African exceptions: democratic development in small island states

Small island states are much more likely to have democratic regimes than large continental states. This trend also holds across Africa, where the five island states with populations of 1.5 million or less are all rated at least ‘partly free’ by Freedom House. In this article we explore what it is about being a small island state that might explain this trend. Building on studies from other small island states, we find that the interaction between the two contextual factors is key to explaining their diversion from mainland trends in the African context. Specifically, ‘smallness’ leads to closer links between citizens and politicians in addition to more effective service delivery, while ‘islandness’ promotes community cohesion and provides a buffer against instability and conflict in neighbouring states. This results in a positive feedback loop that guards against authoritarian excess. Our focus on population size and geography thus adds to the existing studies of the contextual drivers of African democratisation.


Introduction
It has become a cliché to note that the process of democratisation has stalled in many sub-Saharan African countries. It is true that a small number of countries have made progress towards democracy, including Botswana, Ghana and Namibia. But many more states made initial headway only to suffer democratic backsliding, such as Burundi, Tanzania, and Uganda. Another group of countries including Angola, Cameroon, Equatorial Guinea, and Rwanda never achieved democratic status at all. Instead, multi-party elections were introduced in the absence of other meaningful reforms, and have served more to consolidate the position of incumbents than to challenge their hold on power. As a result, only 14% of African states are today rated as 'free', with 49% 'partly free' and 37% 'not free' (Freedom House 2020). Put another way, 86% of African countries feature significant elements of authoritarian rule.
There is one notable exception to this pattern, however. In 2020, none of the five African island states with less than 1.5 million inhabitants are rated as 'not free' by the Freedom House, and other democracy indexes are not far from this assessment. 1 While Cabo Verde, Mauritius and São Tomé and Príncipe are all classified as 'free', the Comoros and the Seychelles are ranked as 'partly free'. Moreover, recent general elections in the latter country resulted in the defeat of the longstanding ruling party and a peaceful transfer of power, and so we expect that the Seychelles will soon be upgraded to 'free'. While it is true that the Comoros do not perform as well as the other islands in democracy indexes, this is mainly due to persistent political instability and inter-island rivalries rather than authoritarian excess. Although our five cases comprise only a small group, it is nonetheless striking that-in contrast to the mainland-Africa's small island states are 60% 'free', while only 40% feature some degree of authoritarianism. This pattern has been apparent for some time, but there has been remarkably little research on them as a group.
The pattern in which small islands outperform large, continental states is not unique to Africa. As Fig. 1 shows, there are currently no island states with populations of less than 1.5 million that are not ranked at least partly free by Freedom House. To be clear, we are not saying that either 'smallness' or 'islandness' are necessary or sufficient conditions for democratisation. There are numerous large continental democracies, just as there are small autocracies. Nor are small island democracies always beacons of liberal freedoms, with some cases, including two in Africa, ranked as partly free. However, the higher likelihood of democracy in small island states and the absence of outright dictatorship suggest that there is something important about the interaction between smallness and islandness that helps to insulate against authoritarian rule.
The aim of this article is to explore the relationship between size, insularity and democracy. The article focuses on the five small African island states-Cabo Verde, the Comoros, Mauritius, São Tomé and Príncipe and the Seychelles-in order to identify the patterns that set them apart from continental and larger states. We explore five effects of being a small island state, two of which relate to population size, two relate to being islands, and one is a product of the interaction between size and location. First, the small size of these countries has enabled a more intimate 1 The Varieties of Democracy (V-DEM) project offers several democracy indexes varying from low (0) to high (1). The 2019 data shows that all island states outperform the regional average scores in the electoral democracy index (0.43) and in the liberal democracy index (0.3). The Comoros are a partial exception to this pattern and displays a more irregular trend: after standing above the regional average roughly between 2004 and 2018 its score slightly decreased in 2019 to 0.39 in the electoral democracy index, and to 0.18 in the liberal democracy index. Without denying the differences in quality of democracy across cases, both the V-Dem and the Freedom House data, suggest that the Comoros are closer to being an electoral democracy than to being an autocracy.
form of politics to emerge in which a greater proportion of the citizenry are in some way connected to the state either because they are directly incorporated in patronage networks and state employment, or because they know someone who is (Corbett 2015). Second, smallness makes it easier-though not always straightforward-for governments to establish effective states. In other words, the centre-periphery divide discussed by Herbst (2014) is neither so prevalent nor so damaging to the capacity of the state to provide public services and maintain order. The result is a positive feedback loop in which the extension of patronage and clientelism can include a greater share of citizens than would be possible in more populous countries. At the same time, the more intimate relationship between clients and patrons-with patronage ties running directly to senior politicians rather than through a series of intermediaries-means that the former often have greater influence in small societies, and are therefore better placed to impose informal checks and balances on what governments can do (Corbett and Veenendaal 2018).
Being an island also has important political ramifications. The first is that limited external influence creates a stronger sense of a shared identity (Anckar 2002(Anckar , 2008) and greater social capital (Congdon Fors 2014). When combined with the fact that a large proportion of the population of these islands arrived relatively recently from other parts of the world, meaning that it was clear that they were not autochthonous, insularity has constrained the emergence of the kind of damaging 'politics of belonging' that has exacerbated political violence in parts of mainland Africa (Nyamnjoh 2005). Secondly, the offshore location of these states means that they have not been impacted by the porous and often arbitrary colonial borders that are often argued to have impeded state building on the mainland (Englebert 2000;Herbst 2014). In turn, this has also insulated them from the destabilising impact of land borders with countries in the midst of civil war or political upheaval (Anckar 2002(Anckar , 2008. More specifically, they have been protected from what democratisation theorists call 'snowball' or 'demonstration effects' (Starr 1991)-although here it is the demonstration of authoritarianism and war rather than democratic government. Finally, we argue that it is the mutually reinforcing interaction between insularity and smallness that is particularly powerful-and can explain the relative absence of authoritarianism in our five cases. The significance of the interaction between these two factors is confirmed by the case of Madagascar, which is an island but has a population of over 26 million-more than Australia-and so does not count as 'small' according to our criteria. Although insularity has had some effects in Madagascar that are similar to our other cases, the large size of the population generates different challenges for democratic consolidation. The centre-periphery divide has more damaging effects, since the state is extremely weak and weakly entrenched (Levitsky and Way 2010: 277), while the development of pronounced social cleavages has contributed to instances of instability and political violence (Marcus 2016).
