Evaluating special representation of non-resident citizens: Eligibility, constituency and proportionality

So far, 21 countries have introduced—and some thereaf - ter withdrawn—reserved legislative seats for their citizens living abroad. Existing literature on this form of special representation has studied this topic either empirically or normatively. We bring these two approaches together based on the main dimensions of institutional variation of special representation: (1) eligibility, (2) constituency struc - ture and (3) electoral proportionality. We first discuss each dimension from a normative perspective. In the second step, we map the range of empirical variation and highlight the most common arrangements. We conclude that the normative justification for special representation is gener - ally weak, but some institutional configurations pose fewer problems. Specifically, we see fewer issues with special representation when electoral inclusion is limited to the first generation of emigrants and when it is used to limit the electoral influence of non-resident populations that make up a large share of the overall electorate. By ground - ing our

based on the main dimensions of institutional variation of special representation: (1) eligibility, (2) constituency structure and (3) electoral proportionality.We first discuss each dimension from a normative perspective.In the second step, we map the range of empirical variation and highlight the most common arrangements.We conclude that the normative justification for special representation is generally weak, but some institutional configurations pose fewer problems.Specifically, we see fewer issues with special representation when electoral inclusion is limited to the first generation of emigrants and when it is used to limit the electoral influence of non-resident populations that make up a large share of the overall electorate.By grounding our normative discussion on an empirical mapping, we bridge two disconnected literatures on special representation of non-resident citizens.

INTRODUC TI ON
In 1976, Portugal was the first country to reserve parliamentary seats for representatives of non-resident citizens.Since then, another 20 sovereign states have adopted provisions for special representation of external voters.Some of these countries have abandoned this institution after implementing it (e.g.Morocco and the Cook Islands) or withdrew it from their legal statutes without ever applying it (e.g.Angola and Mali).Currently, 16 countries provide non-resident citizens with special representation at the national level and several others are considering its introduction.In Chile, for example, this form of special representation was proposed at the Constituent Assembly (2020-2022) yet failed to be incorporated in the constitutional draft that was rejected in the 2022 plebiscite.
While external voting rights have been adopted by ever more countries worldwide, reserving legislative seats for non-resident citizens remains a comparatively rare policy compared to the dominant mode of assimilated representation, that is, merging votes cast by non-residents with those of residents.i We ask whether special representation can be normatively supported by arguments from democratic theories.In contrast with past wholesale evaluations of special representation of non-resident citizens, we make use of a global data set to examine its normatively relevant dimensions.And in contrast with purely empirical analyses, we provide normative arguments that can potentially guide decisions on special representation and its design.Although our combination of normative theory with comprehensive empirical analysis is unusual, we believe that it helps break up academic silos in a field where both normative and comparative political scientists have made important contributions.
We begin by considering general arguments raised in normative evaluations of special representation of nonresident citizens.We then capture differences in its institutional design at the national level, drawing on a subset of the largest data set on migrants' electoral rights that allows to map the development of special representation over time (Umpierrez de Reguero et al., forthcoming).We pursue this analysis of empirical variation combined with an in-depth normative discussion focusing on three dimensions of variation.First, we discuss eligibility conditions determining who is eligible to vote and stand as a candidate in special representation designs.Second, we address variation in terms of constituency arrangements, meaning the design of extraterritorial districts.Third, we focus on electoral proportionality by comparing the share of seats for extraterritorial voters in parliament with their share in the total electorate to measure the extent to which the non-resident citizenry is over-represented or underrepresented in each case.
Our analysis builds on previous work on special representation of non-resident citizens.So far, empirical research has explored the institutional design and enactment of special representation largely by using single case studies ii or small-N comparisons (e.g.Laguerre, 2013;Østergaard-Nielsen & Ciornei, 2019;Palop-García, 2018).In several studies on external voting rights, special representation has been used as an independent or control variable (e.g.Burgess & Tyburski, 2020;Ciornei & Østergaard-Nielsen, 2020;Hutcheson & Arrighi, 2015), or just as an element of the research design (e.g.Collyer, 2014a;Gherghina et al., 2022;Lafleur, 2013).Here, we expand upon the work of Collyer (2014b) who has offered the only previous global analysis of special representation of nonresident citizens.iii While the normative literature has generally been aware of comparative empirical analyses, its conclusions have been confined to arguing for or against the desirability of special representation by considering democratic values that it might promote or undermine (e.g.Bauböck, 2007a;Häggrot, 2022;Owen, 2010;Rubio-Marin, 2006;Spiro, 2006).None of the normative texts that we are aware of has also engaged with the empirical variation in special representation that we focus on.
Accordingly, the contribution of this paper lies in linking an analysis of empirical variation in institutional design to a normative analysis exploring justifications for these different arrangements.It adds to this special issue's general theme of diaspora mobilization and homeland politics by examining a highly visible outreach policy towards diasporas.

