentially beyond the capacity of agencies in lessdeveloped countries (Nuñez & Pauchard 2010). Further, protocols may be undermined by poorly implemented regulations as a result of undertrained or corrupt officials. This may be particularly so in poorer countries where non-shipping of consignments may have an especially detrimental effect on commodity producers. Finally, outgoing biosecurity relies to a certain extent on honesty from institutions and individual personnel. In this paper we hypothesised that commodities imported from poorly governed and corrupt countries would pose a greater risk than those from well-governed, noncorrupt countries. Our results demonstrate that exporting countries’ development and governance are strongly associated with non-native species interceptions. A significantly higher rate of exotic species interception was observed from countries with weaker regulatory quality, inferior ‘rule of law’, and poorer political stability and freedom from violence (Fig. 3). Corruption, though, was the only predictor able to estimate invasive species interception in all three commodity pathways. However, counterintuitively, less corrupt countries are associated with higher invasive species risk (Fig. 2, Fig. 3)

12

Fig. 2 | Effect of corruption on non-native species interceptions for the vegetable, timber and vehicle pathways. The black solid lines are median estimates; shaded regions are 95% confidence intervals. The y-axes range from 0 (least corrupt) to 1 (most corrupt) for indicators. Inner ticks on the x-axis represent individual countries.

Fig 3. | Effect of selected indicators on exotic species interception rates. For each export pathway (A-C), we show the bootstrapped coefficients estimated from the linear mixed-effect model (LMM). As coefficients were derived from standardized values they are directly comparable with each other, and are presented in order from highest to lowest values. Shaded squares indicate the mean, bold lines the standard errors, and non-bolded lines the 95% confidence intervals. Vertical lines mark b = 0, where predictors have no effect. Time was coded to show the changes in non-native species interception from 2002-2011. Significant predictors are labelled with “*”. Regular = regulatory quality

13

We can only speculate as to the reasons for the relationship between increasing corruption and decreasing incursions rates. We hypothesise the following explanations:

Licencing New Zealand’s Ministry for Primary industries operates a licencing system for exports from certain countries. Most organisations importing commodities into New Zealand do so in accordance with Import Health Standards (IHS). These IHS are issued under the Biosecurity Act 1994 and list requirements that must be met before risk goods may be imported into New Zealand. Typically, IHS require that agencies in the origin country undertake management for unwanted species - animals, plants, diseases - associated with the commodity. However, this relies on the institution in the country of origin being able to effectively implement best practice protocols. These may include chemical control, molecular genetic analysis for identification purposes, or authentication by qualified veterinarians of pest free status. Major factors mitigate against poorly governed or developing countries implementing this sort of standard. For instance the lack of a stable scientific community, poor levels of education, low resolution scientific data on local ecosystems and lack of public awareness of the problem of invasive species (Nuñez & Pauchard 2010).

Presumably in recognition of these realities, New Zealand’s Ministry for Primary Industries prescribes that companies importing commodities to New Zealand from certain regions do so under a licence. One of the terms of this licence is that the company is responsible for sanitary and phytosanitary protocols associated with commodities for import. It effectively institutes a dual-layered protective system for New Zealand. The first layer is protocols implemented by government authorities in the country of origin. If these are ineffective, then a second layer of protection is provided by the licenceimposed duty for import/export organisations to institute effective sanitary and phytosanitary protocols. This second layer is particularly effective as there is a strong economic incentive for private entities importing goods to New Zealand to ensure the pest free status of their imports – the risk of licence loss.

14

Unfortunately, empirical validation of this hypothesis is difficult as the details of granted import licences are not readily available. However, this approach has several prima facie benefits: First, it insures against the risk of poor origin country biosecurity protocols. Second, it circumvents the risk of causing offence to, or trade disputes with, the origin country over issues such as poorly implemented biosecurity measures, or the pest free status of certain areas within the country of origin. Finally, costs of invasive species are usually a market externality – their management largely financed by government institutions and ultimately taxpayers. This licencing approach internalises the cost of invasive species in the market system, helping to ameliorate a significant market failure. Sensibly cost internalisation is at the most efficient point in the invasive species management cycle – before entrainment by a transport vector and prior to establishment in a new area.

