South African Parties Hardly Politicise Immigration in Their Electoral Manifestos

ABSTRACT To what extent do political parties in South Africa politicise immigration? We systematically analyse the party manifestos of all major parties in post-apartheid South Africa, using two separate approaches of content analysis: coding all sentences about immigration individually, and coding the electoral manifesto overall using a ‘checklist’. Although we can expect high politicisation of immigrants in new democracies, most party manifestos do not treat immigration at all. If parties in South Africa treat immigration in their manifestos, they tend to take relatively inclusive positions, focus on immigrant integration rather than immigration control, and use instrumental frames. It appears that the nation-building project of a post-apartheid South Africa has not led to an increased politicisation of immigration by political parties qua parties, although individual politicians certainly play a role.


Introduction
With the end of apartheid in South Africa, the nation embarked on an inclusive nationbuilding project, epitomised by the 'rainbow nation'. After decades of racial segregation, the new constitution of 1996 was lauded for its inclusive and progressive stances on intergroup relations: a major step towards uniting different parts of South African society. One way to achieve internal cohesion among South African citizens is by identifying against outsidersagainst non-citizens (Tajfel 1982;Posner 2017). The intuition is that by focusing on whom South Africans are not, commonalities within the nation become emphasised, which in turn strengthens cohesion. Van der Brug et al. (2015) describe how this happens in Western Europe where radical right-wing parties politicise and organise against immigrants as outsiders to strengthen traditional national identities. As a new democracy with great internal diversity, identifying and organising against outsiders may be an 'attractive' option, and indeed Harris (2002) has argued that the figure of the 'foreigner' in South Africa has emerged alongside the new inclusive nation-state discourse. Xenophobia is thus regarded as a perverse upshot of the nation-building project.
In outlining a boundary-making approach, Wimmer (2008) highlights that nation-building can indeed have a tendency to increase negative attitudes to foreigners since this constitutes a potential boundary between insiders and outsiders. Adida (2011) describes how ethnic community leaders actively emphasise such differences, because doing so serves their interests as 'ethnic entrepreneurs' to secure resources for their group (Brubaker Africa are widespread (Gordon 2017;Ruedin 2018), and that appealing to public opinion can be electorally beneficial, we can expect political parties to politicise immigration like many parties in Western Europe and North America do (Mudde 2007; Van der Brug et al. 2015). After all, by politicising against foreigners, they could (try to) strengthen internal cohesion. Similarly, Whitaker and Giersch (2015) observe that opposition to immigration in Africa is larger in more democratic countries, in countries with high ethnic diversity, and in countries with higher economic development. These all suggest that South African political parties should find it electorally advantageous to mobilise voters with anti-immigrant policies.
While the nation-building programme led by the African National Congress (ANC) has been described as clearly anti-immigrant (Neocosmos 2010;Gordon and Maharaj 2015), what has not been studied to date is whether there is evidence for this claim in electoral manifestos, where other political parties stand on immigration, and what role they play in politicising immigration. The focus here is on political parties as organisations that are major actors in the politics of South Africa. If research in Western Europe is anything to go by (Van der Brug et al. 2015), we can expect political parties as formal actors to play a major role in politicising immigration in post-apartheid South Africa. The present systematic content analysis of electoral manifestos, by contrast, reveals that political parties tend not to politicise immigrationnot as parties -, and tend to take relatively inclusive stances when they discuss immigration. It appears that the 'rainbow' is extended to immigrants, if they are considered at all.

Data and methods
Party manifestos are written in the context of competitive elections and formally summarise the policies and positions parties take to mobilise voters. While electoral manifestos are not typically designed to be widely read, they reflect the outcome of internal discussions about the policies and positions parties take during the campaign. Manifestos can be used to identify where parties stand on various issues, and to trace developments backwards in time. This is a major benefit over for instance expert surveys which tend to trace issues only after they have become sufficiently salient, and often only cover the main parties (see Ruedin and Morales 2017 for an extensive discussion comparing different approaches). For instance, in Europe questions of immigration play an increasing role in party competition (Green-Pedersen and Otjes 2017), but the Chapel Hill Expert Surveythe largest of its kindsonly added a question on immigration in 2006 (Bakker et al. 2015). Moreover, South Africa tends not to be included in these expert surveys.
Here we use content analysis to code 17 South African party manifestos from 1994 to the present. The main parties in post-apartheid South Africa are the African National Congress (ANC), the Democratic Party (DP)/Democratic Alliance (DA), Economic Freedom Fighters (EFF), the Inkatha Freedom Party (IFP), the Congress of the People (COPE).
