Vulnerability of Small Island Developing States to Natural Disasters

Abstract

Small island developing states (SIDS) have been identified as particularly vulnerable to natural disasters and climate change. However, although SIDS have similar geographical features, natural hazards produce different outcomes in different states, indicating variation in vulnerability. The objective of this article is to explore the sources of this variation. With the point of departure in theories about how political institutions affect adaptive capacities, this article sets out to investigate whether government effectiveness has an impact on the vulnerability of SIDS. While claims over the importance of institutions are common in the literature, there is a lack of empirical accounts testing the validity of such claims. This shortcoming is addressed by this study’s time-series cross-sectional analysis using data from the International Disaster Risk database and the Quality of Government data set. The results show that government effectiveness has strong and significant effects on the number of people killed and affected by natural disasters.

Keywords

democracy government effectiveness natural disasters political institutions small island developing states vulnerability

Small island developing states (SIDS) are generally held to be particularly vulnerable to natural hazards and disturbing weather events (Wong, 2011). Development challenges associated with exposure to natural hazards—for example, loss of lives, livelihoods, and shelter—have even prompted the proclamation of 2014 as the year of SIDS (UN, 2014). Due to SIDS’ geographical location in hazard-prone regions, but also their relative isolation, limited physical size, and a concentration of population along coastal zones, policy makers and scholars argue that SIDS require special attention and support in adapting to increasingly severe weather events (Pelling & Uitto, 2001; UN, 2014). However, despite a generally dismal outlook, similar natural hazards in fact produce widely different outcomes in different countries, indicating great variation in vulnerability. While some SIDS seem to cope and adapt fairly well, others suffer tremendously. That is, the impact of natural hazards of the same physical magnitude ranges from going more or less unnoticed or causing only small disturbances to resulting in severe catastrophes (Adger, Hughes, Folke, Carpenter, & Rockström, 2005). The overall objective of this article is to explore the sources of this variation further. This is done by critically reviewing existing literature on island states and vulnerability and then testing if the claims put forward in this literature are empirically valid. In particular, the study tests previous claims holding governance and political institutions to be of crucial importance in explaining variation in vulnerability. The chosen methodological approach is a time-series cross-country quantitative analysis where we, in order to explore the sources of variation in vulnerability, use statistical data on disaster outcomes as well as on a number of potential explanatory factors. While such analyses in general—and the use of aggregated data and indices in particular—of course, have limitations as regards analytical depth and nuance, we believe this approach provides a much-needed contribution to an otherwise case-study dominated field of research. While the empirical focus is on SIDS, the article certainly has the potential to also shed light on issues of disaster risk reduction and adaptive capacity more generally.

More specifically, in order to test the validity of the claims put forward in the existing literature, this study aims to look closer at the association between the impact of natural hazards and factors held to act as a buffer and moderate the effects. That is, the empirical analysis sets out to investigate whether islands experiencing milder impacts are relatively well equipped in terms of the factors held to reduce vulnerability, especially in terms of certain characteristics related to governance and political institutions. The choice of small island developing states is motivated by the fact that they constitute a hard test of the explanatory power of previous claims. Investigating variation in impact within this natural hazard prone group of countries hence gives us the leverage to test the strength of the generic claims about adaptive capacity and vulnerability.

The article proceeds as follows: The next section provides a brief overview of extant claims about factors exacerbating or reducing the impacts from extreme weather events. The main drivers and buffers are then summarized and scrutinized empirically. The empirical part investigates the association between the impacts of hazards (number of people killed or affected) and the factors held to increase or reduce the risk of such impacts. While previous literature identifies a number of such factors, empirical tests are scarce.

Vulnerability and Natural Hazards

Natural hazards are normally defined as potentially damaging physical events or weather phenomena that may cause the loss of life or injury, property damage, social and economic disruption, or environmental degradation. For example, hazards often result in disease outbreaks, weather-related and geophysical events including floods, high winds, landslides, droughts or earthquakes, economic volatility, or conflict-related shocks such as outbreaks of fighting or violence (DFID, 2011; Velasques, Bonapace, Srivastava, & Mohanty, 2012). The impact of natural hazards on human well-being has in recent years been unprecedented. Earthquakes, violent winds, floods, and droughts have had severe consequences for millions of people worldwide, presenting an increasingly significant challenge for development and poverty alleviation. For example, in 2010, natural hazards affected more than 200 million people, caused nearly 270,000 deaths, and resulted in 110 billion USD in damages. In 2011, the first famine of the 21st century occurred as drought hit the Horn of Africa (DFID, 2011). Storms alone have in the last decade killed roughly 175,000 people and have affected approximately 400 million (Beck et al., 2012). SIDS, as a group of low-lying coastal territories, have been identified by the United Nations as particularly prone to adverse weather events.¹ For example, many SIDS face special disadvantages associated with small size, insularity, remoteness, and dependency on international assistance, which make them particularly vulnerable (Nath, Roberts, & Madhoo, 2011; Wong, 2011). As such, island states—and especially SIDS—are often considered to be more vulnerable to economic, political, or environmental shocks (Briguglio, 1995; Pelling & Uitto, 2001). In terms of the economy, island states are expected to suffer from greater output volatility and greater volatility in terms of trade, which might spur more intense resource exploitation. Overexploitation of resources may lead to severe and sometimes irreversible environmental damage, as the stories of Easter Island and Pitcairn show (Diamond, 2005). It has also been pointed out that the lack of diversity in the productive base of island states’ economies can be assumed to have negative effects on their resilience to disasters (Pelling & Uitto, 2001). In terms of politics, the personal and informal character of political interaction generally found in small polities might also make SIDS more vulnerable to nepotism, cronyism, patronage, and political clientelism (Baladacchino, 2013; Ott, 2000; Srebrnik, 2004). Finally, the geographical location and environmental conditions of SIDS in some instances even imply that natural hazards threaten the very survival and existence of some small islands (Yamamoto & Esteban, 2014). But how come hazards result in effects of so vastly differing scale in different countries, and how can we understand this?