To substantiate these arguments, the article begins by reviewing the democratisation literature, especially where it is relevant to African states. The second section focuses on the nascent body of scholarship, mainly based on the experience of Caribbean and Pacific Island countries, that seeks to explain the disproportionate success of democratisation in small islands. We note a double absence: the literature on African democratisation tends to overlook smallness and islandness, while the literature on small islands tends to overlook the African cases. The third section introduces our cases, using an historical comparative analysis to trace how population size and island geography shaped political processes over time, and ultimately prevented the emergence of more autocratic forms of rule. However, although our cases are relatively democratic, we also uncover a type of politics characterised by hyper-personalism, pervasive clientelism, and dominant leadership that generates it its own challenges. The penultimate section then discusses why the interaction between size and insularity, rather than either factor on its own, is so important by drawing comparisons with both large island states (Madagascar) and small continental states. In the conclusion we highlight how our focus on hitherto neglected contextual factors-population size and island geography-adds to existing studies of the drivers of African democratization, and consider the implications for democratization processes in larger non-island states.

Democratisation in Africa
Explanations of the limited progress towards democracy in many African states fall into two broad categories. The first draws on the comparative literature on democratisation and emphasises the significance of factors often regarded as the 'pre-conditions' of democratic consolidation, including national wealth, ethnic homogeneity, institutional design, and colonial origin. The second highlights factors seen to be particularly prevalent on the continent, such as neo-patrimonialism and foreign aid.
It is easy to see why many scholars emphasise Cheibub et al.'s (1996) observation that wealth increases the prospects of political stability, with new democracies (and new autocracies) especially likely to survive when they begin life with a GDP per capita in excess of $6000. Very few African states, either in the 1960s or at the time of the reintroduction of multiparty politics in the 1990s, have surpassed this threshold and a number of governments have found their options seriously constrained by a lack of funds (Cheeseman 2015).
The intuitive appeal of the claim that it is easier to build stable democracies in more ethnically cohesive states is also clear (Horowitz 2000). The continent includes some of the most ethno-linguistically diverse societies in the world, while countries such as Nigeria feature over one hundred self-identified groups. Taken together with evidence that many-though far from all-citizens vote along ethnic lines (Bratton et al. 2012;Dulani et al. 2021), and that ethnic identity has played a significant role in some of Africa's most destructive wars (Abbay 2004;Chakravarty 2014), it makes sense that many scholars have worried that the strength of sub-national identities undermines the prospects for political stability and the provision of public goods.
Institutional design has also received considerable comment. Juan Linz's (1990) famous argument that presidential political systems are less likely to be stable than parliamentary ones remains empirically true, although not for the reasons that he originally identified (Sanches 2018a). Although subsequent revisions of Linz's argument suggest that presidential systems tend to be less stable and democratic because they are often introduced in more problematic contexts and not because they descend into institutional gridlock (Mainwaring and Shugart 1997), many Africanists continue to identify excessive presidentialism as a major contributor to democratic backsliding (Cheeseman 2015). The ability of leaders to remove presidential term-limits, for example, is argued to have facilitated the consolidation of one-man rule in countries such as Rwanda and Uganda (Cheeseman 2010).
Finally, a number of historical studies have followed a recent trend in the comparative literature by emphasising the political significance of colonial legacies (Bernhard et al. 2004;Owolabi 2017;Lee and Paine 2019). More specifically, some crossnational quantitative analyses have revealed that the bifurcated legal-administrative frameworks established for 'native' and 'settler' populations had important consequences for long-term development (Owolabi 2017), while others suggest that British colonial rule may have left a less problematic legacy for democratic politics (Lee and Paine 2019). In the African context, this argument is appealing both because colonial governments paid little attention to creating the foundations for democratic self-government, and because the former colonies of two colonisers have experienced particularly troubled post-colonial existences. While Portuguese rule gave way to civil war in its two most significant African territories, Angola and Mozambique, a number of scholars have traced episodes of ethnic cleansing and genocide in Burundi and Rwanda back to the particularly divisive ruling strategies employed by King Leopold II and the Belgian colonial state.
A second set of literature has acknowledged the potential significance of these factors but instead emphasised dynamics that are more specific to the continent, most notably the prevalence of neo-patrimonialism and the impact of foreign aid. Neo-patrimonialism-a kind of politics that emerges out of the fusion of 'traditional' forms of authority and the modern state-is the predominant theoretical framework through which researchers have sought to understand leadership and the state in Africa (Medard 1982). In particular, Africanists have tended to focus on the way in which informal personal networks have undermined the independence and capacity of key democratic institutions (Cheeseman 2015). The prevalence of neo-patrimonialism has therefore been cited as one of the main reasons for the weakness of both the rule of law and of political accountability, two factors with clear implications for the consolidation of democracy. Indeed, even those scholars that have critiqued this terminology have argued that pre-colonial forms of authority effectively prevented the colonial state from becoming embedded in African societies (Chabal and Daloz 1999), resulting in political systems that had the appearance of Western democracies but in reality were anything but.
Finally, the intimate relationship between many African states, foreign aid donors and international financial institutions-along with the fact that the end of the Cold War triggered a wave of political liberalisation on the continent-has also encouraged a focus on the role of Western states in democracy promotion. In line with Levitsky and Way's (2010) argument that democratisation is most likely to occur when Western states have high linkage (i.e. trade) and high leverage (i.e. aid) with a given country, it is often assumed that leaders are more likely to avoid authoritarian strategies if they govern aid-dependent countries. Peiffer and Englebert (2012), for example, demonstrate that the extent of government's economic 'vulnerability' to foreign donors is positively correlated with both earlier transitions away from authoritarian rule and with the quality of democracy between 1995 and 2011.
Although these factors are common in the literature, it is important to note that their significance remains controversial. In addition to contradictory findings concerning the impact of a particular form of colonial rule, or the significance of external actors, it is clear that a number of African countries, including some of our cases, Benin and Ghana, managed to democratise from an extremely challenging context. Indeed, it has been argued that one of the most distinctive features of African states is their capacity to 'democratize against the odds' (Cheeseman 2015). These caveats notwithstanding, however, it is striking that-as we discuss at greater length when we introduce our cases-none of these factors can fully account for the distinctive performance of small island states.