The ANC is the social democratic party that has been ruling the country since 1994 with a clear majority in the National Assembly. Its origins lie in the early twentieth century to fight against the systematic oppression of Blacks. Following the accession of Jacob Zuma as party leader in 2007, COPE split from the ANC, gaining over 7 per cent of the vote in 2009, but under 1 per cent in 2014. The DA can trace its roots to the (liberal white) opposition Progressive Party under apartheid. This liberal party has gained vote shares in every election since 1994, currently holding 22 per cent of the seats in the National Assembly. Like the ANC, the IFP can trace its origin to the Black anti-apartheid struggle in what is today KwaZulu-Natal where it retains its strongest support base. The far-left EFF was established in 2013 and became the third largest party in the National Assembly in 2014.
To code electoral manifestos, we followed Ruedin (2013) and Ruedin and Morales (2017), who have validated different methods to extract party positions from electoral manifestos. In particular, we followed the two-stage coding of first identifying sections of the manifestos that treat immigration and immigrant integration, and second manually coding these sections. The first step identifies whether and to what extent political parties politicise immigration in their manifestos. Human coders manually selected sections of the party manifesto that treat immigration, assisted by a dictionary of keywords. The keywords were used to highlight sections of the manifestos that are likely about immigration; the selection was always done manually by the coders who used written definitions of what constitutes a sentence about immigration. For instance, in the sentence 'Our border controls will be strengthened to improve security and manage immigration effectively', the keywords 'border' and 'immigration' were highlighted by a computer programme, and the coders decided whether the sentence is about immigration and should be included in the content analysis. With the keywords, the task of selecting relevant sections in the manifestos can be sped up without missing relevant sentences. Both immigration control (as in border control) and immigrant integration (of those already in the country) were included as relevant topics. The sections on immigration were cross-validated across three coders: the author and two trained assistants. Of the 17 manifestos examined, only six had substantive contents on immigration, and these six sections were then coded manually to identify the nature of the politicisation. We used two approaches of manual coding because they offer the highest level of validity. On the one hand by coding each sentence according to a codebook, on the other hand by coding the entire section on immigration at once using a 'checklist'. After carefully assessing their suitability to the South African context, we used the codebooks provided by Ruedin and Morales (2017) for both approaches, which were developed in a cross-national (albeit European) context. To ascertain robustness and reliability, each manifesto was coded two or three times. 1 The sentence-by-sentence coding assesses each sentence about immigration in a manifesto individually, and determines the following: whether the sentence is about immigration control or immigrant integration, what aspect of immigration it is aboutsecurity and crime, the economy and the welfare state, politics and institutions, society and culture, and vague issues were differentiated -, whether the sentence is open to immigrants, progressive, cosmopolitan, or multicultural (versus restrictive, on a 5-point scale), and what argument or justification is given (see Table 1 for details). Given the ordinal nature of the response category, we used interpolated median values to summarise the central tendency of party positions on immigration (Revelle 2017). To make the numbers more accessible, the values were then rescaled to a scale from 0 (extensive) to 10 (restrictive) and averaged across coders.
The checklist approach uses 19 statements about immigrationboth positive and negative statementsand the coders decide whether the manifesto overall agrees or disagrees with the statement. The statements cover encouraging skilled immigrants for economic growth, making immigration control more restrictive, immigrants should accept South African culture and values, international terrorism and immigration, removing illegal immigrants, there are too many immigrants, immigrants are necessary for the economy, it is too easy to obtain South African citizenship, it is too difficult to be recognised as a refugee, family reunification has led to uncontrollable immigration, immigrants enrich local culture, criminal immigrants should be removed, a link between crime and immigrants, immigrants lower local wages, immigrants contribute to the welfare state, voting rights for immigrants, immigrants should be able to participate in public life, racism should be combated, immigrants undermine local culture. For instance, 'immigration policies oriented towards skilled workers should be encouraged as a means of fostering economic growth' is a positive statement, whereas 'there are too many refugees (numbers)' is a negative statement (see appendix for full details). The mean of the coded positions is taken as the party position, and then averaged across coders. These positions were then rescaled to a scale from 0 (extensive) to 10 (restrictive) to match those of the sentence-by-sentence coding. Here we report the average position between the sentenceby-sentence coding and the 'checklist' approach as the position of the party, but also present standard deviations to enumerate uncertainty around the estimates (see Ruedin and Morales 2017 for a discussion on uncertainty inherent in party positions on immigration). This extensive averaging was undertaken as a guard against the influence of differences in coding, whichgiven the relatively short sections on immigrationcould have a substantive impact on the calculated position. Table 2 summarises the four main variables. The salience of immigration is obtained by dividing the number of words in the section about immigration by the number of words in the entire manifesto. It is expressed as a percentage. For the issue (immigration control versus immigrant integration), we draw on the sentence-by-sentence coding, and report the mode across all sentences for a given party and year, irrespective of coder. The same approach is used for the frames, which can have the following categories: no argument provided; instrumental, pragmatic, utilitarian or goal oriented arguments; arguments about collective identity (e.g. nationality), ethics, values, community, culture and appropriateness; arguments about universal moral principles and rights (including legal arguments).
