Ethnolinguistic diversity in New Zealand: A socioeconomic analysis

1 Introduction

New Zealand has experienced greater ethnolinguistic diversification of its population over the last three decades than at any other time in the country’s recorded history. Reforms to the immigration policy from 1986 onwards ushered in a system that prioritized individual merit, in particular, skills, qualifications, and links to the New Zealand workforce (Bedford et al. 2002), to replace the previous preference for a narrow range of source countries (i.e., primarily the UK, Ireland and the US). This change of policy, together with the introduction of additional visa categories for temporary workers and investors, led to increases in the proportion of residents from non-traditional countries of origin (Bryant and Law 2004) and in linguistic diversity, in terms of number of languages and of speakers. The languages with the strongest growth in speaker numbers (e.g., Samoan, Chinese, Hindi, Tongan, Afrikaans and Korean) correspond to sustained or increased migration from the corresponding countries. Immigration has thus contributed to a rising proportion of the population that speaks languages other than English (Statistics New Zealand 2013).

The initial immigration policy reforms occurred within a broader context of profound liberal economic reform, and of advancing trade and political relations with the Asia-Pacific region. The conceptualization of the desired migrant that the immigration policy sought to attract reflected the reigning economic model. The system rewarded entrepreneurial individuals likely to succeed in a competitive business-oriented environment (Cain and Spoonley 2013; MBIE 2018). The points system introduced in 1991 favoured professionals with links to New Zealand-based employers, and with demonstrated English-language proficiency. Migrant labour has become crucial for both long and short-term staffing across a variety of higher and lower-skilled sectors including IT, health, agriculture and construction (MBIE 2013, 2017).

Targeted immigration was also intended to benefit New Zealand’s international connectivity by facilitating market access, interpersonal connections and knowledge exchange. Accordingly, New Zealand’s trade sector would experience lower transaction costs by benefiting from immigrants’ intercultural and linguistic skills, international ties, and knowledge of home country markets and business practices (Bryant and Law, 2004). Subsequent studies did indeed identify a positive association between immigrants and New Zealand’s trade with their countries of origin, particularly in the case of countries with a larger migrant stock in New Zealand (Bryant et al. 2004). A later study confirmed that migration stimulates trade (especially imports), and noted that effects were more pronounced for trade with developing countries (Law et al. 2013). Thus, the greater proportion of migrants from non-traditional countries appears to have contributed to New Zealand’s trade diversification.

Migration has fuelled urban growth in New Zealand, particularly in the primary international gateway city Auckland. The 2015 World Migration Report (IOM 2015) classed Auckland among the league of “world cities”, and–following Dubai, Toronto and Brussels–it was ranked fourth with regard to proportion of foreign-born populace. Previous work has noted the formation of “ethnoburbs” in certain areas of Auckland, in which evidence of a concentration of Chinese, Indian, Samoan or South African migrants is visible in the commercial linguistic landscape and the presence of cultural institutions (Spoonley and Meares 2011; Xue et al. 2012). The maintenance of the home language and cultural heritage of ethnic minority communities is supported both by the ethnic concentration of specific districts (Ishizawa and Arunchalam 2014), religious practices (e.g., Kolig 2010; Revis 2016), and the transnational social and professional lives typical of many migrants in Auckland (Feng and Page 2000; Friesen et al. 2005; Ip 2006; Liu and Lu 2015).

Similar to Australia (see Colic-Peisker 2011), however, the New Zealand labour market has been unable to absorb many of the highly skilled migrants it has sought to attract, and many have experienced periods of unemployment or underemployment (Bartley 2013; DOL 2011; Li and Campbell 2009). Long-term underemployment and the resultant deskilling can have serious repercussions on the individuals’ well-being, and represents an underutilization of human resources (Colic-Peisker 2011; North 2007; Watts and Trlin 1999).