Democracy, size and insularity
Given the inability of traditional theories to account for the variations in the degree of authoritarianism that we describe in this paper, we turn to a nascent literature on the impact of both smallness and islandness on democratic politics (Anckar 2002(Anckar , 2006(Anckar , 2008. Both factors have been linked to democratic development, and various scholars have highlighted a statistical correlation between (small) size and insularity and democratic performance (Srebrnik 2004).
The academic literature on the political effects of size can be traced back to the Ancient Greek philosophers (Plato and Aristotle) and the thinkers of the Enlightenment (Montesquieu and Rousseau), who all emphasized the political benefits of a limited population size. According to these theorists, smallness increases the involvement and participation of citizens in politics, and the protection of individual freedoms. In their landmark study, Dahl and Tufte (1973) find that small units tend to be more homogenous, and facilitate direct forms of communication between citizens and politicians, possibly enhancing the quality of political representation.
According to Lijphart (1977: 65) a small population size 'enhances a spirit of cooperativeness and accommodation', which makes democratic development more likely. More recent studies have hypothesised that smallness stimulates the political awareness, efficacy, and participation of citizens, creating a more fertile ground for democratic development (Diamond and Tsalik 1999;Sanches 2020a). Empirical studies of both subnational jurisdictions as well as nation-states also reveal that small units indeed have more homogenous populations, as well as higher levels of political participation (Remmer 2010;Gerring and Veenendaal 2020).
The literature on how being an island affects politics does not have the same classical antecedents as the scholarship on the effects of population size. Rather, a community dedicated to the study of nissology (Grydehøj 2017) has emerged in parallel with the formation of newly independent small island states and territories in the post-colonial period. Island studies scholars define 'islands' as geographic spaces that are isolated from metropolitan centres due to distinct obstacles-typically sea, but also mountains, rainforests, or deserts-that render them relatively insular and inaccessible. These arguments are linked to a broader volume of work that highlights the democracy-stimulating effects of insularity, which typically uses the United Kingdom as the prototypical example.
The first political consequence of insularity is that it facilitates community cohesion, especially in the face of the natural vulnerability of islands to both economic shocks and natural disasters (Anckar 2002). Specifically, the creolisation of society can mitigate against ethnic conflict by reshaping ethnic relations into new identities and cultural frameworks. According to Seibert (2012: 30) the process of creolisation-which occurs through the admixture of different cultural heritages and ethnic groups-facilitated the 'development of a new common culture and collective identity' in Cabo Verde and São Tomé and Príncipe. This, we argue, facilitated future democratic developments. The downside of strong community ties is that islands tend to have powerful cultures of compliance. As scholars of both the Caribbean and the Pacific have long observed, these societies tend to be conservative in the sense that they are resistant to change because, regardless of the institutional structure, politics will always be an intimate, face-to-face affair dominated by key community leaders (Sutton 2007).
The natural border between land and sea also means that islands tend to be on the periphery of historical processes that affect other states (Congdon Fors 2014). Many islands never industrialised and the very smallest often decolonised later than larger mainland states due to a perception that independence was economically unviable. In relation to democratisation, the positive effect is that small islands do not share land borders and thus are less likely to be destabilised by wars or revolutions in neighbouring states. This benefit appears to be particularly significant in the case of sub-Saharan Africa given the porous and often arbitrary borders inherited from colonial rule (Herbst 2014). Englebert et al. (2002) find that where colonial borders partitioned pre-existing political groupings-a phenomenon they call 'dismemberment'-this increased the proportion of international disputes. By contrast, where borders brought together different precolonial political cultures-which they term 'suffocation'-there has been a greater likelihood of 'civil wars, political instability, and secession attempts' (Englebert et al. 2002(Englebert et al. : 1093. Island states typically were not subdivided, and so did not experience either of these processes. They have also been more insulated from problematic international 'demonstration effects'. While the idea that democratisation can be facilitated by imitation is often advanced to explain the relative strength of liberal institutions in Europe (Teorell 2010), this logic can be reversed in Africa; authoritarian neighbours make consolidating democracy more difficult. As a result, the insulation of sea borders can help explain diversion from continental trends (Veenendaal 2020).
Empirical analyses reveal that both smallness and insularity are statistically correlated with democracy, and some scholars have argued that the higher incidence of democracy in small states is in fact a product of their island status (Anckar 2008;Veenendaal 2014). We cannot resolve this debate here: the hypothesised effects are not easy to disentangle, and our five cases share both characteristics. Instead, we argue that it is the mutually reinforcing interaction between islandness and smallness that can help us explain the absence of the worst forms of authoritarian excess. Specifically, we argue that smallness leads to more authentic political representation and effective service delivery, while islandness promotes community cohesion and provides a buffer against instability and conflict in neighbouring states. This results in a positive feedback loop that guards against authoritarian excess.

Africa's small island states
We examine our claims by means of a small-N comparative study that includes Cabo Verde, the Comoros, Mauritius, the Seychelles and São Tomé and Príncipe. There are remarkably few similarities between these cases aside from their smallness and islandness, and so none of the factors that are commonly hypothesised to promote democratic governance can explain the distinctive performance of this group (Table 1). There is no common colonial heritage (Lee and Paine 2019), for example, that would explain the political fortunes of these states. Cabo Verde and São Tomé and Príncipe were Portuguese colonies and the Comoros a French colony. The Seychelles and Mauritius were French territories first, and later British.
The same is true of wealth. None of our five cases met Cheibub et al.'s (1996) threshold ($ 6000) at the point of independence. Additionally, there is little evidence that economic success has driven democratisation thereafter. Indeed, while Mauritius and the Seychelles have become relatively wealthy, this is not yet true of Cabo Verde, the Comoros and São Tomé and Príncipe. Moreover, the GDP per capita of our wealthiest case (the Seychelles) is more than ten times that of the poorest (the Comoros). Partly as a result, there is also considerable variation in the degree of aid dependency among our states.
There is also little convergence when it comes to institutional design. In part due to the different colonial heritage of the five states, their political systems are considerably different. Mauritius operates a parliamentary system that reflects the influence of British colonisation, the Seychelles' and the Comoros' presidential constitutions are perhaps closest to the French model under 'cohabitation', and finally Cabo Verde and São Tomé and Príncipe have semi-presidential systems following the Portuguese model (Sanches 2020b). Again, there is no evidence that the design of the political system represents a common variable that could explain the lower levels of authoritarianism in these states. The impact of neo-patrimonialism is harder to measure as it requires a qualitative assessment of complex political landscapes. In all five cases we find that politics is extremely personalised and that governments rely heavily on patronage and clientelism to retain control. However, we also find considerable variation in the extent to which this generates the kinds of damaging corruption that is often said to undermine progress towards democracy. While Cabo Verde, Mauritius, and the Seychelles feature low levels of corruption, São Tomé and Príncipe performs less well, and the Comoros stand among some of the world's most corrupt states (Table 1). While corruption is not a perfect proxy for neo-patrimonial politics, this suggests that there is significant variation within our cases. Moreover, we argue below that in many ways the better performance of some of these states on issues such as corruption is a product of small island status, rather than an independent driver of accountable and responsive government.