In general, whether a hazard turns into a disaster, a “serious disruption of the functioning of a community or a society causing widespread human, material, economic, or environmental losses which exceed the ability of the affected community/society to cope using its own resources” is argued to depend on the vulnerability of a system or society (Brown, Crawford, & Hammill, 2006, p. 3). There are certainly plenty of definitions of vulnerability (see Birkmann, 2006). In fact, already in 1981, Timmerman posited that “vulnerability is a term of such broad use as to be almost useless for careful description at the present, except as a rhetorical indicator of areas of greatest concern” (p. 17). Similarly, Liverman (1990, p. 27) argued that vulnerability “has been related or equated to concepts such as resilience, marginality, susceptibility, adaptability, fragility, and risk,” and Füssel (2007, p. 155) adds “exposure, sensitivity, coping capacity, criticality, and robustness” to this list. Other conceptualizations emphasize “multiple exposure,” implying that all forms of threats—for example, climate change, globalization, poverty, and epidemics—tend to converge and affect those with the least resources or capacities to deal with the challenges they confront (Kelman, Gaillard, & Mercer, 2015, p. 23).

Nevertheless, although turned into somewhat of a buzzword and being subjected to significant conceptual stretching, definitions tend to converge around a definition similar to the one put forward by the Intergovernmental Panel on Climate Change (2014, p. 5):

The propensity or predisposition to be adversely affected. Vulnerability encompasses a variety of concepts and elements including sensitivity or susceptibility to harm and lack of capacity to cope and adapt.

In a similar vein, Wisner—a pioneering scholar within risk management—argues that vulnerability is “the characteristics of a person or group and their situation that influence their capacity to anticipate, cope with, resist and recover from the impact of a natural hazard” (Wisner, Blaikie, Cannon, & Davis, 2004, p. 11).

While several definitions exist, there is hence a general consensus that natural hazards do not cause disasters by themselves: “It is the combination of an exposed, vulnerable, and ill-prepared population or community with a hazard event that results in a disaster” (Shaw, Pulhin, & Jacqueline Pereira, 2010, p. 2). Disaster vulnerability hence determines a society’s ability and capacity to cope with disturbances and moderates the outcome to ensure benign or only small-scale negative consequences (Manyena, 2006).

More than discussing various definitions of the term vulnerability, a lot of research has focused on assessing the factors determining levels of vulnerability. To start with, there has for quite some time been a clear move toward trying to take the “naturalness” out of natural disasters. And while the research field still suffers from under-theorization and a lack of causal, much less predictive, factors, some causal claims and roots of vulnerability have been suggested (Roe & Schulman, 2012).

First, there is a large literature suggesting that vulnerability is determined by social and structural factors such as levels of inequality, marginalization, and social injustice. Schröter, Polsky, and Patt (2005) illustrate this through the example of a famine and argue that it is more informative to look at the social, economic, and political marginalization of individuals and groups as the causes of that famine rather than focusing on the physical stress of the system as the cause for famine, such as drought. This example in turn follows the work of Sen (1981, 1990), who initiated an influential research agenda on the capacity of different political systems to deliver freedom from famine and other disasters. According to this literature, a political contract model of disaster prevention involves a political commitment from the government, recognition of the disaster as a political scandal by the people, and lines of accountability from the government to the people that enable this commitment to be enforced. In addition, repression of freedoms of expressions and association is argued to prevent civil society organizations from mobilizing to protest against or prevent potential disasters. It is also argued that the reverse case is clear: Human rights abuses are invariably an intimate part of disaster creation (De Waal, 1997).

According to Sen’s original argument about famines,

the diverse political freedoms that are available in a democratic state, including regular elections, free newspapers and freedom of speech, must be seen as the real force behind the elimination of famines. Here again, it appears that one set of freedoms—to criticize, publish and vote—are usually linked with other types of freedoms, such as the freedom to escape starvation and famine mortality (1990).

That is, democratic institutions are held to ensure that credible information exists and that governments respond to this information in a timely and effective manner. According to this logic, freedom of speech gives civil organizations a critical role calling to account those politicians and administrators who are failing to respond. In turn, elected representatives are expected to have a powerful incentive to respond: Their positions depend upon popular support and a member of parliament who allows a famine or other disasters to occur in his or her constituency is likely to face serious difficulties in getting reelected (De Waal, 1997).