Research into socioeconomic characteristics of ethnolinguistic minorities, in particular migrants, has considered the contribution of language and ethnicity-related factors to the phenomenon of lower income levels or underemployment (e.g., Grin 1996, 2003). A second line of research has considered the economic or professional benefits to employees of being multilingual, whether in languages which have official status in the country or in particular foreign languages. An important motive for this work was to assess whether investment in the acquisition and maintenance of additional languages can be associated with professional opportunities and income-related benefits. Such studies have demonstrated that particular languages are favoured in the job market (Stöhr 2015), with preferences for different languages in the public and private sectors (Limon and Novak-Lukanovič 2017), and in particular economic sectors (Grin et al. 2011).

While previous studies have identified the types of challenges or barriers experienced by ethnolinguistic minorities in New Zealand, in particular skilled immigrants from non-English-speaking backgrounds (e.g., North 2007), the demographic and socioeconomic indicators that characterize New Zealand’s multilingual population, and how these have evolved over a period of considerable demographic change, remains unexplored. This study focuses specifically on non-official languages in New Zealand – that is, languages other than English, Māori and New Zealand Sign Language.^[1] It thus encompasses languages that were either learnt as foreign languages, as heritage languages, or were learnt as a primary home or community language (otherwise termed mother tongue or first language).

I employ data from five censuses between 1996 and 2018 to explore the circumstance of being a speaker of non-official languages and being an English-speaking monolingual in New Zealand across a period of 23 years.^[2] The study considers the socioeconomic indicators that characterize different degrees of multilingualism – that is, cohorts of individuals with competence in a certain number of non-official languages–and compares the results with the English-speaking monolingual cohort. An additional focus concerns constructs which are typically employed to describe identity: birthplace, ethnicity and religion. These indicators provide insight into individuals’ origin and cultural orientation, and can also provide some indication of the motivation for acquiring and maintaining their linguistic repertoire. For instance, knowledge of one or more languages may be widespread in an individual’s country of origin or valued by the heritage culture. The two main research questions I address are as follows:

1. What is the proportion of the population of the following demographic groups? (a) Individuals who know one or more non-official languages; (b) individuals who are monolingual English-speakers.

2. How can these individuals be characterized with regard to the following socioeconomic variables: religious and ethnic affiliations, birthplace, education, occupation and income?

2 Multilingual skills and socioeconomic indicators

In adulthood, language acquisition and maintenance requires long-term commitment, considerable investment (time and material resources) and usually entails opportunity costs. In turn, language proficiency may generate benefits with regard to employment outcomes, earnings, and improved life opportunities (or consumption benefits) (Grin 2003). In certain instances, possessing skills in selected languages may provide a competitive advantage for job seekers or greater job security in periods of economic downturn (Grin et al. 2011). In light of this, language competence can be viewed as possessing economic value (Chiswick and Miller 2014).^[3] The anticipation that competence in a specific language will contribute some form of economic or lifestyle benefit is, for many, a strong motivational factor for language acquisition, and for intergenerational language maintenance (Tawalbeh 2018).

Research in different geographic contexts has considered the economic or professional benefits of being multilingual in particular languages. It is important to distinguish between the status of the languages focused on, however. The focus of much research has been on migrants’ competence in the dominant language in prominent migrant-receiving countries such as Germany, Spain, the UK, the US, Canada, Australia and Israel. In most cases, a positive correlation between proficiency in the host country’s dominant language and earnings was found (Chiswick and Miller 1995). This association is mediated by other variables such as ethnicity or education levels. For instance, ethnic origin can contribute to enduring disadvantage in workplace earnings among migrant workers proficient in the dominant language (Dustmann and Fabbri 2003).

A second line of research has been on competence in non-official or non-indigenous languages, typically English, and the implications this has for income levels. Discrete studies employing sample survey data from German-speaking Switzerland, Israel, Germany and China have shown that strong English competence can usually be associated with higher earnings in certain skilled professions, but the relationship is weaker in low-skilled professions (e.g., Grin 2001; Stöhr 2015). These studies have pointed to a complementarity between (English) foreign-language competence, education levels and occupational skills in these countries.