In the following, we present descriptive statistics, exploring the extent and nature of politicisation by political parties in South Africa.

Findings
Parties in South Africa often do not mention questions of immigration control and immigrant integration in their electoral manifestos. When they do mention immigration, they typically refer to immigration in a few sentences. The share of the section on immigration in the manifesto is given in the third column of Table 3. Only one manifesto -DA 2009contains a longer discussion of immigration and integration policies. With 2.5 per cent of the manifesto dedicated to the topic, the 2009 DA manifesto is comparable to the roughly 5 per cent Ruedin and Morales (2017) report for typical European parties, while the other manifestos treat immigration less extensively than the typical European manifesto.
With the low salience of immigration in party manifestos, there is no sign that political parties (as parties) drive the politicisation of immigration in South Africa. If anything, following the argument that immigration is likely politicised in new democracies, we would expect greater politicisation closer to 1994, when the nation-building process was still 'new'. The situation summarised in Table 3, by contrast, highlights that before 2004 none of the parties mentioned immigration in their electoral manifesto at all. While this absence of politicisation in party manifestos contradicts the theoretical expectation, it is in line with Segatti's (2011) observation that government focused on the economic transition, and barely treated immigrationas reflected in the low pace of adopting immigration policies despite pressures from a changed environment, including refugees from countries in the region. Segatti (2011) outlines how apartheid-era policies on immigration remained in place despite the fact their inadequacy to deal with the labour shortages and refugees from neighbouring countries has been obvious even before the democratic transition. While immigration policies were revised in 2002, 2004, 2007, and 2011, these revisions have failed to address the recruitment of skilled workers, despite a new preamble to the Immigration Act that claims to encourage immigration of highly skilled worker (Nkomo 2014). The revisions since 2007 strengthened enforcement mechanisms, although it is questionable whether there is sufficient state capacity to enforce these policies. With an increasing immigrant population and an immense backlog in asylum applications (Segatti 2011;Ruedin 2018), we cannot conclude that there was no political 'need' to deal with immigration. The slow and limited changes in immigration policies, in line with the limited politicisation in electoral manifestos, suggest that South African politics have been preoccupied with different issues. The chronologically first manifesto to treat immigration in post-apartheid South Africathe DA 2004includes one long sentence on the need to attract and encourage skilled immigration and abolishing restrictions. There is no reference to the limited policy changes introduced in the same year, and the manifesto does not provide policies to achieve the outlined aim. Five years later, the DA provide the only instance of proper policies on immigration and integration in any of the 17 manifestos examined. This manifesto includes a concrete plan of how the DA propose to reorganise immigration services and allocate clearer responsibilities. They repeat the difficulty of recruiting skilled immigrants mentioned in the 2004 manifesto, but the 2009 manifesto proposes a points system instead of the existing quota system. With the xenophobic violence of 2008 fresh in memory, the 2009 DA manifesto mentions immigrant integration by suggesting that immigrant 'assimilation' should actively be encouraged. The suggestion to work with NGO and community organisations to provide language training constitutes an actual policy, as does the proposal to grant certain students and workers permanent residence after 5 years. This expansive policy is countered by a commitment to fingerprint illegal immigrants and establishing permanent border control at high-risk areas. This means  Notes: Given is the year of the election, the party abbreviation (the DA and DP are differentiated to give the party name at the time of the election), the share of the manifesto about immigration calculated as the percentage of the section on immigration as a total of the manifesto (in words); the position the party took based on the average between manual sentence-by-sentence coding and manual 'checklist' coding of the manifesto as a whole (higher numbers denote more restrictive positions), along with the standard deviation of this position to indicate uncertainty around the position (StDev); the most common issue (mode, 1 = immigration control, 2 = immigrant integration) based on sentence-by-sentence coding; the most common frame (mode, 1 = instrumental, 2 = identity, 3 = moral principles) based on sentence-bysentence coding. No position can be calculated when immigration is not mentioned in the manifesto, hence the empty cells.