Second, there is a lot of research on how environmental mismanagement and collective action failures severely worsen vulnerability. For example, research on ecosystem services shows that well-managed ecosystems act as strong buffers against harsh weather events. On the contrary, loss of biodiversity or environmental degradation increases the severity of natural hazards significantly. The reasons for why some societies or communities manage their ecosystems in a more sustainable manner than others in turn point to factors such as social capital and trust as crucial (Adger, 2003). Similarly, the common property research tradition shows that norms of trust and reciprocity can seriously reduce the risk of ending up in so-called social dilemmas or tragedy of the commons situations (Agrawal, 2001; Ostrom, 1990).²

Insights from this research tradition in turn lead us to the third category of potential explanations to variation in natural hazard impact, that is, governance and institutional and organizational features. This category in fact captures many of the claims accounted for earlier. The common property resource tradition, for example, emphasizes the importance of social, political, and economic organizations with institutions as mediating factors that govern the relationship between social systems and the ecosystems on which they depend (Dolésak & Ostrom, 2003). In addition, the research focusing on social marginalization and injustice clearly highlight the importance of governance and functioning political institutions. In the growing adaptive capacity literature, there is hence an emerging consensus about the integral role that institutions, governance, and management play in determining a system’s ability to adapt to climate change and severe weather events (Agrawal, 2008; Bele, Sonwa, & Tiani, 2014; Berman, Quinn, & Paavola, 2012; Brooks, Adger, & Kelly, 2005; Brown, Nkem, Sonwa, & Bele, 2010; Eakin & Lemos, 2006; Engle & Lemos, 2010; Gupta et al., 2010; Haddad, 2005; Ivey, Smithers, de Loë, & Kreutzwiser, 2004; Valdivieso & Andersson, in press; Weis et al., 2016; Yohe & Tol, 2002). This literature tends to build on the insights stemming from the seminal work of Ostrom (1990, 2010) and sets out to explore the multilevel governance challenges posed by increasingly complex and dynamic socioecological systems. Yet, although this literature has contributed greatly to our understanding of adaptive capacities—especially at the local level—it has also been criticized for not taking the role of the state seriously. More specifically, while local-level institutional arrangements certainly are crucial for increasing adaptive capacities, these arrangements tend to be nested within larger governance framework (Agrawal, 2001; Mansbridge, 2014). As such, while the literature on institutional adaptive capacities provides valuable knowledge about institutional design and diversity, it tends to overlook more fundamental political determinants and drivers of vulnerability and adaptation. For those reasons, this article focuses explicitly on the role played by institutions at a higher level within the political structure.

As Lebel, Tan Sinh, and Nikitina (2010, p. 132) put it, “reality is often much more political than the technical descriptions of disaster.” Characteristics of the political system are hence said to reduce or enhance vulnerability and thus directly affect the number of people killed and affected by natural hazards (Adger et al., 2005). Micromechanisms such as individuals’ decisions to engage in long-term investment in housing and infrastructure as well as taking stewardship of ecosystem’s buffering capacities, and their ability for voice and participation are, for example, held to depend on political order and institutional organization.

Taking our point of departure in this institutional argument, we aim to test the role of institutions on vulnerability. We follow Douglass North’s definition of institutions as “the rules of the game in a society or, more formally, the humanly devised constraints that shape human interaction” (1990, p. 3). Previous attempts to empirically investigate the influence of political institutions on vulnerability are mostly limited to case studies, while generalizable large-N analyses are scant. One of the first statistical studies was conducted by Kahn (2005), where he finds that in countries with lower income inequality, higher levels of democracy, and good institutions, including high regulatory quality, voice and accountability, rule of law, and stronger control of corruption, fewer people die as a result of natural and industrial disasters. Similarly, Raschky (2008) showed that government stability and higher investment climate result in lower human and economic losses. Our analysis aims to explore if the found patterns are applicable to SIDS, the UN-selected group of low-lying coastal countries, which particularly suffer from their exposure to adverse weather events and, therefore, the consequences of climate change.

Our hypothesis to test is:

Hypothesis 1: Higher institutional quality is related to fewer casualties as a result of natural disasters in SIDS.

The next section spells out the empirical strategy for investigating how institutional features affect vulnerability.

Data and Method

We keep the test simple and focus on establishing whether there is any association between the proposed explanatory factors and the outcomes in terms of number of people killed or affected by natural hazards. In short, we aim to find evidence of whether countries with stronger institutions and more robust governance systems have fewer people killed or affected by adverse weather events.

Of course, the validity of the results depends on how we measure and operationalize our concepts of interest. As regards the dependent variables—number of people killed or affected by natural hazards—we rely on the data from International Disaster Database, gathered by Centre for Research on the Epidemiology of Disasters. The database is compiled from various sources, including the United Nations agencies, nongovernmental organizations, insurance companies, research institutes, and press agencies (Guha-Sapir, Below, & Hoyois, 2016). The two main indicators we use are total number of people affected by disasters per year, which is a sum of all people injured, homeless, and affected,³ and number of people killed as a result of disasters per year. The data for both indicators are positively skewed, as there are substantially more cases with fewer human suffering and casualties than there are cases of large-scale disasters. To improve model fit and distribution of the residuals, we log-transform the variables and conduct the analyses only for nonzero observations—that is, when people did suffer.⁴

The explanatory variables include government effectiveness, an indicator developed by the World Bank, the level of democracy which is a combined score of measures by Freedom House and Polity IV, suggested by Hadenius and Teorell (2005), gross domestic product (GDP) per capita and population, both taken from Penn World Trade statistics (Heston, Summers, & Aten, 2012),⁵ geographical position of a country from La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1999),⁶ and a time trend. Table A1 in the Appendix provides summary statistics for the variables used in the analysis.