THE EMERG EN CE OF S PECIAL REPRE S ENTATI ON
Special representation is neither a new electoral instrument, nor is it specific to non-resident citizens.Reynolds (2005) provides a global overview of different electoral mechanisms that ensure the representation of ethnic, racial, national or religious communities.As his work demonstrates, 'the practice of seat reservation, or gerrymandering for distinct communal groups, has long existed' (Reynolds, 2005, 302).Reynolds points to examples from colonial times when these electoral mechanisms were often used as an instrument of control by containing voting power.Identity-based conditions for reserved seats, such as religion, ethnicity or language, have also emerged in post-conflict consociational scenarios or where establishing a social contract was rather challenging (Collyer, 2014b).
Non-resident citizens' special representation is a type of discrete representation iv that has been adopted by national legislatures from the 1970s onwards.The spread of special representation mirrors the rise in diaspora institutions since the 1980s, as documented by Gamlen (2014, 182).In terms of geographical spread, with the exception of Asia, special representation arrangements appear at some point in all world regions.With four out of 19 countries, Latin America is the world region with the highest share of countries with special representatives of non-resident citizens in their legislatures.
The provision of reserved seats for overseas voters has been enacted either simultaneously with external voting rights or thereafter.In Table 1, we list the adoption and implementation of special representation separately.Unlike previous studies on emigrant enfranchisement and dual citizenship acceptance (e.g.Turcu & Urbatsch, 2015;Vink et al., 2019) that have found a pattern of policy diffusion, our data suggest this only with regard to Lusophone states and possibly also Francophone states in Africa (although Algeria adopted special representation earlier than France).Following in the footsteps of Portugal's pioneering role in special representation, most African Lusophone countries (Angola, Cape Verde, Guinea Bissau and Mozambique) have adopted laws creating reserved legislative seats for their emigrants (Hartmann, 2015).This pattern includes cases with an unusual trajectory of special representation, such as non-implementation (Angola) or retraction (Guinea Bissau).
Implementation of legislation introducing special representation has been uneven across countries.In nearly half of all cases, legislation and implementation happened within the same or the following year (Cook Islands, Dominical Republic, Ecuador, Guinea Bissau, Morocco, Niger, Peru, Portugal, Senegal and Tunisia).In a second group of nine countries, implementation was delayed for longer (Algeria, Cape Verde, Colombia, Croatia, France, Italy, North Macedonia, Mozambique and Romania) and in Angola and Mali the relevant laws have not been implemented until today.Table 1 presents more details in a chronologically ordered list of countries that have adopted special representation.
In most countries, the institutional provisions of non-residents' special representation have remained unchanged since the initial adoption.But six out of 21 countries that had at one point introduced special representation subsequently reduced the number of allocated seats or abolished special representation for non-resident voters altogether.Four countries (Colombia, Croatia, Italy and Mozambique) have reduced the number of reserved seats in their legislatures across parliamentary elections.We have not found any case where the number of seats has been increased since the initial adoption.In two countries (Morocco and the Cook Islands), special representation for citizens abroad was ended after having already been implemented.It is noteworthy that many of the earliest adopters of special representation for non-residents citizens are among those countries that reduced the number of seats or abolished special representation altogether.Seeing that roughly a third of cases subsequently restricted special representation suggests that this institution remains politically controversial after the initial implementation.This highlights the need for exploring normative arguments for and against particular arrangements.

NORMATIVE CONS IDER ATIONS ON S PECIAL REPRE S ENTATION OF NON -RE S IDENT CITIZENS
In this section, we discuss three dimensions of the special representation of non-resident citizens from a normative perspective: eligibility-who is eligible to vote and stand as a candidate in special representation designs; constituency arrangements-the institutional design behind extraterritorial districts and electoral proportionality-the comparison between the share of seats for extraterritorial voters in parliament and their share in the total electorate.