The story is similar when it comes to ethnic diversity. Cabo Verde, São Tomé and Príncipe and the Seychelles feature mixed-race communities and a dominant language and culture (Seibert 2012;Campling et al. 2013). The Comoros' inhabitants are a blend of various peoples from the Indian Ocean shoreline. In contrast to the other cases, the predominantly Muslim population already inhabited this archipelago before the onset of colonisation. Moreover, Mauritius 'is one of the most ethnically heterogeneous states in the world' for its size (Srebrnik 2002: 278). The main ethnic groups include Indo-Mauritians (dominant), Creoles, Sino-Mauritians, and Franco-Mauritians. These variations suggest that ethnic homogeneity, on its own, cannot be an explanation for the democratic performance of these five countries. Furthermore, we argue that-as with lower levels of corruption-the fact that many of these countries ultimately built more cohesive societies was not simply because they started with more homogenous population but was in part a product of the evolution of a more inclusive and stable form of politics. This, and other features of small island life described below, encouraged the creolisation of society in some cases, and the accommodation of ethnic differences in others. In the Comoros, for example, the distinction between some groups have been blurred over time due to intergroup marriage (Ojo 2016: 258-259).
By a process of elimination, we therefore arrive at the alternative explanation for why these states became democratic introduced above: their small size and insular geographies. We conduct an historical comparative analysis to identify similarities and differences between countries based on a combination of interviews held primarily with politicians in some of our cases, 2 qualitative sources and secondary 2 Seventeen interviews were conducted in Cabo Verde (in 2017), fourteen interviews were conducted in the Seychelles (in 2011), and participant observation and informal interviews were conducted in São Tomé and Príncipe (in 2019). The interview respondents primarily consisted of (former) politicians, as well as journalists, academics, and representatives of non-governmental organisations (NGOs). The participant observation in São Tomé happened during a two-week event that aimed at promoting citizenship and political participation in the country. The interviews were conducted as part of different research projects, and unfortunately we have not yet been able to conduct interviews in the Comoros and Mauritius. However, this deficiency is partially remedied by the existence of a more extensive secondary literature on these (larger) cases, from which we have drawn insights. literature. We knew that commonalities were significant if they diverged from established studies of other African states, both large and small, continental and island, but resonated with studies of democracy in Caribbean and Pacific small islands states (Corbett and Veenendaal 2018). Thus, we provide a form of comparative analysis that is both sensitive to contextual nuances but at the same time draws out and puzzles with affinities between otherwise disparate cases (Boswell et al. 2019).

The political life of African small island states
It is important to consider the democratic development of our five island states in the longue durée for two reasons. First, the different features of the five countries mean that while they have all tended to avoid the worst excesses of authoritarianism, their political systems have followed different pathways. Second, the historical evolution of political practices has itself played an important role in shaping contemporary dynamics and the prospects for a more open political landscape. This section traces the development of political institutions and norms of behaviour from the colonial era to the present day. In particular, we demonstrate that the combination of a less extractive colonial experience and relatively inclusive form of authoritarian rule post-independence created a more hospitable environment for the introduction of multiparty politics in the early 1990s. As we will show, both developments can be linked to these countries' small size and archipelagic geography.

The colonial legacy
The relative ease of extending control over small island states, along with their strategically important location and 'more benign disease environment' led to a form of colonial rule that was both more pervasive and longer lasting. While most mainland African territories were only fully colonised towards the end of the 19th century, Cabo Verde and São Tomé and Príncipe had already become part of the Portuguese Empire in the 15th century, whereas Mauritius and the Seychelles were colonised in the late 18th century and the Comoros were annexed by France in 1843. Conversely, by the time that the Comoros, Cabo Verde, São Tomé and Príncipe, Mauritius and the Seychelles gained independence, most countries across the African continent had already acquired that status. Yet, in contrast to the 'extractive, inegalitarian institutional heritage' (Bertram 2007: 245) of mainland countries, colonial governments in small islands relied less heavily on divide-and-rule policies and when independence came it rarely involved violent struggle.
The longer period of colonial rule also had profound implications for the societal development of our cases. While the Comoros had their own indigenous cultures before the imposition of European rule, the other island territories experienced processes of creolisation similar to the Caribbean. Being previously uninhabited, the contemporary populations of the Seychelles and Mauritius are all descendants of immigrants who were brought to the islands-often against their will-at some point during colonial rule. Although the populations of Cabo Verde and São Tomé and Príncipe retain some indigenous groups, the majority of these countries' citizenry now consists of the descendants of immigrants (Seibert 2012). In Cabo Verde, the Seychelles and São Tomé and Príncipe the creolisation process led to a blend of different heritages and to the development of shared identities while in Mauritius, the distinct populations with Indian, French, British and Creole heritage resisted incorporation into a common identity. Significantly, however, the comparative ease of governing a small territory meant that even in this case, the divide and rule strategies of colonial authorities were not as deleterious to forming a coherent national political community as they were in countries such as Nigeria and Rwanda. Overall, the combination of tiny populations and islandness facilitated the creolisation of society, or at least (with the exception of the Comoros) the kind of consensual politics that is conducive to democracy.
This did not mean that colonial rule facilitated democratic government in any straightforward way. In the Comoros, the French colonisers only permitted the formation of legal political parties in the archipelago after the 1968 student strike (Banks et al. 2007: 205). This concession resulted in the emergence of various proindependence political parties that gained representation in the Comorian Chamber of Deputies in 1972. Moreover, Cabo Verde and São Tomé and Príncipe were ruled under particularly restrictive conditions, with few freedoms granted by the Portuguese regime, itself an authoritarian state under the Salazar regime. It was therefore only in the Seychelles and Mauritius where aspects of self-government emerged relatively early. In the Seychelles, the first legislative elections occurred in 1948 under restricted suffrage rules, and the first elections under universal suffrage in the 1970s resulted in the rise of two newly formed parties: the Marxist Seychelles People's Progressive Front (SPPF) 3 led by France-Albert René and the liberal-conservative Seychelles Democratic Party (SDP) led by James Mancham. Similarly, in Mauritius a series of protests led by agricultural and labour movements and the Mauritian Labour Party in the 1930s pressured the British colonial government 'to expand subordinate group access to the government' (Lange 2010: 2). But even in Mauritius and the Seychelles, the legacy of colonial rule included first, slavery, and later, indentured labour (Vaughan 2005).