The main independent variable, government effectiveness, is an aggregated index consisting of multiple individual indicators. It assigns a government effectiveness score for each country, reflecting opinions of various stakeholders, including public, private, and nongovernmental organization-sector experts. It captures “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, Kraay, & Mastuzzi, 2010, p. 4). Therefore, apart from bureaucratic quality and effectiveness of public administration, the indicator includes outcomes of government actions, such as delivery of public services. This aspect provides an opportunity to use this measure in evaluating governments’ readiness to cope with natural disasters, protect its populations, and provide public goods that decrease vulnerability, such as reliable infrastructure. Despite that the measure is often criticized for being rather broad,⁷ it nevertheless has a strong advantage over the existing alternatives⁸ regarding its coverage of SIDS countries. It is in fact the only indicator measuring government effectiveness in SIDS that extends wide enough across countries as well as long enough over time to perform a statistical analysis.⁹

The indicator gauging countries’ levels of democracy ranges from 0 to 10, where 0 stands for authoritarian states, while 10 corresponds to most democratic regimes. A general principle in measuring countries level of democracy is: The more democratic attributes a country has, the higher score on the democracy scale it receives. The specific democratic attributes covered by the indicator are determined by two original measures of democracy: Freedom House (Freedom House, 2013) and Polity IV (Marshall & Jaggers, 2013), which are averaged into a single index (Hadenius & Teorell, 2005). The democratic characteristics measured by Freedom House incorporate the amount of political rights enjoyed by the population, including the right to vote in elections, choosing representatives to decision-making authorities, the right to join political parties and organizations, and the right to compete for public office. They also include the extent of civil liberties that provide the opportunities for freedom of expression, freedom to participate in civil society organizations, demonstrations and engage in open public discussions, rule of law, and personal freedom from state interference. Democratic characteristics measured by the Polity IV project additionally account for the existence of constraints on the executive power and the presence of a system of checks and balances. The combined index has shown to perform better in terms of validity and reliability than its constituent parts separately (Hadenius & Teorell, 2005; Teorell et al., 2016). Including democracy in the list of explanatory variables allows us to control for freedom of expression, political representation, strength of civil society, and human rights, which in previous research have been found to decrease the number of disaster victims (see, e.g., Congleton, 2006). As such, taken together, these indicators capture the concept of a robust governance system and make it possible to test the theoretical claims about the role of institutions in reducing vulnerability.

GDP per capita in turn captures economic abilities of the state to invest into the necessary infrastructure, which can decrease vulnerability. Population size is expected to positively correlate with vulnerability. A measure of a country’s geographical position accounts for the proneness to certain weather conditions as well as helps to control for the unobserved heterogeneity. We also include a time trend to account for any effect time could have on our data.

As we are analyzing the political situation on SIDS, the data on the political variables are only available for independent countries, not territories under foreign rule, such as, for example, French Polynesia (France), Anguilla (the United Kingdom), or Aruba (the Netherlands). Therefore, due to data availability, we only use a sample of independent SIDS, which originally comprises 38 countries, but the data availability is limited to 35 countries in this study (see Table A2 for a full list of countries used in the analysis). Although choosing only SIDS constrains our sample regarding variability in geography, institutions, and vulnerability, there is still considerable variation within the SIDS group. The group of SIDS includes both regimes with developed democratic institutions, such as Barbados and Seychelles, and authoritarian states, such as Singapore and Cuba¹⁰; it includes both wealthy and poor nations and states with different geographical characteristics. For example, GDP per capita in the SIDS group ranges from 900 USD in Comoros and Tuvalu to 30,000 USD in Singapore, Trinidad and Tobago, Barbados, and Bahamas. Similarly, land areas vary from Tuvalu’s 30 km² to 452,860 km² in Papua New Guinea.

We use the data available across countries and over time to increase the sample size and obtain more precise estimates.¹¹ By employing time-series cross-sectional analysis, we have to deal with a number of problems inherent to panel data. Since we perform our analysis exclusively for the sample of vulnerable countries, adverse weather events happen more or less independently from previous occurrences, and therefore, we expect little serial correlation within the data. Indeed, Langrage multiplier tests confirmed that autocorrelation is not a problem in our models. Therefore, we can make use of the pooled ordinary least squares regression with panel corrected standard errors suggested by Beck and Katz (1995). This methodological technique uses all variation in the data for the analysis.

The equation to be estimated is

\ln Y it = α + β j X jit + e it,

(1)

where i denotes country, t denotes year, j denotes number of independent variables,

\ln Y it

is a natural logarithm of the dependent variable, α is an intercept,

X it

stands for the vector of independent variables, and e is an error term. All models successfully passed standard regression diagnostic.

Results and Discussion

We start analyzing the relationship between government effectiveness and the number of people killed and affected by natural disasters by looking at cross-country scattergrams. To do so, we construct the means for each country over the period available and plot the countries according to their aggregated scores. The scattergram in Figure 1 shows the distribution of countries according to their government effectiveness and the total number of people affected by natural disasters. It suggests that SIDS with high government effectiveness such as Singapore, Bahamas, and Barbados have fewer people suffering the consequences of natural disasters than countries with low government effectiveness, such as Haiti, Dominican Republic, Comoros, and Guinea-Bissau. The linear prediction is significant¹² and explains 17% of the variation in the data, as shown by the R².

Figure 1.

Government effectiveness and the number of people affected by natural disasters in SIDS.

The scatterplot in Figure 2 reveals a similar pattern, with respect to the number of people killed in natural disasters. It shows that in countries with high government effectiveness, such as Singapore and Antigua and Barbuda substantially less people die as a result of disasters than in countries with low government effectiveness, such as already mentioned Haiti, Comoros, and Guinea-Bissau. The relationship between government effectiveness and the number of people killed in disasters is significant, and the linear prediction explains 14% of the variation in the data.

Figure 2.

Government effectiveness and the number of people killed as a result of natural disasters in SIDS.