that overall, the 2009 DA manifesto provides a brief but complete vision of how immigration should be tackled: tough borders, open to 'wanted' workers, and 'assimilating' those already here. Probably relating to the 2008 xenophobic violence, the 2009 manifestos by the ANC and the IFP mention immigrant integration. The 2009 ANC manifesto includes a vague commitment to reducing xenophobia and acknowledges the contribution of immigrants to South African economy but stops short of outlining actual policy. Similarly, the 2009 IFP manifesto briefly discusses xenophobia, which it relates to there being too many 'illegal' immigrants present. The manifesto contains no outline how this problem could be resolved. Overall, it appears that the 2008 xenophobic violence led the main parties to discuss xenophobia and immigration, but only the DA proposed policies on immigration and immigrant integration. Whether this is because only the DA considered the issue important enough, whether there are strategic reasons for the differences, or whether the other parties were internally divided over immigration and chose not to take a clearer position the present analysis cannot tell.
If the DA provided a vision on immigration and immigrant integration in 2009, in 2014 they chose not to discuss it at all. This decision is in line with the general observation that political parties in South Africa do not politicise immigration and immigrant integration much in their electoral manifestos: not as parties. In 2014 the ANC dedicated a single sentence to strengthening border control to manage immigration better. Immigrants already living in the country are not mentioned, suggesting that the xenophobic violence of 2008 was no longer salient, despite continuing violence against immigrants (Misago 2016). The then still new EFF briefly discussed immigrant integration by including immigrants in a list of groups that should not be discriminated against in the labour market and enjoy equal rights in labour law. They focus on the unity of the working class, suggesting that this unity would help reduce xenophobiawithout providing reasons or direct policy proposals. All these observations lead to the same conclusion drawn from the limited policy changes over time: political parties did not politicise immigration as unified political actors as can be observed in Western Europe and as could be expected in new democracies like South Africa.
Where immigration is mentioned in the manifesto, we can calculate the position parties take on immigration. The positions shown in Table 3 are the mean of the sentence-by-sentence coding and the 'checklist' approach, and theoretically range from 0 (most extensive or positive to immigrants) to 10 (most restrictive or negative to immigrants). Most of the positions are on the extensive/positive end of the scale (fourth column in Table 3), suggesting that in the party manifestos the rainbow in the 'rainbow nation' is extended to include immigrants. Because of the limited datashort sections of the manifestos on immigrationthe positions vary noticeably between elections and come with great uncertainty as visible by the large standard deviations (fifth column in Table 3), if they can be calculated at all. With this it seems difficult to classify South African parties on their position on immigration between 1994 and 2014, but in their manifestos, they tend to be relatively extensive or positive towards immigrants. For comparison, Ruedin and Morales (2017) give the position of the French Front National in 2012 as 9.4, the British Conservative Party in 2001 as 7.3, or the Austrian Greens in 2008 as 2.2. Despite the relatively large uncertainty around their exact estimate, the positions South African parties take in their manifestos can be described as relatively moderate extensive/positive. When parties discuss immigration, they can discuss immigration controlborder control -, or the civic integration of immigrants already present in the country. Table 3 shows that parties refer to immigrant integration (coded 2, second column from the right) more often than immigration control (coded 1). This focus on immigrants already in the country is line with the liberal/extensive stance and the notion of an inclusive 'rainbow nation'. With a focus on the nation-building project and politicising against outsiders, by contrast, we would have expected greater emphasis on immigration control: keeping 'them' out. Put differently, by looking at the issue focused on, we have additional indications that South African parties do not use restrictive immigration as a means to bolster the new democracynot in their electoral manifestos.
Looking at the frames or justifications given for the positions the parties take, we note that the most common frames are instrumental (right-most column in Table 3). In none of the manifestos was the most common frame related to (national) identity, which is what we would expect from parties politicising in support of an inclusive nation-building project. Of all the frames identified across all sentences, only 6 per cent invoke identitybased arguments. The dominance of instrumental frames, however, is in line with the politicisation of immigration in Western Europe (Van der Brug et al. 2015). Put differently, looking at the frames or justifications, we once again find no evidence that South African parties would use immigration as a means to bolster the new democracynot in the electoral manifestos that reflect the position of the parties as formal political actors.