The key feature of colonial rule in our cases was therefore not that it laid down a blueprint for democratic government, but rather that it avoided some of the most problematic aspects of colonialism on the mainland. Because the island states were so small, it was comparatively easy for colonial regimes to establish and retain control. Along with the absence of large European settler communities, this reduced the incentive for authoritarian abuses, and meant that there was less extensive development of the kind of coercive legal apparatus and security forces that evolved during this time in countries such as Kenya. In addition, although the end of colonial rule came relatively late, the five island territories experienced comparatively peaceful and negotiated transitions to independence, which required the holding of elections (Cabo Verde, São Tomé and Príncipe, Mauritius) or referendums (the Comoros) and the establishment of compromise between political parties (the Seychelles). This was a very different experience from those countries that gained independence through armed conflict such as Angola, Mozambique and Zimbabwe. Partly as a result, most nationalist elites were not rebel leaders before they became civilian presidents. The one partial exception to this is Cabo Verde, where many leaders were involved in the liberation struggle in Guinea Bissau, but this conflict never reached the islands (Chabal 2002). The absence of conflict is significant, as research on other African states finds that sustained political violence undermines the prospects for democratisation (Cheeseman 2015).
Thus, while colonial rule did not necessarily leave a strong foundation for democratic governance, it was not as destructive to that goal as in other parts of the continent. Consequently, the small island states entered independence with comparatively cohesive societies and limited experiences of violent struggle. However, the comparison of colonial legacies also points to an interesting discrepancy between the Comoros and the other four cases. As mentioned above, the predominantly Muslim population of the Comoros consists of various ethnic groups whose presence predates the onset of colonialism (Walker 2007). The absence of creolisation and a shorter period of colonial rule means that the Comoros are closer to the experience of mainland African states than our other cases, which could explain their greater democratic difficulties.

The 1970s and 1980s: struggle for political stability
The shallow roots of the democratic elements introduced in the pre-independence period are demonstrated by the fact that in all of the countries bar Mauritius the government either fell to a coup or moved to extend control and sideline the opposition-and even Mauritius suffered a state of emergency. The post-colonial administrations were also dominated by a small number of individual leaders, leading to heavily personalised political systems that were at times chaotic. However, when political violence occurred it tended to be directed between elites and not at ordinary citizens. Moreover, when coups and political infighting gave way to the formation of one-party states, these proved to be relatively stable and inclusive, and so served as a more promising springboard for the process of democratisation that occurred in the late 1980s. In addition to the less problematic colonial inheritance, the emergence of less repressive authoritarian governments in the 1970s owed much to the ability of leaders to co-opt support, and the greater traction of clients, in small societies.
Cabo Verde had perhaps the least eventful post-colonial era. After independence in 1975, the government introduced a one-party state led first by the African Party for the Independence of Guinea and Cabo Verde (PAIGC) and then from 1980 onwards by the African Party for the Independence of Cabo Verde (PAICV). A similar process occurred in São Tomé and Príncipe, where Manuel Pinto de Costa ruled a oneparty state from independence in 1975 all the way through to the reintroduction of multiparty politics in 1990. Although Da Costa's Prime Minister, Miguel Trovoada, was arrested in 1979 and accused of a complicity in a coup attempt, Trovoada was allowed to go into exile rather than being executed (Seibert 1999;Sanches 2020b). The Seychelles followed a similar path, after a greater period of initial turbulence when a Marxist coup within the first year after independence replaced President James Mancham with his Prime Minister, France-Albert René. The country became a one-party state in the following year, however, and despite an army mutiny and an attempt by South African mercenaries to return Mancham to power it remained so until 1991, when René agreed to a process of political liberalisation.
The situation was rather more complicated in the Comoros, first because of tensions between the four islands that make up the country, and second because of the role of European mercenaries. In 1974, three of the islands voted for independence, but a fourth, Mayotte, voted to retain links with France. One year after the attainment of independence of the three islands, President Ahmed Abdallah was deposed by Ali Soilih who sought to introduce a socialist one-party state. This plan was upset by the intervention of European mercenaries in 1978, who removed Soilih from power and restored Abdallah to the presidency. In his second stint in office, Abdallah overcame a series of challenges to lead his single-party system for a decade, only to be assassinated by a different group of European mercenaries in 1989. France subsequently intervened to remove the mercenaries and set out a timetable for democratisation (Banks et al. 2007;Baker 2009).
Mauritius took a different path still, managing to maintain high quality democracy and competitive elections for almost all of its post-colonial existence (Miles 1999). This seemed unlikely in 1971, when a combination of economic difficulties and a series of by-election victories for the opposition Mauritian Militant Movement (MMM) led the government to cancel the first independence elections. However, instead of changing the constitution to introduce a one-party state, the ruling party eventually lifted these restrictions and held uninterrupted elections from 1982 onwards. This impressive political stability owed much to the combination of the country's parliamentary political system and party system, which effectively required parties to form power-sharing coalitions in order to secure a majority. Along with the Fabian socialist beliefs of early Prime Ministers, this encouraged a politics of compromise and mutual toleration.
With the exception of Mauritius, then, the 1970s and 1980s did little to create or consolidate democratic institutions. Indeed, even in Mauritius political life continued to be dominated by a small number of figures. Since independence in 1968, Mauritian politics has been a family affair, with only five different individuals holding the post of Prime Minister-and two of these (Navin Ramgoolam and Pravind Jugnauth) are the sons of former PMs (Sir Seewoosagur Ramgoolam and Sir Anerood Jugnauth, respectively) (Kasenally 2011). This reflects a broader pattern across our five countries, namely that the linkage between citizens and politicians has tended to be clientelistic.