Trade partners and economic activity Lack of corruption has been shown to be strongly associated with high economic activity (Mauro 1995). Many authors have convincingly demonstrated that high levels of economic activity and trade increases regions’ invasibility (Margolis et al. 2005; Westphal et al. 2008; Hulme 2009). The volume of commodity demand, length and variety of transport networks and number of ports of entry interact to change the risk of unintentional non-native species incursions. It may be that the relationship we describe here between decreased corruption and an increased non-native species interception rate is not to do with the levels of corruption of exporting states per se. Instead, it may be the relationship is driven by high levels of economic activity, which is in turn strongly associated with low levels of corruption. Put another way, countries with high levels of economic development and trade have more invasive species able to be entrained in trade pathways; effectively acting as reservoirs from which invasive species spill over into the global trade system.

Organised vs disorganised corruption Corruption can operate at different levels. It may be an increased likelihood to lie by members of the public (Gächter & Schulz 2016), or high level, institutionalised corruption by well-organised groups 15

such as crime syndicates or political cadres. The corruption data used in this study were taken from Kaufmann et al. (2010) where corruption is defined as ‘capturing 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’. It is possible that our results reflect a difference between organised, and disorganised corruption. For instance, countries may be classified in the 30th percentile in terms of corruption, but also be highly organised, major trading nations with good governmental effectiveness. Therefore, although a country may be classified as corrupt, this may not have an effect on the implementation of regulatory policy such as biosecurity protocols for goods being exported. In fact, in such countries it would be in the interests of corrupt officials to ensure an effectively functioning trading system in order to provide opportunities to siphon profits on a larger scale.

Future research The field of risk assessment and management using biological and environmental data is welldeveloped and improving rapidly. Socio-economic indicators such as levels of corruption are less utilised in the scientific literature, particularly in invasion biology. This is surprising given invasive species risk sits at the interface of disciplines including economics, ecology, law and sociology. The corruption pattern observed in our study was unexpected, future research in this area could usefully focus determining the causative driver of this pattern. Utilising data from disparate disciplines will only serve to better calibrate models that inform biosecurity agencies’ management of risk, thereby reducing the environmental and economic costs of invasive species.

16

References Aukema, J.E. et al., 2011. Economic impacts of non-native forest insects in the continental United States. PLoS ONE, 6, p.e24587. Biosecurity New Zealand, 2005. $80 Million Pest Eradicated. Biosecurity, 1. Bates D. et al., 2015 lme4: Linear mixed-effect models using Eigen and S4. R package version 1.1-8, URL: http://CRAN.R-project.org/package=lme4. DiTomaso, J., 2000. Invasive weeds in rangelands: Species, impacts, and management. Weed Science, 48, pp.255–265. Fisman, R. & Svensson, J., 2007. Are corruption and taxation really harmful to growth? Firm level evidence. Journal of Development Economics, 83, pp.63–75. Fox J. & Weisberg S. 2010. An R Companion to applied regression, Second Edition. Sage Publication, California. Fredriksson, P.G. & Svensson, J., 2003. Political instability, corruption and policy formation: the case of environmental policy. Journal of Public Economics, 87, pp.1383–1405. Fritts, T.H. & Rodda, G.H., 1998. The Role of Introduced Species in the Degradation of Island Ecosystems : A Case History of Guam. Annual Review of Ecology and Systematics, 13, pp.113– 140. Gächter, S. & Schulz, J.F., 2016. Intrinsic honesty and the prevalence of rule violations across societies. Nature, pp.1–11. Gordon, D.R. et al., 2008. Consistent accuracy of the Australian weed risk assessment system across varied geographies. Diversity and Distributions, 14, pp.234–242. Hulme, P.E., 2009. Trade, transport and trouble: managing invasive species pathways in an era of globalization. Journal of Applied Ecology, 46, pp.10–18. Jepson, P. et al., 2001. The end for Indonesia’s lowland forests? Science, 292, p.859. Kaufmann, D., Kraay, A. & Mastruzzi, M., 2010. The Worldwide Governance Indicators Methodology and Analytical Issues, Keller, R.P., Lodge, D.M. & Finnoff, D.C., 2007. Risk assessment for invasive species produces net bioeconomic benefits. Proceedings of the National Academy of Sciences of the United States of 17