We proceed by formally testing the pattern of the relationship controlling for other factors that might stand behind the relationship. Table 1 presents the results of the tests from Equation 1. Models 1 and 2 show the results for the relationship between our independent variables and the total number of people affected by disasters in a given year, while Models 3 and 4 illustrate the link between the explanatory factors and the number of people killed as a result of disasters per year.

Table 1.

The Effect of Government Effectiveness on the Number of People Killed and Affected by Natural Disasters in SIDS.

	Number of people affected by disasters (log)		Number of people killed by disasters (log)
	Model 1	Model 2	Model 3	Model 4
Government effectiveness	−0.635***	−1.203*	−1.190***	−0.538*
	(0.180)	(0.534)	(0.185)	(0.247)
Number of disasters (log)		0.847*		0.750⁺
		(0.378)		(0.385)
Democracy		0.009		0.044
		(0.111)		(0.054)
GDP per capita (log)		0.784⁺		−0.297*
		(0.469)		(0.150)
Population (log)		0.243		0.207
		(0.165)		(0.142)
Latitude		3.860		1.408
		(4.021)		(2.306)
Year		−0.025		−0.027
		(0.063)		(0.033)
Constant	8.241***	49.249	2.565***	56.495
	(0.228)	(125.156)	(0.114)	(65.116)
Observations	k161	152	133	126
Adjusted R²	0.028	0.161	0.221	0.293
Number of countries	35	35	35	35

Note. Panel corrected SE in parentheses. SIDS = small island developing states.

***

p < .001. **p < .01. *p < .05. ⁺p < .1.

For each of the dependent variables, we start with the bivariate correlations first. Models 1 and 3 show the bivariate relationship between government effectiveness and the number of people affected and killed in adverse weather events. The relationship is negative and significant, indicating that higher government effectiveness is associated with lower number of people killed and affected by natural disasters. Adjusted R² is higher in the model with the number of people killed in natural disasters as dependent variable, implying that government effectiveness explains more variation in the data on the number of people killed than as regards the number of people affected by disasters.

In the next step, we investigate the results further, introducing the rest of the explanatory factors in the models. Models 2 and 4 show the relationship between government effectiveness and our dependent variables, controlling for number of disaster events occurring per year, democracy level, GDP per capita, population size, countries’ latitude, and a time trend.¹³ The results show that government effectiveness still exerts a negative significant effect on both our dependent variables despite the fact that other explanatory factors, deemed important in previous research, are accounted for. The effects are substantial. The model predicts that an improvement in the government effectiveness by 1 unit on the World Bank measurement scale from −2.5 to 2.5 is associated with a decrease in the number of people affected by natural disasters by 233 percent,¹⁴ and the number of people killed by 71 percent,¹⁵ holding the rest of the variables constant. Interestingly, however, among the control variables, it is only GDP per capita that has a significant effect. Contrary to the influential argument by Sen (1981, 1990), democracy does hence not seem to affect to what extent people are affected or killed by natural hazards in SIDS.

The predicted power of both models rises substantially compared with the bivariate regressions in Models 1 and 3, as indicated by the change in the adjusted R², and reaches 0.16 in the model with total number of people affected as dependent variable and 0.29 in the model with the total number of deaths as dependent variable. This implies that the selected set of predictors explains 16% of the variation in the number of people affected and 29% of the variation in the number of people killed as a result of disasters. The models have different number of cases due to the different number of nonzero observations in the data across years. However, most importantly, the amount of countries stays the same throughout the analysis. The composition of the country sample is slightly different in Models 1 and 2 compared with Models 3 and 4 due to data availability issues. In models with total number of people affected by natural disasters as dependent variable, Tuvalu is missing from the analysis, while in models with the number of people killed as a result of a disaster as dependent variable, Kiribati is missing.

We made a number of robustness checks on our results to make sure that the results are trustworthy. We performed the regressions with our dependent variables weighted by the population size and number of natural disasters occurring in a country per year. The results remained significant. We also performed the analysis with high-leverage observations—that is, Haiti and Tuvalu, which could potentially pose problem for the results—deleted. These two countries had been shown to have high influence on the results according to the leverage tests; however, they had not been identified as residual outliers. Additional tests showed that the results remained significant even when the high-leverage countries are removed from the models; therefore, we decided to keep them in the original analysis in order not to reduce the sample size. The results also remain robust in random effects models, where the estimates are averaged over the time-series and cross-sectional information in the data. The results from the robustness checks are documented in Table A3.

Taken together, the results thus lend support to our hypothesis and allow us to infer that in SIDS, higher government effectiveness is associated with substantially fewer people killed and affected by natural disasters.

Conclusion

SIDS have in recent decades been identified as particularly vulnerable to natural disasters and climate change. Violent winds, floods, and droughts have had severe consequences for millions of people and currently present an increasingly significant challenge for development and poverty alleviation in small island states. However, although islands tend to have similar geographical features, natural hazards produce widely different outcomes in different island states, indicating great variation in vulnerability. While some islands seem to cope and adapt fairly well, others suffer tremendously. That is, the impact of natural hazards of the same physical magnitude ranges from going more or less unnoticed or causing only small disturbances to resulting in severe catastrophes. The overall objective of this article was to explore the sources of this variation further. More specifically, with the point of departure in theories about how institutions affect collective action and adaptive capacities, this article set out to investigate how political institutions such as democracy and government effectiveness impact on the overall vulnerability of island states. The results lend support to our expectations, indicating that higher government effectiveness is indeed associated with a lower number of people killed and affected by natural disasters in the group of SIDS countries and the difference that government effectiveness makes seems to be quite substantial. Even small positive changes in government effectiveness are associated with twice the amount of people remained safe after the disaster and half the amount of people killed. Contrary to the influential argument by Sen and others, democracy, however, does not seem to affect the number of people killed or affected by natural hazards in SIDS when government effectiveness is also included in the model.