Despite these limitations, however, the forms of government that emerged in the post-independence era were often less abusive or oppressive than those on the mainland. This was in part due to the fact that governance was less challenging in smaller and more cohesive states, but it was also related to the way in which patron-client relationships play out in small island societies. In the larger states of the mainland, clientelism often led to the emergence of distant and indirect relationships brokered by multiple tiers of intermediaries (Barkan 1984). The situation was very different in our five countries, generating a set of more intimate relationships. To start with, direct contact and face-to-face communication between citizens and politicians in small societies facilitate enhanced citizen engagement and a sense of being closer to government. Indeed, kinship and family relations have a greater effect on clientelistic exchanges in small states-tellingly, the acronym of São Tomé and Príncipe (STP) is often jokingly interpreted as 'somos todos primos' (we are all cousins). In turn, this 'closeness' helps to maintain political stability and places greater constrains on political leaders.
These more direct clientelistic ties have important consequences for the delivery of public services. In contrast to larger countries, where state resources are distributed to citizens by means of complex, hierarchical networks that involve many intermediaries or 'brokers', in small societies citizens have direct access to politicians, and can directly express their grievances, preferences, or demands (Veenendaal 2019;Veenendaal and Corbett 2020). This not only gives clients greater power vis-à-vis patrons than the clientelism literature tends to envisage, but also means that politicians can address the concerns and needs of their constituents much more directly. Since electoral districts are much smaller than in larger countries, a few votes might make the difference between winning and losing, meaning that individual citizens can exert greater pressure on politicians to bestow them with certain goods or services (Veenendaal 2019). These electoral realities often coincide with social pressures, as clients and patrons are more likely to be related or otherwise connected, so that ignoring constituents' wishes can generate social sanctions for politicians. For example, in São Tomé and Príncipe, '[t]he personal bonds between rulers and their opponents, which inevitably existed in such a small society, prevented the regime from becoming as violent as other similar regimes' (Seibert 1999: 150). Even under single-party rule, therefore, politicians in our five cases could not easily ignore the demands of their supporters.
This more inclusive governance style is important, because civilian one-party states represented a stronger foundation for the evolution of democratic multiparty political systems (Bratton and van de Walle 1997). By both establishing norms of political participation and institutionalising elections, single party systems helped to strengthen democratic practices and values (Cheeseman 2015). It was therefore significant that the civil authoritarian regimes that emerged in Cabo Verde, São Tomé and Principe and the Seychelles all held regular plebiscitary elections in which citizens were asked to either approve or reject the single party lists. Turnout rates in these elections were generally high, even though electoral competition was restricted by the government (Clemente-Kersten 1999;Thibaut 1999a, b). This was in part because ruling parties were keen to encourage mass participation to demonstrate their own legitimacy. Prior to the elections in Cabo Verde, for instance, rallies were held throughout the country to discuss the PAICV candidates, while in São Tomé and Principe, the MLSTP candidates were first elected by district assemblies (Sanches 2018b(Sanches , 2020bSeibert 1999). Other forms of mobilisation included the public discussion of key public policies, as occurred during the implementation of the agrarian reform in Cabo Verde in the mid-1980s (Sanches 2018b: 76).
As with the colonial period, then, the 1970s and 1980s did not so much institutionalise democracy as avoid the worst aspects of authoritarianism. With the exception of Mauritius, none of these political systems enabled citizens to select their governments in any meaningful way. Yet, while many mainland states suffered divisive episodes of military or one-man rule, the small island states tended-after bouts of inter-elite struggle-to establish relatively stable and participatory forms of politics. In this way, the politics of insularity facilitated the transition to multiparty politics in the 1990s. Indeed, it is striking that in Cabo Verde and São Tomé and Principe constitutional amendments introduced in 1985 and 1987, respectively, allowed independents and members of associations authorised by the state to run for elections (Branco and Varela 1998;Seibert 1999;Sanches 2020b). As such, the process of democratisation could be said to have begun considerably earlier than on the mainland.

The 1990s and 2000s: (re)introduction of multi-party politics
Three trends combined to force the reintroduction of elections across Africa in the late 1980s (Cheeseman 2015). First, economic stagnation undermined the popularity of governments and rendered them increasingly in need of international financial assistance. Second, citizens and marginalised political elites became evermore frustrated with the absence of economic and political choice, especially when the nationalist struggle began to fade in the popular memory. Third, the end of the Cold War enabled Western powers to focus on human rights and democracy as conditions for their support, as opposed to loyalty. Between 1989 and 1995, these three pressures resulted in the reintroduction of multiparty elections in the vast majority of African states.
What happened during and after the introduction of multiparty politics varied considerably, however. As set out in the Introduction, following the transition many of the countries on the mainland-though by no means all-became 'competitiveauthoritarian' regimes in which elections were held without the other trappings of democracy (Levitsky and Way 2010). The experience of the small island states was quite different. All now hold regular elections that are relatively open and competitive. Indeed, it is striking that, following the first victory of the Seychellois opposition in the 2016 parliamentary elections and the 2020 general elections, all five cases have experienced alternation in office through the ballot box. This stands in strong contrast to the mainland, where a majority of countries have yet to change the party of government through an election. These transfers of power were not isolated examples, but part of a wider trend: most notably, in the 1990s Cabo Verde and São Tomé and Principe quickly joined Mauritius in the very select category of African liberal democracies. Yet, as in the two previous eras, the comparative lack of authoritarian tendencies in these countries has not been driven by the strength of democratic institutions. Instead, the politics of insularity has continued to be characterised, as in other parts of the world, by a strong discrepancy between formal institutional structures and an influential set of informal political dynamics, which can chiefly be attributed to these countries' small size and insular geographies (Corbett and Veenendaal 2018).
The prevalence of elite and personalised politics was evident during the process of transition from authoritarian rule itself (Cheeseman et al. 2018). Cabo Verde, for example, underwent a pacted transition in which the main negotiators embraced the new democratic agenda. The reform process was initiated in the late 1980s by the PAICV, and later spurred by the emergence of the Movimento para a Democracia (MpD) in the early in 1990s, with talks between the two parties leading to the emergence of a new, stable, political dispensation (Sanches 2020b). As a Cabo Verdean politician explained the process of reform was constantly ongoing: '[i]t was not like one day we remembered that we should open up politically' (personal interview by author, 1 February, 2017).