Less attention has been devoted to migrants’ languages (other than English) and the implications competence in these has on employment outcomes. Stöhr (2015) examined the income differentials of migrants and Germans within the German workplace and found no advantage for migrants’ languages. However, migrants’ attained greater economic advantage from proficient English than their German counterparts, provided this was combined with dominant language proficiency (German). This advantage was limited to particular professions with an international orientation, which Stöhr (2015) attributes to additional advantages that migrant workers may possess such as international experience.

The level of proficiency in the language, and in particular skills (listening comprehension, speaking, reading and writing), will affect the extent to which language competence has beneficial outcomes for individuals (Grin 2003). For instance, according to Grin et al. (2011), foreign-language competence in the Swiss labour market only had economic benefits if a certain proficiency threshold was reached. Other studies have pointed to the particular importance for income levels of literacy skills in the dominant language (e.g., Chiswick 1991; Dustmann 1994). Due to the absence of information about overall or skill-specific proficiency levels in the datasets used to investigate socioeconomic variables and language competence, however, most studies do not make these distinctions.

Census data may be used to investigate employment outcomes for migrant groups, compared with native-born individuals. Employing a single dataset from the Australian 2006 census, Colic-Peisker (2011) demonstrated that employment outcomes are worse for immigrants from a non-English-speaking background, compared with Australian-born individuals and UK-born immigrants, although a hierarchy of relative disadvantage related to source country was discernible among the migrant groups.

In the only such study undertaken in the New Zealand context, Winkelmann and Winkelmann (2002) investigated the extent to which migrants successfully integrated into the domestic labour market from the perspective of employment status and income. The authors distinguished between migrants from English-speaking and non-English-speaking backgrounds, by using birthplace and self-reported language competence. The initial income disadvantage experienced by migrants (when compared with New Zealand-born individuals) was stronger for Asian and Pacific Island migrants, and diminished in most cases over several decades. Data employed in this study were from three censuses (1981, 1986, 1996) which preceded and followed the immigration policy reforms beginning in 1986. This timespan also coincides with a period of radical economic restructuring initiated in 1984, which resulted in escalating levels of unemployment and socioeconomic inequality, only moderated by a change of government in 1999 (Gamlen 2013). From the mid-1990s, small private enterprise (“ethnic businesses”) run by migrants became more common in New Zealand’s largest urban centres (Cain and Spoonley 2013; Meares et al. 2011; Spoonley and Meares 2011). In light of this changing socioeconomic context, more recent census data might demonstrate better labour market integration of ethnic minorities.

Research into associations between competence in non-official or migrant languages and earnings or occupation are not available within the New Zealand context. The information provided by New Zealand census data does not distinguish between proficiency levels or different linguistic skills (beyond the fairly basic level stipulated in the census question), and cannot therefore be used to identify whether an individual’s proficiency level can be associated with socioeconomic indicators. Nevertheless, qualitative studies have generally affirmed that employees’ competence in certain languages is advantageous in business contexts; this is particularly the case of Chinese for businesses with an international orientation or which target local “ethnic” customers (Meares et al. 2011; North 2007), and for selected Asian and European languages in the tourism sector (Watts and Trlin 1999).

Learning or maintaining one’s competence in one or more non-official languages as an adult indicates that individuals engage in social relationships or pursue interests or activities for which knowledge of such languages is advantageous. That is, one’s linguistic competence is usually a combined consequence of one’s domestic arrangements, social engagement, and professional, educational or leisure pursuits. In contrast to official languages, such languages do not receive direct state endorsement in the New Zealand context, and exposure to such languages through state-sponsored channels is usually marginal or absent. While some may be taught in the secondary school system, the limited provision for foreign-language learning in New Zealand’s mainstream schools typically results in modest proficiency levels among school leavers who opt to learn an additional language (East et al. 2012). Moreover, school enrolments in foreign languages have weakened in recent decades, due in part to the emphasis on STEM subjects for university entrance (East and Tolosa 2014).