Eligibility
Evaluating the special representation of non-resident citizens normatively presupposes a generally positive attitude towards extraterritorial electoral rights as permissible, recommendable or even required.As is often the case, TA B L E 1 Adoption and implementation of non-resident citizens' special representation.Note: The information for each country-year reflects the state of legislation in effect on 1 January of that year.Where enactment and implementation occurred in the same year, as in Ecuador and Tunisia, we did not code the enactment of the electoral law the year after but the same year to avoid an artificial mismatch between the policy stages.See Table S1.
Source: Authors' own elaboration.ing rights but will find it difficult to restrict them to those carrying the citizenship of the country where the election is held (e.g.Goodin, 2007).A principle of including those and only those comprehensively subjected to the law leads to the contrasting recommendation of excluding non-residents from the franchise (e.g.Beckman, 2009Beckman, /2022;;López-Guerra, 2005).A purely procedural view of democratic legitimacy, like Schumpeter's (1942), would instead entail that democratic legislatures' decisions on whether to enfranchise citizens residing abroad are normatively unconstrained.
Some authors claim that disenfranchising citizens living abroad creates a kind of second-class citizenship (e.g.Rhodes & Harutyunyan, 2010).Yet, this critique begs the prior question of whether emigrants and their descendants have a right to retain their citizenship of origin in the first place.Bauböck (2007aBauböck ( , 2018) ) proposes a citizenship stakeholder principle according to which all those and only those who have claims to citizenship and voting rights whose individual interests in autonomy and well-being are linked to the collective interests of a polity in its self-government and flourishing.In Bauböck's view, this principle mandates external voting rights for coerced emigrants and temporary absentees as a matter of justice.This principle also provides general support for non-resident voting for the first generation of emigrants, as long as the non-resident citizen vote does not threaten to overwhelm the domestic one.
Simultaneously, this principle requires phasing out citizenship and/or voting rights for later generations born abroad who have not established sufficiently strong links with their ancestors' country of origin (Bauböck, 2007a).
Non-residents' special representation can be conceived either as a group right or as a mode of territorial representation (see Hutcheson & Arrighi, 2015, 898-899).Citizens living abroad, despite differences in their interests, depending on which generation of emigrants they belong to, which country or region of residence they live in, what ties they have to a country of origin, share special interests as a group, such as diplomatic and consular protection, citizenship law reform, facilitation of remittances and re-integration (Owen, 2010;Spiro, 2006).However, as pointed out by Häggrot (2022), it is unclear why non-resident citizens' interests are more special than those of other groups of voters who do not get reserved seats.Special interest representation in assemblies with general legislative powers can be criticized as a corporatist feature that challenges liberal principles associated with individual equality of citizens and their freedom to determine themselves which of their interests they foreground in their electoral choices.
Moreover, the delegates elected by non-resident voters for reserved seats have a mandate to vote on all issues debated in the legislature and not just on those concerning the interests of emigrants.The stakeholder principle suggests therefore that citizens living permanently abroad should be enfranchised solely if they can be assumed to share with all citizens an interest in the common good.General representation and assimilated voting should therefore be the default mode of non-resident citizens' representation (Bauböck, 2007a).In other words, we do not find that the group rights perspective offers a convincing justification for special representation.This does not imply that special interests of non-resident citizens should never count for their electoral inclusion.For example, Owen (2010) has suggested that a principle of including all subjected to the laws requires enfranchising them in referendums on constitutional changes and legal reforms that will determine whether they can retain their citizenship and voting rights.However, this argument cannot ground a claim to special group representation This leaves open the second interpretation of special representation as a mode of territorial representation, which we address in the next section.