Since SIDS, as low-lying coastal territories, are prone to be hit by adverse weather events and the risk is only expected to grow due to climate change, the findings of this article have important policy implications. More specifically, when planning developing aid to countries suffering the risk of adverse weather events, donors have to obtain information not only about the potential ways to improve infrastructure, building codes, or other technical aspects of resilience, but they also need to take the political context where the projects are to take place into account. Policy actors have to be aware of the particular institutional challenges in the contexts where they are implementing their projects for the projects to deliver effective results. Projects aimed at reducing vulnerability and developing successful mitigation and adaptation strategies to climate change hence have to adapt to the political contexts in target countries. These political contexts, as shown by the present study, vary greatly between SIDS and have different implications for the success of reducing vulnerability. The implementation strategies, therefore, should bear different routines, depending on countries’ institutional quality and their capacity to deliver the results. In cases where institutional frameworks are weak, the usual methods of vulnerability reduction might not bring the desired success and have to be adjusted to better fit the context. For example, short-term solutions could imply disaster prevention through nongovernmental channels, such as adaptation and mitigation work of nongovernmental organizations and civil society. Long-term strategies in contexts with weak institutions should, first of all, aim at improving governance quality, as this is a root cause of countries’ vulnerability to natural disasters.

Our findings also imply that studies of vulnerability should focus more explicitly on the role played by governments and state authorities. As such, while previous research has provided important insights as regards local-level institutional design and diversity, there are reasons to believe that the workings of these local-level arrangements are crucially dependent upon the larger governance structures in which they are embedded. Although local-level institutions may be important, the role of the government may hence be even more important. Similarly, while scholars and policy makers are increasingly concerned with governance at the global level, that is, Earth systems governance, the findings of our study indicate that such a focus should also include an increased understanding of the role played by national governments. That is, despite that some scholars and policy makers tend to argue that increasing globalization and interdependence substantially affect the capacity of SIDS to govern the impact of disasters within its borders,¹⁶ our results indicate that political institutions at the national level do in fact indeed play an important role. However, while our study certainly lends support to the idea of “bringing the state back in” within studies of vulnerability, the specific causal mechanisms and dynamics behind how government institutions affect adaptive capacities need to be even further explored and specified. For such an endeavor, the methodological approach of this study, however, has some limitations and need to be complemented with more qualitative and focused comparisons and case studies. While the quantitative approach of this article has strengths as regards generalizability and the identification of general patterns of correlations, the aggregate indicators used risk losing out in terms of specificity and nuance. For example, while being a standard control variable, GDP at the aggregate level does not capture variation in the structure and coping capacities of local or traditional economies. Studies focusing more explicitly on causal mechanisms would therefore benefit from acknowledging the within-country variation in vulnerability—as well as in sources of vulnerability—that our study design cannot capture accurately. Finally, while our study focuses on how political institutions affect outcomes from natural hazards, we have not considered the fact that natural hazards also potentially have long-run effects on political systems. On this note, recent research has shown that while natural hazards certainly can have devastating effects on political development, causing increased instability and turmoil, natural hazards also have the potential to constitute an impetus for positive political changes such as democratization. Current research on the interplay between political institutions and natural hazards is, however, still in an infant stage and the question of what to expect politically from future increasingly extreme weather is in need of both further theoretical development and rigorous empirical analysis.

Footnotes

Appendix

Acknowledgments

The authors are grateful to Carola Betzold and Ilan Kelman for many helpful comments and suggestions at the early stages of the article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Swedish Research Council (grant number 2009-01866 and 2016-02119) as well as the Centre for Collective Action Research (CeCAR) at the University of Gothenburg, Sweden.

Notes

Author Biographies

Martin Sjöstedt is an Associate Professor at the Department of Political Science at the University of Gothenburg, and affiliated with the Quality of Government (QoG) institute and the Centre for Collective Action Research (CeCAR). His research is in the area of comparative politics and comparative political economy of developing countries and deals with politics of development, institutional theory, and natural resource management.

Marina Povitkina is a PhD candidate at the Department of Political Science and the Quality of Government Institute, University of Gothenburg. Her research is in the area of comparative environmental politics and focuses on the roles of democracy and quality of government in achieving sustainable development.

References

Adger

W. N.

(2003) Social capital, collective action, and adaptation to climate change. Economic Geography 79: 4.

Adger

W. N.

Hughes

T. P.

Folke

Carpenter

S. R.

Rockström

(2005) Social-ecological resilience to coastal disasters. Science 309(5737): 1036–1039.

Aginam, O. (2011). Climate change diplomacy and small island developing states. Development & Society. Retrieved from http://ourworld.unu.edu/en/climate-change-diplomacy-and-small-island-developing-states.

Agrawal

(2001) Common property institutions and sustainable governance of resources. World Development 29(10): 1649–1672.

Agrawal, A. (2008). The role of local institutions in adaptation to climate change. International Forestry Resources and Institutions Program (Working Paper No. W081-3). University of Michigan, School of Natural Resources and Environment.

Alliance of Small Island States. (2016). Alliance of small Island states. Retrieved from http://aosis.org/.