3 Capturing identities through the census

The census is undisputedly an instrument of political and socioeconomic bearing. The selection and formulation of cultural categories (typically ethnicity, language and religion) entail choices which are sociopolitically motivated and have repercussions for the social groups identified (Kertzler and Arel 2002). These may include the official recognition of particular social groups (with a concomitant extension of legal and political rights), or may affect the distribution of state resources, such as the allocation of funding.

Individuals often possess a degree of agency with regard to how facets of their identity are recorded, and their identity-related affiliations are sensitive to local contextual logic. Owing to the inherently subjective and circumstantial nature of cultural categories, shifts in the immediate sociopolitical environment or a change of audience will influence individuals’ perception of the relative saliency of culture-related components of their identity (Kertzler and Arel 2002; Nagel 1994). Nevertheless, individuals’ self-identification with cultural categories in a census is circumscribed by the inclusion of predetermined categories, the optional or mandatory nature of these, and restrictions on selecting multiple categories. Thus, both structure and agency are implicated in the construal of individuals’ cultural identity by the census (Nagel 1994).

The cultural category “ethnicity” has received considerable attention in research into the use of the census. It is one of the principal categories employed in the formulation of a country’s national identity descriptors. Frequently, this category is cross-tabulated with socioeconomic indicators to establish the relative social advantages of recognized ethnic groups, and thus contributes to decision-making concerning the allocation of state resources. As a social construct, the factors used to determine ethnic affiliation change over time. For instance, Kukutai (2011) records the evolving conceptualization of indigenous Māori identity in the New Zealand census from a blood quantum measure in the early 1900s, to a declaration of ancestry or decent, or tribal affiliation in the early 2000s. In the current New Zealand census, an ethnic group is described as possessing a proper name, and shared elements of common culture, geographic origin and ancestry.

In the last two iterations of the New Zealand census (2013, 2018), the form provided eight pre-determined ethnicities (New Zealand European, Māori, Samoan, Cook Island Māori, Tongan, Niuean, Chinese, Indian), down from the 14 ethnicities listed in the 1996 census, and an open field for self-nominated affiliation. Multiple responses are possible, and a growth in multiple affiliations has been observed across recent censuses in New Zealand. In census output, ethnic group affiliations are individually counted. The same procedure of providing selected predetermined categories and an open field, with a provision for multiple responses, is followed for items concerning religion and language.

The formulation of census items concerning language varies greatly across countries. Items may distinguish between home or mother tongue, additional languages known, and skills (oral language or literacy) and levels of competence. Information derived from queries into whether a person “knows” or “uses” particular languages are hard to interpret, as they give no indication of the respondent’s communicative ability (see Holmes 1997). Greater informational value is attained if the question identifies the skill and the context or purposes of use, as this information can indirectly indicate the level of competence expected. Questions formulated too narrowly risk limiting results. For instance, the Australian census enquires into language(s) used in the home (intended to shed light on intergenerational maintenance), but thereby disregards alternative public domains (Clyne and Kipp 2006; Holmes 1997). Census items concerning language, especially mother tongue, can generate tension, particularly in contexts with a history of conflict related to political rights and the distribution of resources (Duchêne and Humbert 2018). For instance, census results may support dominant linguistic minorities in their bid to assert linguistic rights (Perry 2015).

The New Zealand census has included one item on the knowledge of languages (including the official languages) since 1996 and, unlike the Australian and Canadian census item (see Clyne and Kipp 2006), it has remained unchanged: “In which language(s) could you have a conversation about a lot of everyday things?” In the formulation of this item, the skill (speaking), the event (a conversation), and the general topic (everyday things) are identified. The latter facet signals a relatively informal or casual register and, indirectly, the level of proficiency. The potential informational value of responses to this question is thus relatively high.