The situation was somewhat similar in São Tomé and Príncipe, where the regime initiated the move towards democracy. Pressured by internal critics, President Da Costa held a national conference in 1989 to set the liberalisation agenda (Seibert 1999;Sanches 2020b). Before any other country in Africa, Da Costa put the issue of reintroducing multiparty politics to a referendum, where it secured 95% support. The Seychelles also experienced a 'transition from above', as the SPPF controlled the process through various phases, despite the fact that the opposition Seychelles Democratic Party (SDP) was allowed to take part in the commission that drafted the constitution (Thibaut 1999b). Meanwhile, in the Comoros, elite negotiations took place in Paris under the auspices of the French government, leading to a National Conference of 24 political parties and a new constitution that was approved by 74% of voters in a referendum in 1992 (Ali et al. 2011).
In all of the countries that underwent a democratic transition in the early 1990s, then, the main negotiations took place within a small political elite and the role of the wider public was limiting to approving the deal that emerged from those talks. As Rachel Rield (2014) has argued, when this happens, far-reaching political reform and liberalisation are less likely to occur. Instead, the quality of democracy will depend on the ruling elite, their commitment to pluralistic norms and values, and the informal constraints that they face. On the mainland, 'transitions from above' in countries such as Tanzania typically resulted in the maintenance of tight authoritarian control (Cheeseman 2015). That this has not happened in the five island states speaks to the more promising context for democratisation and the greater constraints placed on elites. Indeed, it is the particular incentives facing small island leaders, and the relationship between them, that best explains the variation between our cases over the last 30 years.
In São Tomé and Príncipe, for example, '[t]he reluctance to attack and prosecute political opponents reflects the nature of elite formation […] which centers on shifting alliances between the key political leaders and their factions. Periodic seemingly bitter quarrels are resolved with new deals being made' (Frynas et al. 2003: 17). Personalistic power struggles are also a key component of Cabo Verdean politics, but have proved manageable because '[a]mong the elite, democracy is not simply a device to attract donors, but appears to be a deep-rooted commitment' (Baker 2006: 509). In interviews, politicians lionise the willingness of their parties to accept defeat, lauding both national democracy and inclusive mechanisms to select party leaders.
Not all of our countries have had such a smooth transition, however. Inter-elite relations in the Comoros and the Seychelles have been significantly less harmonious. In the Seychelles, the durability of the SPPF-renamed the Parti Lepep (People's Party) in 2009 and re-branded as the United Seychelles Party in 2018has stymied the pace of political change. More specifically, the fusion of the SPPF and the state has enabled it to dominate the bureaucracy, civil society and the media (Baker 2008). Relations between the SPPF leadership and the opposition have been hostile, and opposition leaders and their supporters have long been marginalised and victimised by the ruling party. The government has also sought to mobilise support through clientelistic strategies (Veenendaal 2014), which are more feasible in a country with less than 100,000 citizens. However, despite these challenges, politics remains considerably more open than in most mainland states, which is demonstrated by the fact that the opposition was able to win the 2016 general elections-a watershed moment in the country's history that led to the immediate resignation of President Michel. In the 2020 presidential election, Michel's successor Danny Faure was convincingly defeated by the longstanding opposition leader Wavel Ramkalawan, meaning that the Seychelles have now also experienced alternation in office at the executive level. These developments will almost certainly change Freedom House's ranking of the Seychelles from 'partly free' to 'free' in the near future.
Inter-elite relations have also been strained in Comorian politics, which continues to be something of an outlier among our cases due to the degree of inter-island hostilities and secessionist aspirations, and the instability generated by the fact that one island remains an overseas department of France (Walker 2007;Mohamed 2001;Hassan 2009;Ali et al. 2011). The attempted secessions of the two smaller islands of Nzwani (Anjouan) and Mwali (Mohéli) from the largest island of Ngazidja (Grande Comore) have fundamentally destabilised Comorian politics, and pose an enduring obstacle to economic and democratic development in the archipelago (ibid.). While the implementation of a federal arrangement and a rotating presidency attempted to generate more stability, these changes failed 'to address the almost intractable problem of both meeting the need for island cooperation to ensure viability, and the need to recognise island difference and freedom from domination by any other island' (Baker 2009: 217). The weaker national identity generated by rival islands means that '[o]nly at its boundaries, its interface with the outside world, is the Comorian state real; from within all dissolves as the form reveals itself to have no content' (Walker 2007: 600). Moreover, while the Comorian political elite now consists of politicians from all three islands, the antagonistic relations between the islands entail that the elite remains fragmented. Yet, although this has produced endemic instability, it has also limited politicians' opportunities to concentrate powers in their own hands, meaning that in contrast to many mainland African states, no single politician has been able to entirely dominate Comorian politics. As a result, even the Comoros have avoided the worst excesses of authoritarian rule.
As we have already noted, the intimate nature of politics in small island states has some positive consequences in terms of citizens feeling connected to the system and limiting the abuse of power. This is not a straightforward boon for democracy, however, as the clientelistic relationships that have emerged have often weakened key checks and balances institutions. In São Tomé and Príncipe, '[d]emocratic institutions have merged with the political attitudes and clientelist models of resource distribution which have characterized all previous regimes. […] Party competition has only resulted in new opportunities for creating and exploiting patron-client relationships' (Seibert 1999: 322). Similarly, in the Seychelles ' [t]here are few in the country who are not convinced that the patronage system continues despite the transition to democracy' (Baker 2008: 289).
In sum, therefore, patron-client linkages in the five African small island states appear to have simultaneously led to the maintenance of formal democratic institutions while constraining the quality of democracy itself. Even in Mauritius, whose exceptional democratic experience has drawn the greatest attention from comparative politics scholars, politics exhibits a 'dynastic, ailing, and closed culture' (Kasenally 2011: 43). In this sense, the personalisation of politics has been both a blessing and a curse. The greater constraints placed on political elites in small societies have encouraged leaders to limit their abuse of power while simultaneously sustaining their popularity through personalistic and clientelistic appeals that constrain the evolution of more transparent and accountable government.

Reconsidering the interaction between size and insularity
Our explanation for how the interaction between small population size and islandness reduces the risks of authoritarianism can also be observed if we take away either one of these factors. Consider Madagascar, for example, which is an island but is anything but small. The continent's most populous island is currently rated as 'partly free' following a transition back to electoral politics in 2013, but the quality of democracy has regularly fallen well below this level during a number of authoritarian episodes, most notably a coup in 2009. The large size of the country's population-and of the island itself-has played an important role in this process, because it has restricted creolisation and facilitated social divisions. Although these are complex, the most notable cleavage falls between the Merina-roughly the two main groups that live on the highlands and plateau area-and the côtiers, who reside in the coastal areas (Marcus 2016).