Baladacchino

(2013) Global tourism and informal labour relations: The small scale syndrome at work, London, England: Routledge.

Barnett

Campbell

(2015) Climate change and small island states: Power, knowledge, and the South Pacific, London, England: Routledge.

Beck

Katz

J. N.

(1995) What to do (and not to do) with time-series cross-section data. American Political Science Review 89(03): 634–647.

10.

Beck, M. W., Shepard, C. C., Birkmann, J., Rhyner, J., Welle, T., Witting, M., … Mucke, P. (2012). World Risk Report 2012. Alliance Development Works. Berlin.

11.

Bele, M. Y., Sonwa, D. J., & Tiani, A. M. (2014). Local communities vulnerability to climate change and adaptation strategies in Bukavu in DR Congo. The Journal of Environment & Development, 23(3): 331–335.

12.

Berman

Quinn

Paavola

(2012) The role of institutions in the transformation of coping capacity to sustainable adaptive capacity. Environmental Development 2: 86–100.

13.

Birkmann

(2006) Measuring vulnerability to natural hazards: Towards disaster resilient societies, Tokyo, Japan: United Nations University.

14.

Bolt, J., & van Zanden, J. L. (2013). The first update of the maddison project; reestimating growth before 1820. Retrieved from http://www.ggdc.net/maddison/maddison-project/home.htm.

15.

Briguglio

(1995) Small island developing states and their economic vulnerabilities. World Development 23(9): 1615–1632.

16.

Brooks

Adger

W. N.

Kelly

P. M.

(2005) The determinants of vulnerability and adaptive capacity at the national level and the implications for adaptation. Global Environmental Change 15(2): 151–163.

17.

Brown, O., Crawford, A., & Hammill, A. (2006). Natural Disasters and Resource Rights. Building resilience, rebuilding lives. International Institute for Sustainable Development.

18.

Brown

H. C. P.

Nkem

J. N.

Sonwa

D. J.

Bele

(2010) Institutional adaptive capacity and climate change response in the Congo Basin forests of Cameroon. Mitigation and Adaptation Strategies for Global Change 15(3): 263–282.

19.

Congleton

R. D.

(2006) The story of Katrina: New Orleans and the political economy of catastrophe. Public Choice 127(1–2): 5–30.

20.

De Waal

(1997) Famine crimes: Politics & the disaster relief industry in Africa, Bloomington: Indiana University Press.

21.

DFID (2011) Defining disaster resilience: A DFID approach paper, London, England.

22.

Diamond

J. M.

(2005) Collapse: How societies choose to fail or succeed, New York, NY: Viking.

23.

Dolésak

Ostrom

(2003) The commons in the new millennium: Challenges and adaptation, Cambridge, MA: MIT Press.

24.

Eakin

Lemos

M. C.

(2006) Adaptation and the state: Latin America and the challenge of capacity-building under globalization. Global Environmental Change 16(1): 7–18.

25.

Engle

N. L.

Lemos

M. C.

(2010) Unpacking governance: Building adaptive capacity to climate change of river basins in Brazil. Global Environmental Change 20(1): 4–13.

26.

Freedom House. (2013). Freedom of the world and freedom of the press. Retrieved from http://www.freedomhouse.org/.

27.

Füssel

H.-M.

(2007) Vulnerability: A generally applicable conceptual framework for climate change research. Global Environmental Change 17(2): 155–167.

28.

Guha-Sapir, D., Below, R., & Hoyois, P. (2016). EM-DAT: The CRED/OFDA International disaster database. Retrieved from http://www.emdat.be.

29.

Gupta, J., Termeer, C., Klostermann, J., Meijerink, S., van den Brink, M., Jong, P., Nooteboom, S., & Bergsma, E. (2010). The adaptive capacity wheel: A method to assess the inherent characteristics of institutions to enable the adaptive capacity of society. Environmental Science & Policy, 13(6), 459–471.

30.

Haddad

B. M.

(2005) Ranking the adaptive capacity of nations to climate change when socio-political goals are explicit. Global Environmental Change 15(2): 165–176.

31.

Hadenius, A., & Teorell, J. (2005, August). Assessing alternative indices of democracy (C&M Working Papers 6). Montreal, Canada: IPSA.

32.

Heston, A., Summers, R., & Aten, B. (2012). Penn World Table Version 7.1. Philadelphia, PA: Center for Comparisons of Production, Income and Prices at the University of Pennsylvania.

33.

Intergovernmental Panel on Climate Change. (2014). Intergovernmental panel on climate change. Climate change 2014. Impacts, adaptation and vulnerability. Part a: Global and sectoral aspects. Retrieved from http://www.ipcc.ch/ipccreports/tar/wg2/index.php?idp=8.

34.

Ivey

J. L.

Smithers

de Loë

R. C.

Kreutzwiser

R. D.

(2004) Community capacity for adaptation to climate-induced water shortages: Linking institutional complexity and local actors. Environmental management 33(1): 36–47.

35.

Kahn

M. E.

(2005) The death toll from natural disasters: The role of income, geography, and institutions. Review of Economics and Statistics 87(2): 271–284.

36.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The Worldwide Governance Indicators. Methodology and Analytical Issues. (World Bank Policy Reserach Working Paper No. 5430).

37.

Kelman

Gaillard

J. C.

Mercer

(2015) Climate change’s role in disaster risk reduction’s future: Beyond vulnerability and resilience. International Journal of Disaster Risk Science 6(1): 21–27.

38.