The final identity-related category, religion, may be examined from three different perspectives: affiliation, beliefs and practices. Of these, affiliation is considered of primary public interest and, where this topic is included in the census, this is the focus of the corresponding question (Field 2014). This facet of religiosity does not indicate the extent of religious commitment, and it may merely signal ancestry or empathy (Field 2014).

The efficacy of measuring religious affiliation through the census depends on the formulation of the question item (Field 2014). The question item included in the New Zealand census (“What is your religion?”) suggests that respondents have a religion, and it is more likely to lead respondents to attest a religious affiliation than a question such as “do you have a religion” (Hackett 2014), although the latter has the inconvenience of requiring a follow-up query to determine affiliation. The position of the option “no religion” is also considered to affect results. The position in first place among predetermined options (as seen in the New Zealand census) is found to induce more respondents to select this category (although this result is considered to be more accurate) than when it is placed at the end of listed possible religions (Hughes 2007).

Multiple responses to this question are permitted in the New Zealand census; previous research has suggested that this approach is advisable as mixed-faith identities are increasing among particular ethnic groups (Weller 2004). In this sense, religious affiliation is similar to ethnic identity in that it is often found to be mobile (i.e., individuals may migrate between categories), susceptible to fusions, and influenced by social aspirations and personal interpretations of ancestry.

Monitoring religious affiliations through the census is an important facet of mapping the relative importance of institutionalized religions and the extent of religious diversity on national and local levels (Bouma and Hughes 2014). Census-generated information on affiliation can be cross-tabulated with other census indicators (e.g., ethnicity, language, education, income, place of residence) to deepen understanding of differences between demographic categories. For instance, the introduction of an item concerning religion in the 2001 and 2011 censuses in the UK enabled insight into the socioeconomic disparities between specific demographic groups subsumed into the broader “Muslim” or “South Asian” categories (Hussain and Sherif 2014).

4 Methods

This study employs a specially commissioned dataset from the five recent censuses (1996, 2001, 2006, 2013 and 2018)^[4] and which is limited to the “usually resident population” (i.e., excluding visitors) located in private dwellings on census night.

The census question that enquires into language competence was used to create cohorts of individuals who claim to know one or more non-official language(s). Individuals who claimed to know only English (i.e., who did not stipulate any additional language) formed the “monolingual” cohort. Multilingual people were grouped according to the number of non-official languages they claimed to know. Due to the low numbers of people in the more multilingual cohorts, and the resultant risks this posed to the anonymity of individuals, five non-official languages was the maximum number captured at the initial data extraction stage. When examining the association between the language variable and socioeconomic variables, the cohorts involving between three and five foreign languages were conflated (and relabelled as LG3+), as an additional measure to conserve anonymity. Only when reporting overall numbers, and for the questions permitting multiple responses (ethnicity and religion) are the cohorts disaggregated (see Tables 1–3).

A cross-tabulation analysis was undertaken of these cohorts of multi- and monolingual people using additional census data variables: religion, ethnicity, birthplace, income, occupational status and occupation type. This identified the proportion of people from each cohort with a particular characteristic.

When calculating the proportion of particular variables (e.g., birthplace), the total figure against which the proportion is measured is the total number of respondents for the respective category or question item, rather than the total population. This is due to the existence of empty or invalid responses and the restricted nature of some items. For instance, sections of the census related to education, income and occupation are only completed by individuals over the age of 15. With regard to employment status, the census only collects information from individuals who are in the labour market (whether employed or not) and excludes individuals who, whether due to retirement or incapacitation, do not work.^[5]

5 Results

This section presents the findings concerning the proportion of the population that knows one or more non-official languages, and the socioeconomic variables (place of birth, education, income and occupation, and religious and ethnic affiliations) that characterize speakers of non-official languages, in comparison to monolingual English-speaking individuals.