Constituency structure
Territorial representation of electoral districts or subnational polities, a feature of many electoral systems, is another possible justification for special representation.Extraterritorial (also called overseas) constituencies for non-resident citizens seem to fit this logic.Häggrot (2022)  Contestation over a territorial interpretation of extraterritorial constituencies for non-residents has occasionally not been confined to debates within the country where the election is held.In some instances, host states have also considered special representation schemes as infringing on their territorial sovereignty.This was a background for Canadian objections against Tunisian and French legislative elections on its soil in the early 2010s (see e.g.Lafleur, 2013).Canada worried that an elected candidate representing a geographical constituency that includes Canada could claim to represent Canadian (dual) citizens residing in Canadian territory in a foreign legislature.As Mégret and Girard (2015) have argued, banning dual citizens from participating in elections of their second country of citizenship does not protect Canada against foreign interference in its domestic affairs.Such concerns are more reasonable where the enfranchised extraterritorial population is not of migrant origin but consists of territorially concentrated ethnic-kin minorities (e.g.ethnic Croats in Bosnia and Herzegovina).If such a territorial minority is represented through special seats in the parliament of a neighbouring kin state, this is likely to be regarded as a threat to the territorial integrity of the state of residence and to undermine internal accommodation through minority rights and territorial autonomy (Bauböck, 2007b).
Thus, both for domestic and international reasons, the territorial logic has shortcomings.We conclude that neither the group rights nor the territorial perspective provide convincing normative arguments for special representation.This is a pro tanto judgement that does not imply that the practice can never be justified contextually or that it somehow violates a basic democratic principle.As we will argue in the next section, special representation is indeed preferable in some circumstances.Before that, we still need to address the question of whether normative considerations can also guide the choice between alternative constituency designs.
A halfway special representation that counts all external votes in a single domestic district (usually the capital city), seems most difficult to justify.The only problem that it resolves is how to allocate external voters to domestic constituencies in assimilated representation schemes if these voters have never resided in the territory.This problem emerges, however, only in over-inclusive systems that produce citizens without any relevant biographical link to the country.Where only those non-residents who have a genuine link to the country, such as previous residence (e.g.Chile, see Finn, 2023), are enfranchised, it will be easy to allocate their vote to a domestic constituency based on their last residence or that of their family.Objections against counting all extraterritorial votes in one district carry even more weight.Doing so creates arbitrary inequalities among domestic voters depending on whether they live in the district where due to their concentration external votes may impact strongly on electoral outcomes, or in any of the other districts that are kept free of external voters even if these have roots there.
The choice between single and multiple extraterritorial constituencies is more contextual and less exposed to normative critiques.Where the overall size of the extraterritorial electorate and number of reserved seats is small and where the 'diaspora' is overwhelmingly concentrated in a particular country or world region, merging them all into a single constituency seems to be the obvious solution.Where it is large and widely dispersed, Häggrot's democratic reasons why territorial subdivisions promote democratic values may apply to some extent also among the extraterritorial electorate: multiple constituencies can balance out demographical distortions of geographical representation and introduce greater diversity of interests and perspectives among the special representatives, undermining the problematic notion of a homogenous national diaspora.

Electoral proportionality
Our argument in Section 'Constituency structure' that assimilated representation is generally preferable to special representation does not imply that the latter is illegitimate.In certain contexts, special representation may be preferable: specifically, where (1) the share of non-residents in the total electorate is very large v ; (2) there is a risk that the voting system will be manipulated to mobilize an extraterritorial electorate (Umpierrez de Reguero, 2022; Wellman, 2021) and (3) there is thus a danger that the external vote may not only tip the balance (which is always a possibility) but could structurally dominate the domestic vote (Bauböck, 2007a).
Hungary provides an example of successful incumbent mobilization and manipulation of external votes.After offering citizenship to Hungarian-speaking kin minorities in neighbouring countries in 2010, Viktor Orbán also granted them voting rights in national elections.His Fidesz Party has successfully mobilized this external vote in all national elections since then.In the 2014 elections, Fidesz achieved a constitutional two-thirds majority for the first time-a milestone for which the external kin minority vote was crucial.Among the manipulative elements in Hungarian elections is a difference in voting methods for ethnic kin minorities who cast their votes by mail ballot and for expatriates in the rest of the world who remain registered at a Hungarian address and need to travel to vote at embassies or consulates (Pogonyi, 2017;Waterbury, 2023).The latter group is much less supportive of Orbán's government than the former.
As pointed out by Spiro (2006), in such contexts, assigning a limited number of seats to external votes can have the effect of containing their impact, which in turn reduces incentives for incumbents to mobilize a 'diaspora' block vote.The reduction of extraterritorial seats in the Croatian Parliament at the beginning of the 2000s provides a relevant example.Croatia has a very large non-resident citizen vote, a significant share of which consists of a co-ethnic kin population in neighbouring Bosnia-Herzegovina.Before the so-called 'non-fixed quota reform' of 1999, the parliamentary majority of Tudjman's HDZ party was strongly bolstered by the special representatives of Bosnian Croats (Kasapović, 2012, 785).Multiple extraterritorial constituencies may offer additional opportunities for gerrymandering by assigning more seats to geographical districts where non-resident citizens are more likely to vote for the parties or candidates in power.