The tensions between these two groups is politically significant, not least because a common complaint of the côtiers is that the central government does not respond to their needs. As Rosabelle Boswell (2008: 64) has argued, the distinct and segregated location of these groups has encouraged 'a hegemonic discourse of "rootedness"'. In turn, the politics of belonging has led 'certain groups to forge links with actual and fictive "homelands"', undermining the evolution of a more cohesive national identity. These divisions cannot be neatly mapped onto all of the political crises that have engulfed Madagascar over the last twenty years, because many have been driven by intra-family struggles within the Merina elite (Jütersonke et al. 2010: 13). The popular narrative of a conflict between Merina and côtiers has, however, undermined social cohesion and contributed to a divisive understanding about the source of the country's problems. The large size of the population has also proven to be problematic in a second way, namely that political elites are far removed from ordinary citizens. Rather than the close connection between political leaders and their societies that we describe in small island states, Malagasy politics is said to play out 'amongst a small number of around 20 elite families who […] show little interest in the welfare of the broader population' (Jütersonke et al. 2010: 13). In turn, this disconnect has contributed to a series of protracted intra-elite disputes that have led to chronic political instability (Marcus 2016).
The case of Madagascar might be taken to suggest that size is what really matters, and that insularity is of secondary import. But while there are a number of cases on the mainland that could be read as supporting this interpretation, it turns out to be deeply misleading when the whole universe of cases is considered. It is true that the vast majority of countries that are typically classified as full democracies in mainland Africa have small populations. Botswana and Namibia have less than 3 million inhabitants, while Benin and Senegal have less than 16 million, comparable with mid-sized American states. By contrast, of the continent's ten most populous countries only one-South Africa-has achieved democratic status. This speaks to the difficulty of managing large populations across vast territories, especially when these challenges are compounded by high levels of diversity and a limited state infrastructure. But there is powerful evidence that size alone does not insulate a country against authoritarianism. While less populous African countries are on average more democratic than more populous ones, the ten smallest countries on the mainland include a number of states that have a long history of abusive and exclusionary governments. This includes five countries-Djibouti, Equatorial Guinea, eSwatini, Gabon and Mauritania-that are currently rated as 'not free' by the Freedom House.
The key point then is that neither size nor insularity on their own can fully explain our cases-and there are good reasons to think that this pattern may be relevant beyond the African context. Small islands in the Caribbean and the Pacific are among the most stubbornly persistent post-colonial democracies in the world (Corbett and Veenendaal 2018). Europe is the other region with a high proportion of small states. Here the evidence is more mixed but again the small islands all rank as highly democratic (Malta, Cyprus and Iceland) while some of the small continental states (e.g Monaco and Liechtenstein) are among the oldest remaining monarchies in the world. In any case, our claim is not that being a small island is a necessary or sufficient pre-condition for democracy. Rather, these patterns, both in African and elsewhere, indicate that the interaction between these two factors appears to facilitate or contribute to democratisation in hitherto underappreciated ways. They do not determine the fate of democracy but they can help us better understand similarities between otherwise very different countries.

Conclusion
The five small island states of Cabo Verde, the Comoros, Mauritius, São Tomé and Príncipe and the Seychelles have consistently recorded comparatively low levels of authoritarianism compared to mainland Africa, and three of them are fully fledged democracies. Neither the classic accounts in the democratisation literature, nor the arguments that have been developed specifically to describe and explain African realities can fully explain their greater insulation from authoritarian rule. Instead, we have argued that their distinctive pathway is best viewed as a product of their being small islands. Specifically, smallness brings clients closer to patrons and eases the task of government, while islandness promotes community cohesion and provides a buffer from instability and conflict in neighbouring states. All these factors are mutually reinforcing. Smaller populations and geography have enabled governments to more effectively engage with citizens, and increased the pressure on leaders to maintain inclusivity. They have also, by facilitating creolised societies with a greater degree of homogeneity, led to more cohesive societies that are better placed to demand public goods. Meanwhile, being offshore has insulated these states from some of the potential negative political and economic consequences of having 'bad neighbours'. Taken together, these developments militated against the use of divisive winner-takes-all politics, boosted political stability, and encouraged more responsive and accountable government.
This does not mean, however, that all five states will inevitably become and maintain high quality democracies. In most cases, the political systems that have evolved are not strongly institutionalised, which makes them vulnerable to external shocks. Moreover, we have documented the emergence of intensely personalised political systems in which leaders have at times enjoyed considerable leeway to operate as they see fit. Presidents and Prime Ministers have faced strong incentives to avoid authoritarian excesses and these structural conditions remain in place, but good and forward-thinking leadership has also played an important role in mediating the impact of political and economic crises. This suggests that some of our states-most notably the Comoros-are still at risk of democratic rollback, especially if individuals are elected who do not buy in to the norms of inclusive government that have gradually evolved over the past sixty years. As we have discussed, the absence of creolisation and the shorter period of colonial rule make the Comoros more similar to mainland African states. In present times, persistent inter-island tensions continue to pose a formidable obstacle to democratic development and political stability, explaining the deviant trajectory of this case. Nonetheless, our cases appear to have much better prospects for democratic consolidation than the average mainland state.
The significance of these findings extends beyond our five cases. Even following the advent of new institutionalism in the early 1990s, there is still a tendency to focus on the strength of formal political institutions in explaining the emergence of high quality democracy. Against this, our five cases demonstrate that more democratic forms of government can grow in the context of relatively weak formal checks and balances so long as supportive contextual factors and constraining informal institutions are in place. The comparison between these states, those on the mainland, and larger islands, also highlights the importance of the interaction between being an island and having a small population-neither of these factors is nearly as powerful in isolation.
The significance of the interaction between size and insularity for the prospects for democratic transition and consolidation is an important finding, but not one that easily translates into policy prescriptions. The emphasis we place on creolisation points to the importance of nation-building, and to inclusive nation-building strategies. If being a small island is conducive to democratisation we might also reasonably expect greater progress from countries that feature this kind of society, and alter policy interventions accordingly. But at the same time we recognise that neither reducing the size of states nor creating more ethnically homogenous units will necessarily improve the prospects for democratic consolidation on their own. Moreover, it is impossible to translate over a hundred years of political and social development that has been heavily shaped by insularity into remedies that can easily be applied to mainland countries with very different histories. In this sense, we might conclude that whether or not a country enjoys the benefits of proximity and insularity is more an issue of luck than design.