America, 104, pp.203–207. Kovacs, K.F. et al., 2010. Cost of potential emerald ash borer damage in U.S. communities, 2009– 2019. Ecological Economics, 69(3), pp.569–578. Kraus, F. et al., 1999. Eleutherodactylus Frog introduction to Hawaii. Herpetological Review, 30(1), pp.21–25. Kumschick, S. & Richardson, D.M., 2013. Species-based risk assessments for biological invasions: advances and challenges. Diversity and Distributions, 19, pp.1095–1105. Lester, P.J., 2005. Determinants for the successful establishment of exotic ants in New Zealand. Diversity and Distributions, 11, pp.279–288. Liebhold, A.M. et al., 2012. Live plant imports: The major pathway for forest insect and pathogen invasions of the US. Frontiers in Ecology and the Environment, 10, pp.135–143. Loo, J., 2008. Ecological impacts of non-indigenous invasive fungi as forest pathogens. Biological Invasions, 11, pp.81–96. Margolis, M., Shogren, J.F. & Fischer, C., 2005. How trade politics affect invasive species control. Ecological Economics, 52, pp.305–313. Mauro, P., 1995. Corruption and Growth. The Quarterly Journal of Economics, 110, pp.681–712. Nuñez, M.A. & Pauchard, A., 2010. Biological invasions in developing and developed countries: Does one model fit all? Biological Invasions, 12, pp.707–714. Pejchar, L. & Mooney, H. A, 2009. Invasive species, ecosystem services and human well-being. Trends in Ecology & Evolution, 24, pp.497–504. Pheloung, P., Williams, P. & Halloy, S., 1999. A weed risk assessment model for use as a biosecurity tool evaluating plant introductions. Journal of Environmental Management, 57, pp.239–251. Pimentel, D., Zuniga, R. & Morrison, D., 2005. Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecological Economics, 52, pp.273– 288. R Core Team. 2015 R: A language and environment for statistical computing. R Foundation for Statistical Compution, Vienna, Austria. URL http://www.R-project.org. Roman, J. & Darling, J. A., 2007. Paradox lost: genetic diversity and the success of aquatic invasions. 18

Trends in Ecology and Evolution, 22, pp.454–464. Roy, B.A. et al., 2014. Increasing forest loss worldwide from invasive pests requires new trade regulations. Frontiers in Ecology and the Environment, 12, pp.457–465. Scalera, R., 2010. How much is Europe spending on invasive alien species? Biological Invasions, 12, pp.173–177. Schmidt, J.P., Springborn, M. & Drake, J.M., 2012. Bioeconomic forecasting of invasive species by ecological syndrome. Ecosphere, 3(5), p.Art 46. Smith, R.J. et al., 2003. Governance and the loss of biodiversity. Nature, 426, pp.67–70. Springborn, M., Romagosa, C.M. & Keller, R.P., 2011. The value of nonindigenous species risk assessment in international trade. Ecological Economics, 70(11), pp.2145–2153. Transparency International Corruption Perceptions Index 2015 (Transparency International, Berlin, 2015). Vilà, M. et al., 2010. How well do we understand the impacts of alien species on ecosystem services? A pan-European, cross-taxa assessment. Frontiers in Ecology and the Environment, 8(3), pp.135–144. Vitousek, P. & D’Antonio, C., 1997. Introduced species: A significant component of human-caused global environmental change. New Zealand Journal of Ecology, 21, pp.1–16. Wei, S.-J., 2000. How taxing is corruption on international investors? The Review of Economics and Statistics, 82, pp.1–11. Westphal, M.I. et al., 2008. The link between international trade and the global distribution of invasive alien species. Biological Invasions, 10, pp.391–398. Whitten, T., Holmes, D. & MacKinnon, K., 2001. Conservation biology: A displacement behavior for academia? Conservation Biology, 15, pp.1–3. Wickham H. 2015 Scales: Scale Functions for Visualization. R package version 0.2.5. http://CRAN.R-project.org/package=scales. van Wilgen, B. et al., 2008. A biome-scale assessment of the impact of invasive alien plants on ecosystem services in South Africa. Journal of Environmental Management, 89(4), pp.336–49. Work, T.T. et al., 2005. Arrival rate of nonindigenous insect species into the United States through 19

foreign trade. Biological Invasions, 7, pp.323–332.

20