Containing the impact of external votes through special representation
Our empirical measure for the proportionality of non-resident citizens' special representation in Table 3 indicates that in two-thirds of the cases, special representation has the effect of under-representing non-residents compared to resident citizens.This suggests that legislatures choose special over assimilated representation less because they want to strengthen the influence of the non-resident citizenry but because they want to represent them while also containing their influence.
How should one evaluate such containment strategies from a perspective of democratic norms?Underrepresentation as well as over-representation can be regarded as a violation of the ideal of equality of political influence, which seems to provide another reason for preferring assimilated representation.However, if we start from the need for justifying the contested inclusion of non-residents in the demos, concerns about the danger of domination of domestic voters by external ones seem strong enough to reverse this judgement where the non-resident citizenry is very large and where incumbents have means to mobilize non-residents that the opposition is lacking.
Before endorsing special representation on this ground, one must also consider other ways of containing an excessive impact of the non-resident citizen vote.As an alternative (or addition) to reducing the number of seats elected by non-resident citizens, the number of eligible voters residing abroad can also be reduced by ( 1) limiting voting rights to a certain maximum period of residence abroad (e.g.Germany requires prior residence in the last 25 years) or to first-generation emigrants; (2) withdrawing citizenship from non-residents who voluntarily acquire a foreign nationality through naturalising in their residence country or (3) phasing out the extraterritorial transmission of citizenship jure sanguinis beyond the second generation, that is, the first-generation born abroad to emigrant parents, as long as this does not create statelessness.
The first of these strategies attributes to states a problematic power of disenfranchising individuals who have in the past legitimately held voting rights.Cut-off points for losing the franchise appear arbitrary and difficult to justify in individual cases.Once non-residents have gained a right to participate in elections, it is hard to justify that voters who have made use of their franchise in the past and have stayed informed about politics in their origin country or visited it frequently will lose their voting rights after a certain time.By contrast, the option of not enfranchising the next-generation born abroad seems justifiable if these persons have not established sufficiently strong links with their parents' countries of origin.While citizenship is acquired automatically at birth, citizens are enfranchised only when they reach voting age and qualify as active members of the demos.It seems therefore legitimate to add a condition of stakeholdership for the enfranchisement of citizens born and raised abroad.Not enfranchising those who do not qualify also does not deprive them of a right they have held in the past.
The second containment strategy is problematic if emigrants have a claim to citizenship based on their biographical ties to countries and if national citizenship is conceived as a lifelong status that states should have very limited powers to withdraw.In this view, migrants have strong claims to dual citizenship in countries of origin and residence.We do not find concerns about dual voting by dual citizens on grounds of equality convincing, since their votes in origin and residence countries are not aggregated and count only once in each election (see Blatter et al., 2022;Seubert et al., 2019; but see Goodin & Tanasoca, 2014).
The third strategy can be fully endorsed by a stakeholder principle.Phasing out jus sanguinis transmission after the first-generation born abroad ensures that extraterritorial citizenship is not over-inclusive for populations of emigrant ancestry or ethnic kin minorities.Since non-resident voting is contingent upon citizenship status, this prevents also an over-inclusive electorate.
To sum up, the choice of special over assimilated representation can be defended as a mitigation strategy against over-inclusion, but the need for such a 'special remedy' can in many cases be avoided by making sure that citizenship laws are not extraterritorially over-inclusive.There will still be a smaller number of cases where firstgeneration emigrant communities are so large that weighing down their electoral impact justifies assigning their representatives a limited number of reserved seats.