Komai, M. (2014). Pacific leaders share their views on climate diplomacy. Retrieved from https://www.sprep.org/sids/pacific-leaders-share-their-views-on-climate-diplomacy.

39.

Kostakos, G., Zhang, T., & Veening, W. (2014). Climate security and justice for small island development states: An agenda for action. Den Haag, Netherlands: The Hague Institute for Global Justice.

40.

La Porta

Lopez-de-Silanes

Shleifer

Vishny

(1999) The quality of government. Journal of Law, Economics, and organization 15(1): 222–279.

41.

Lebel

Tan Sinh

Nikitina

(2010) Adaptive governance of risks: Climate, water, and disasters. In: Shaw

(eds) Climate change adaptation and disaster risk reduction: Issues and challenges, Bingley, England: Emerald Group Publishing Limited, pp. 115–142.

42.

Liverman

D. M.

(1990) Vulnerability to global environmental change. In: Kasperon

R. E.

(ed.) Understanding global environmental change: The contributions of risk analysis and management, Worcester, MA: Clark University, pp. 27–44.

43.

Mansbridge

(2014) The role of the state in governing the commons. Environmental Science & Policy 36: 8–10.

44.

Manyena

S. B.

(2006) The concept of resilience revisited. Disasters 30(4): 434–450.

45.

Marshall, M. G., & Jaggers, K. (2013). Polity IV project: Political regime characteristics and transitions, 1800–2012. Retrieved from http://www.systemicpeace.org/.

46.

Nath

Roberts

J. L.

Madhoo

Y. N.

(2011) Saving small island developing states: Environmental and natural resource challenges, London, England: Commonwealth Secretariat.

47.

North

D. C.

(1990) Institutions, institutional change, and economic performance, Cambridge, England: Cambridge University Press.

48.

Ostrom

(1990) Governing the commons: The evolution of institutions for collective action, Cambridge, England: Cambridge University Press.

49.

Ostrom

(2010) Beyond markets and states: Polycentric governance of complex economic systems. Transnational Corporations Review 2(2): 1–12.

50.

Ott

(2000) Small is democratic: An examination of state size and democratic development, New York, NY: Garland.

51.

Pelling

Uitto

J. I.

(2001) Small island developing states: Natural disaster vulnerability and global change. Global Environmental Change Part B: Environmental Hazards 3(2): 49–62.

52.

Raschky

P. A.

(2008) Institutions and the losses from natural disasters. Natural Hazards and Earth System Sciences 8(4): 627–634.

53.

Roe

Schulman

P. R.

(2012) Toward a comparative framework for measuring resilience in critical infrastructure systems. Journal of Comparative Policy Analysis: Research and Practice 14(2): 114–125.

54.

Schröter

Polsky

Patt

A. G.

(2005) Assessing vulnerabilities to the effects of global change: An eight step approach. Mitigation and Adaptation Strategies for Global Change 10(4): 573–595.

55.

Sen

(1981) Poverty and famines: An essay on entitlement and deprivation, Oxford, England: Oxford University Press.

56.

Sen, A. (1990, June 14). Individual freedom as a social commitment. New York Review of Books. Retrieved from http://www.nybooks.com/articles/1990/06/14/individual-freedom-as-a-social-commitment/.

57.

Shaw

Pulhin

J. M.

Jacqueline Pereira

(2010) Climate change adaptation and disaster risk reduction: An Asian perspective. In: Shaw

Pulhin

J. M.

Pereira

J. J.

(eds) Climate change adaptation and disaster risk reduction: An Asian perspective, Bingley, England: Emerald Group Publishing Limited, pp. 1–18.

58.

Srebrnik

(2004) Small island nations and democratic values. World Development 32(2): 329–341.

59.

Teorell, J., Dahlberg, S., Holmberg, S., Rothstein, B., Khomenko, A., & Svensson, R. (2016). The quality of government standard dataset. Retrieved from http://www.qog.pol.gu.se.

60.

Timmerman

(1981) Vulnerability, resilience and the collapse of society. A review of models and possible climatic applications, Toronto, Canada: Institute for Environmental Studies, University of Toronto.

61.

UN. (2014). The international year of small island developing states 2014. Retrieved from http://www.un.org/en/events/islands2014.

62.

Valdivieso, P., & Andersson, K. P. (forthcoming). Local politics of environmental disaster risk management: Institutional analysis and lessons from Chile. Journal of Environment and Development.

63.

Velasquez, J., Bonapace, T., Srivastava, S., & Mohanty, S. (2012). Reducing vulnerability and exposure to disasters: Asia-Pacific disaster report 2012. UNESCAP and UNISDR. Bangkok.

64.

Weis, S. W. M., Agostini, V. N., Roth, L. M., Gilmer, B., Schill, S. R., Knowles, J. E., & Blyther, R. (2016). Assessing vulnerability: An integrated approach for mapping adaptive capacity, sensitivity, and exposure. Climatic Change, 136, 615.

65.

Wisner

Blaikie

Cannon

Davis

(2004) At risk: Natural hazards, people’s vulnerability, and disasters, 2nd ed. London, England: Routledge.

66.

Wong

P. P.

(2011) Small island developing states. Wiley Interdisciplinary Reviews: Climate Change 2(1): 1–6.

67.

Yamamoto

Esteban

(2014) Atoll Island States and international law, Berlin, Germany: Springer.

68.

Yohe

Tol

R. S.

(2002) Indicators for social and economic coping capacity—Moving toward a working definition of adaptive capacity. Global Environmental Change 12(1): 25–40.