6 Discussion and conclusion

Major migration policy changes from the late 1980s onwards that replaced the country of origin preference with a focus on qualifications and expertise contributed to “a profound transformation” of New Zealand society (Bedford et al. 2002: 86). The resultant increased ethnolinguistic diversity of immigrants, concomitant with sustained emigration of native-born individuals (Bedford et al. 2002), has contributed to an increasing proportion of people in New Zealand who know non-official languages. That is, the proportion of monolingual English speakers consistently fell between 1996 and 2013 (albeit rising again in 2018).

The notion of an intensified form of demographic diversity characteristic of many migrant-receiving countries in the early 21st century was intended to capture a multiplicity of social attributes (Vertovec 2015). From the perspective of the migrant, an important motive for transnational mobility is the desire to improve socioeconomic outcomes of the individual or family. Thus, an inquiry into the trajectories, or the degree of social mobility of particular demographic groups needs to consider the interplay between identity-related features such as language, ethnicity and birthplace, and social indicators such as education and employment.

This investigation into the characteristics of individuals who know non-official languages or who are monolingual English speakers uncovered changes in the cultural identity and the socioeconomic composition of this population between 1996 and 2018. The vast majority of multilingual individuals were born abroad (tendency increasing), and an average of 41% in this time period had been resident for less than a decade. Results for multilingual people were contrasted with those for individuals who claim to know only English. Over 80% of these monolingual individuals were born in New Zealand and the majority identified with the New Zealand European ethnicity. Individuals born in Europe constituted the largest minority of English-speaking monolinguals; a reflection of sustained immigration from the UK (MBIE 2018).

Knowledge of non-official languages is a progressively “Asian”-influenced phenomenon. That is, the more multilingual cohorts (speakers of two or more non-official languages) are increasingly dominated by individuals who identify with the Asian ethnicity, and whose place of birth is in the Asian region. In recent years, the main source countries of migrants in the skilled and temporary worker categories include countries where multilingualism is relatively widespread (e.g., India, China, Philippines) (MBIE 2018), and languages commonly used in these countries were among the strongest growing non-official languages in New Zealand between 1996 and 2018 (Statistics New Zealand n.d.). This constitutes a change from the earlier census years of 1996 and 2001, where “other European” ethnicity and born in Europe was the dominant profile of individuals with competence in non-official languages.

Multilingual cohorts in this study tended to be affiliated to a religion. They contribute to the maintenance of faith-based cultural practices in New Zealand, as the growth towards “no religion” is considerably slower than among monolingual English-speaking individuals. The maintenance of Christianity as the most common affiliation is supported by the growth of migration from countries such as South Korea and the Philippines (see Park and Anglem 2012), but eastern religions (Hinduism and Buddhism) are the fastest-growing. The opportunities for multilingual services in religious institutions are common in the main urban centres of New Zealand, and these constitute an important domain for the maintenance of language and cultural practices by minority communities (Barkhuizen 2006; Kollig 2010; McCaffery and McFall-McCaffery 2010; Park and Anglem 2012).

Evidence was found for underemployment in the multilingual cohorts, apparent in the lower proportion of individuals in paid employment and lower income levels, despite the higher level of qualifications. This is a well-documented experience of migrants in New Zealand, particularly those from non-traditional migrant source countries (e.g., Li and Campbell 2009), despite an immigration policy expressly designed to attract migrants with the skills and experience required by the local labour market. A large number of individuals in the skilled migrant category is employed in sectors such as hospitality, retail and agriculture, and occupations in these sectors typically only provide modest remuneration levels (MBIE 2017), irrespective of the individual’s qualifications.