EMPIRIC AL VARIATION OF S PECIAL REPRE S ENTATION OF NON -RE S IDENT CITIZENS Eligibility
Non-resident citizens' special representation is shaped, to a large extent, by the eligibility conditions for voting for or becoming a special representative.Voters for special representation seats or candidates running for such seats are usually not required to be first-generation emigrants.In most cases, voting rights of non-residents are granted to all adult citizens, including emigrants' descendants (born abroad with jus sanguinis citizenship) and extraterritorial kin minorities.That is why we prefer to use the broader term 'special representation of non-resident citizens' instead of 'special representation of emigrants.'vi Generally, the selection of special representatives for non-resident citizens is limited to those living abroad, but there are exceptions or anomalies.For example, in the 2002 legislative election of Colombia, the one-seat representative for the international constituency was elected not only by non-resident voters but also by resident citizens (Escobar, 2007).Neither are representatives themselves necessarily non-resident citizens.In some cases, they can register a dual residence, such as in Romania (Vintila & Soare, 2018) and the Mexican province of Zacatecas.In other cases, after winning seats as challengers, representatives are no longer required to reside abroad to get re-elected (Fliess, 2021;Umpierrez de Reguero & Navia, 2024).
From a stakeholder perspective, which assumes a weakening of relevant ties across generations born abroad, it is remarkable that none of the 16 countries that currently apply special representation at the national level restrict voting rights to first-generation emigrants.However, at least one country (Italy) differentiates electoral rights between migrant generations when it comes to access-based conditions for exercising these rights (Arrighi et al., 2019).

Constituency structure
There is considerable variation in the design of special constituencies for non-residents.Generally, extraterritorial constituencies are created.vii Countries can establish a single global extraterritorial constituency (which may have one or multiple seats) or several constituencies defined by geographical subdivisions.As discussed above, the latter design is more likely where the aim is to represent geographical diversity among non-resident citizens and to avoid the demographically strongest sections of the 'diaspora' dominating representation in parliament.Another motive may be to give more weight to non-resident citizens in specific countries based on historical ties (especially colonial ones).For instance, Algeria reserves two out of six extraterritorial districts for Algerians living in France (Benzenine, 2019).In a similar vein, Portugal has created two districts that reflect within and cross-regional emigration-the Europa and Fora da Europa districts (Lisi et al., 2019).
In some cases, the conceptual distinction between special and assimilated representation is not straightforward.If in assimilated representation all extraterritorial votes are counted within one specific domestic constituency, this could arguably be considered a halfway form of special representation, particularly when the number of extraterritorial voters is sufficiently high to impact the party choice of candidates.More precisely, such arrangements do not amount to reserved seats or special representation of extraterritorial voters, but they imply their special non-representation in all other electoral districts that do not include nonresident votes.For example, in Poland, extraterritorial votes are added to those cast in the electoral district of Central Warsaw (Lesińska & Popyk, 2023).seats that are reserved for delegates elected by non-residents as part of the respective legislature.The seat share for special representation remains mostly constant within a specific country over time (see Figure S1).Croatia stands out as the only country with a sharp change, specifically a steep decline, in this measure.This change was implemented, first, with the introduction of a non-fixed quota, and thereafter, in 2010, when 'the two strongest parties in Croatia agreed in closed-door negotiations to elect three fixed diaspora representatives in future' (Kasapović, 2012, 785).On average, special representation seats make up 2.9% of all legislative seats, with the lowest share in Morroco in 1984 (0.3%) and the highest in Croatia in 1995 (8.6%).
By comparing with the share of registered voters residing abroad as part of the total electorate, we can calculate proportionality (see  As pointed out, reserving seats for non-resident citizens' representation varies across country cases and elections, but also between extraterritorial districts, generating relevant differences from a district-based perspective (see Table S2).