Discussion and conclusion
Drawing on a boundary-making approach, Wimmer (2008) highlighted how new democracies may politicise immigration to increase internal cohesion. In the South African context, this is often read as nation-building (necessarily) translating into xenophobia and the politicisation of immigration. With negative portrayals of immigrants in the media and purportedly high levels of xenophobia and xenophobic violence, there is evidence of this happening (Neocosmos 2010). Here we examined to what extent political parties politicise immigration in their electoral manifestosdocuments that reflect the position parties take as formal political actors. Contrary to what can be expected in a new democracy, and contrary to what can be observed in Western Europe (Van der Brug et al. 2015), political parties do not seem to be major players in politicising immigration in South Africa in their electoral manifestos. The near absence of politicisation by political parties in South Africa highlights that new democracies do not necessarily politicise immigration, and that the boundary between 'citizens' and 'immigrants' as the outsider is not necessarily emphasised across all contexts or by all political actors.
Rather than focusing on immigration, the political agenda is dominated by other issues, especially the economic transition and service delivery (Segatti 2011). This is in line with the observation by Knigge (1998) who highlighted that in Western Europe anti-immigrant parties are more successful at the polls when economic growth is high. The argument is that when the economy is going relatively well, there are few other pressing issues on the political agenda. This pattern can also be found at the macro level, where attitudes to immigrants are more negative in countries with higher levels of economic development (Whitaker and Giersch 2015). A preoccupation with other issues at the level of the formal organisation, however, does not preclude members of the party from engaging in the politicisation of immigrants, but there is no evidence that this is done strategically by the parties. While there is clear evidence that the media frequently take positions against immigration (Kariithi, Mawadza, and Carciotto 2017), they do this as part of a general focus on crime and issues close to the everyday lives of ordinary South Africans.
In this context, it should be emphasised that especially middle and low-ranked politicians do take negative positions on immigration (Neocosmos 2010;Segatti 2011;Whitaker 2017),as do community leaders (Adida 2011)but they may not do so systematically, and they do not do so as parties. These studies are in line with research in Western Europe and the United States that highlights that while the positions of parties may echo the concerns of the population, in the field of immigration it is particularly parties and the political elite who influence voters by legitimising certain positions and verbalising vague feelings of unease (Statham and Geddes 2006;Morales, Pilet, and Ruedin 2015).
Contrary to what happens in Western Europe and the United States, the situation in South Africa highlights that immigration need not be politicised by political parties as formal actors for xenophobia and xenophobic violence to take place: elite-led need not mean party-led politicisation. There is clearly some tension in South Africa that politicians can pick up in public statements (Segatti 2011;Ruedin 2018), and the party leadership is implicit in that they generally tolerate xenophobic statementsoften reacting late, and only when there is public pressure (Misago 2016). This suggests that anti-immigration positions may be welcomed to distract from the failures of government and parties to deliver, but they are not actively pursued as an official programme to mobilise voters. If politicising against immigrants were such a clear strategy to mobilise voters, this would be reflected in the party manifestos as we can find it in the manifestos of the radical right in Europe, for example (Ruedin and Morales 2017). The party manifestos suggest that as formal actors, South African parties remain committed to the rainbow nation, and that the rainbow is extended to include immigrants.
Drawing on party manifestos, we could see that the nation-building project of postapartheid South Africa has not led to an increased politicisation of immigration by political parties in their electoral manifestos. It appears that political parties are preoccupied with other issues that may be more pressing to voters, notably the economic transition and service delivery (Segatti 2011). While parties do not greatly politicise immigration in their electoral manifestos, where they do, they tend to take extensive/positive stances (rather than restrictive/negative ones). This suggests that the rainbow nation tends to be extended to include immigrants, akin to what Morales, Pardos-Prado, and Ros (2015) observed in Spain. In addition to generally extensive or positive positions, we also observed a focus on immigrant integration (rather than immigration control), and instrumental justifications (rather than identity or norms). This suggests a pragmatic approach to immigration rather than symbolic politics in support of the inclusive nation-building project. A change towards more post-material values where identity politics play a larger role rather than more immediate material needs tends to take many years after economic prosperity (Inglehart, Ponarin, and Inglehart 2017), suggesting that the politicisation of immigration by South African parties may not change much in the near futuredespite its potential in supporting an inclusive post-apartheid national identity.