Additionally, support was found for the socioeconomic disadvantage experienced by Pacific Island migrants identified by Winkelmann and Winkelmann (2002) and more recent studies. For instance, Ishizawa and Arunachalam (2014) found that Samoans (New Zealand’s largest Pacific Island community) were more likely to reside in socioeconomically deprived areas of Auckland, and the community possessed a high ratio of first language maintenance (relative to other large ethnic minority groups). In the current study, the LG1 cohort, the only cohort with a high proportion of individuals identifying with the Pacific Islands (by ethnicity or birthplace), displayed a higher proportion of individuals in the low income category and a lower proportion of individuals in paid employment than was found for any other cohort or the population as a whole, across almost all census years (although the 2018 figures displayed some improvement).

In contrast, previous research has shown that the employment and income outcomes of UK-born migrants, New Zealand’s traditional primary migrant source country, are not different from native-born individuals (Winkelmann and Winkelmann 2002). This outcome is assisted by their (usually) English native-speaker status, their cultural proximity to the New Zealand European population, and the relative ease with which the official recognition of professional qualifications from countries such as the UK occurs. As the second main birthplace of the monolingual cohort is Europe, the positive employment outcomes reported in Winkelmann and Winkelmann (2002) contribute to explaining the somewhat stronger likelihood evident in the present study of English-speaking monolingual people to be in paid employment than individuals in the more multilingual cohorts.

Poorer employment and income outcomes of most multilingual cohorts, when compared to the monolingual cohort, cannot be attributed solely to the effects of prejudice in the labour market. Niches of “anti-Asian” or “anti-immigrant” prejudice certainly endure, and the Asian ethnic group is more likely to experience race-related discrimination (particularly in work-related contexts) than other ethnicities in New Zealand (Statistics New Zealand 2012). However, discrete public surveys have also identified a largely accepting attitude towards New Zealand’s growing Asian-heritage populace, and thickening ties with Asian economies (Butcher et al. 2015; DOL 2011). The immigration category of new arrivals also has an impact on work-related outcomes. For instance, principal applicant economic migrants typically experience better work-related outcomes than secondary applicants, or those with refugee or family reunification status (Chiswisk and Miller 2014; Colic-Peisker and Tilbury 2006; DOL 2011). New Zealand receives refugees from countries with traditionally high levels of societal multilingualism but low levels of English-language ability, and the difficulties in transferring qualifications hinder work-related integration (Marlowe et al. 2014).

The overall rise in levels of education, employment and income among the more multilingual cohorts over the last two census years suggests that individuals with this characteristic are becoming better situated in mainstream New Zealand society, at least in large urban centres such as Auckland. The increase in the proportion in the higher income category in the LG2 and LG3+ cohorts is comparable to the increase found for the total population, but the proportion in paid employment increases at a faster rate. In this sense, the results from this study reflect an increase in the proportion of ethnolinguistic minorities in the “middle-class” category, as identified in Colic-Peisker (2011).

While many who identify as ethnolinguistic minorities continue to experience obstacles in securing employment at levels commensurate with their qualifications and pre-immigration work experience, many nonetheless appear to be successful in reorienting their professional lives towards opportunities offered by the local labour market. Chinese and Indian ethnic minorities in particular display a strong propensity to engage in entrepreneurial activities (de Vries et al. 2015; Spoonly and Meares 2011). This expansion of small businesses managed by ethnic minorities in particular urban localities enables more recent migrants, particularly from non-English-speaking backgrounds, to benefit from co-ethnic relational networks for employment or entrepreneurial purposes (Cain and Spoonley 2013). Although this occupational category does not necessarily favour high earnings, it provides an opportunity to acquire local work experience and improve English language skills.

With the growing ethnolinguistic diversity of New Zealand’s population, concomitant with greater socioeconomic diversity of recent migrants to New Zealand, concrete state-sponsored measures to support the ongoing socioeconomic integration of ethnolinguistic minorities are warranted (Colic-Peisker 2011), in particular English-language support and re-training for the local workforce, in parallel with measures aimed at addressing institutional and attitudinal impediments to employment (Bartley 2013).