DISCUSS ION
In this section, we bring together the normative discussion and empirical observations.Our main point is that special representation of non-resident citizens is not the same everywhere and empirical differences matter also for normative evaluation.The most important variation is on the dimensions of eligibility, constituency structure and proportionality.From this starting point, we ask which of these institutional design features are normatively most justifiable.And to what extent does the observed empirical variation in special representation reflect what we identified as the normative ideal?
By grounding our normative discussion in empirical variation, we move beyond the question of whether special representation is desirable towards considering which specific features make it more or less acceptable.Regarding eligibility, we first rejected an 'interest group' justification for special representation and argued that assimilated representation is more in line with democratic inclusion principles.Second, eligibility for external voting should generally be restricted to first-generation emigrants.In this respect, our empirical findings indicate a pervasive pattern of over-inclusiveness.In all countries that grant special representation to emigrants' descendants (born abroad with jus sanguinis citizenship) or extraterritorial kin minorities are eligible to vote.Some countries, such as Italy, do however, differentiate between categories of non-resident citizens in access-based conditions for exercising these rights.
With respect to territorial constituency design, we concluded that allocating all external votes to one domestic electoral district, as is done in a few cases of assimilated representation, is hard to justify.By contrast, the choice between single and multiple extraterritorial constituencies with reserved seats should depend on the size and dispersion of the non-resident citizenry.Our empirical analysis shows that the allocation to single (42.9%) and multiple (57.1%) extraterritorial constituencies is slightly unbalanced, favouring geographical fragmentation to either better represent diversity within the 'diaspora' or to control the weight of the diaspora in contrast to the domestic vote.either at the aggregate national level, to a single pre-defined district within the country, or (most commonly) into different electoral districts, usually based on biographical ties such as a voter's last place of residence or ancestral home' (Hutcheson & Arrighi, 2015, 898).
iii Global comparisons on emigrant enfranchisement by, for instance, Wellman et al. (2022) provide some insights on the special representation of non-resident citizens but their main focus is on voting rights. iv The existing literature sometimes describes non-resident citizens' special representation as a form of 'direct representation' (e.g., Collyer, 2014b;Spiro, 2006).We refrain from this terminology as it wrongly suggests that assimilated representation is 'indirect'.Discrete and assimilated representation are both 'direct' in the sense that all members of a constituency can directly vote for their representatives-they differ in the drawing of constituency borders and the allocation of seats.Merging non-citizens' votes with those of resident citizens generates direct representation of transterritorial constituencies, rather than indirect representation of non-residents.The French Senate is a rare case of indirect special representation.The Assemblée des Français de l'étranger is elected by non-residents French citizens in 15 extraterritorial constituencies and sends 6 delegates to the Senate.
v Interestingly, Häggrot (2022) says his argument for special representation applies only to contexts where the share of expatriate voters is relatively small.By contrast, Bauböck (2007a) argues that special representation can be conditionally justified where this share is relatively large because of the possibility of weighing down the impact of external votes through reserved seats.
vi Hypothetically, special representation of emigrants could also be broader than that of non-resident citizens if noncitizen emigrants could also vote.However, we have not found any case where non-citizens without residence status are enfranchised.As the case of New Zealand shows, however, where non-citizen permanent residents are eligible to vote in national elections, they may also be allowed to do so from abroad.For this, they must have stayed on the electoral roll and must have visited New Zealand at least once a year-in contrast with non-resident citizens who must return to the country only once between elections (Barker & McMillan, 2016, 6).
vii In proportional representation systems where the whole country is a single electoral constituency (e.g., in Israel), parties could include a certain number of special representatives of non-residents on their lists or a party could be formed that claims to represent the diaspora vote.However, Israel does not allow for extraterritorial voting and we have not found cases where special representation occurs without extraterritorial constituencies.

14682435, 0 ,
Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/imig.13263 by European University Institute, Wiley Online Library on [10/05/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 5 SPECIAL REPRESENTATION OF NON-RESIDENTS however, different normative theories of democracy lead to contrasting conclusions on the legitimacy of external voting rights.Those defending the principle of including all affected interests in the demos can embrace external vot- offers a sophisticated defence of special representation of expatriates on the ground of democratic values promoted in general by geographically 14682435, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/imig.13263 by European University Institute, Wiley Online Library on [10/05/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License sub-dividing the demos into electoral constituencies.Doing so enables political campaigning and partisan mobilization; promotes deliberation among voters because of geographical proximity and among delegates in the legislature because of representation of geographical diversity within the demos and strengthens the value of equality of political influence of voters by securing a more even attention of legislators to geographically unevenly distributed groups among the electorate.Applying these criteria, Häggrot finds that-apart from deliberation among extraterritorial voters, which is unlikely to be enhanced among a geographically widely dispersed electorate-special representation of non-resident citizens contributes more to achieving these values than assimilated representation.The problem with this analysis lies in applying internal features of electoral systems that promote democratic values through the design of electoral representation of residents to the electoral representation of non-residents.The logic of territorial representation is grounded in a view of the polity as an internally differentiated jurisdiction and cannot be simply extended to territories outside its jurisdiction.What is represented in a parliament through reserved seats is not foreign territories but (usually small numbers of) citizens residing there.The decisive question is what justifies their inclusion in the electorate in the first place.On a stakeholder account, neither their special interests nor the fact that they are citizens suffice.What matters is the strength of their ties to their country of origin.The issues of who should be awarded citizenship and voting rights outside of a state's jurisdiction are thus inevitably contested.The citizenship stakeholder principle provides an answer to these concerns by justifying the inclusion of non-residents if they can be assumed to share with domestic voters a sufficiently strong interest in the future of the polity.This justification spills over into the question of general or special representation.It suggests that Häggrot's reasons for endorsing internal geographical divisions of electoral constituencies are only secondary considerations when it comes to designing the representation of external voters.Since the justification for inclusion lies in the contingent commonality of their interests with resident citizens, this provides a reason for assimilated representation that overrides the contribution of territorial subdivisions to a better quality of democratic representation.This conclusion becomes even stronger if some of the positive effects of splitting up domestic electorates into territorial constituencies are anyhow diluted or absent in large and widely dispersed extraterritorial constituencies.

14682435, 0 ,
Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/imig.13263 by European University Institute, Wiley Online Library on [10/05/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License A similar arrangement existed in Peru until 2017 (Umpierrez de Reguero, 2022).Since a recent expansion of non-resident citizens' political rights in the Netherlands in 2022, Dutch citizens living abroad can participate in elections of the non-residents' electoral college for the Upper House of Parliament (Eerste Kamer), registering their extraterritorial vote via the Municipality of The Hague (Law N° 316, 18 August 2022).

14682435, 0 ,
Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/imig.13263 by European University Institute, Wiley Online Library on [10/05/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License is limited to descriptive representation and does not address whether overseas voters are substantively represented; nor does our measure take into account different turnout among voters in the country and abroad.In most countries, overseas citizens are (slightly) under-represented.The most extreme case is the Cook Islands in 1999, with a proportionality measure of −32.3.Only six out of the 16 countries for which we have complete data overrepresent their non-resident citizens.

of special representation a (A) Share of extraterritorial electorate b (B) Difference (A − B)
Table3).Positive differences indicate over-representation of external citizens compared to domestic ones, while negative differences indicate under-representation.It should be noted that our measure TA B L E 3 Average proportionality of non-resident citizens' special representation (N = 60 elections).Note: The figures represent averages across elections.The column 'difference' is calculated as the strength minus the share of the extraterritorial electorate, indicating a plus (+) sign for over-representation and minus (−) sign for underrepresentation.Countries are ranked with regard to proportionality of representation.Extraterritorial (reserved) seats multiplied by 100, divided by the total number of seats.Non-resident citizens' electorate multiplied by 100, divided by the total electorate.14682435, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/imig.13263 by European University Institute, Wiley Online Library on [10/05/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License a b While countries such as the Dominican Republic, Senegal and Tunisia provide different numbers of extraterritorial seats per geographical region depending on where the estimated non-resident population is higher, other cases, such as France, North Macedonia and Romania, just allocate the same number of extraterritorial seats over time independently of the estimated population by extraterritorial district.In the latter scenario, this is usually a product of the overall electoral system design replicating the seat allocation domestically.In Ecuador, for instance, domestic districts with a reduced number of inhabitants, such as Las Galápagos or Morona Santiago, receive the same attention as each extraterritorial seat since the Electoral Law (2009) established a minimum allocation of two seats per constituency.As a result, Ecuadorians in the Europe, Asia and Oceania district get the same number of seats as those who reside in Latin America, the Caribbean and Africa constituency regardless of different demographical sizes.This produces an under-representation in the former district and an over-representation in the latter (Umpierrez de Reguero & Navia, 2024).Furthermore, there are countries like Italy that have reduced considerably the number of domestic and extraterritorial seats in the latest parliamentary election, affecting the proportionality of two subdistricts: Europe (from 5 seats up to the 2018 elections to 3 in 2022) and South America (from 4 seats to 2), where non-resident Italian citizens reside in great numbers.There are also cases like Cape Verde whose electoral system design prevents a future adjustment depending on aggregated voter turnout because it has standardized the number of extraterritorial seats across districts.