1
The Economics of Minority Language Identity
Peter S. Li
University of Saskatchewan
Commissioned by the Department of Canadian Heritage for the
Ethnocultural, Racial, Religious, and Linguistic Diversity and Identity Seminar
Halifax, Nova Scotia
November 1-2, 2001
Available on-line at
The views expressed in this paper do not necessarily reflect
those of the Department of Canadian Heritage.
This paper is written for the Policy-Research Seminar on Identity, Halifax,
November 1-2, 2001, under the auspices of the Multiculuralism Program, Department of Canadian Heritage. Harley Dickinson, Li Zong and several anonymous reviewers read a draft of this paper and provided helpful comment. The author is solely responsible for the analysis, interpretation and views expressed in the paper.
Abstract
It is unclear whether ethnic identity constitutes an advantage or a disadvantage for social mobility. The inconclusive debate largely arises from the complexity and confusion in framing and measuring “ethic identity.” This paper focuses on a more refined dimension of minority language identity, as an aspect of ethnic minority identity, to see how it changes and what kind of market returns it brings. Using microdata of the 1996 Census, this study finds substantial variations in the adoption of a non-official language as mother tongue and home language, but such variations are more pronounced among foreign-born Canadians than among the native-born. By and large, non-official home languages and mother tongues produce net market penalties, whereas English mother tongue or home language yields positive net returns for both men and women. The paper suggests that the market disincentives associated with non-official languages and the incentives associated with English mother tongue or home language probably explain why minority language identity declines in Canada in favour of the English language. Although the study shows that minority language identity jeopardizes market outcomes, the debate regarding whether ethnic identity helps or hurts social mobility cannot be resolved without further studies treating ethnic identity as a multidimensional concept and determining how each dimension influences economic performance.
Social scientists have used different approaches in defining and measuring ethnic identity, and have come to conflicting conclusions regarding whether ethnic identity is a resource or a penalty for social mobility in Canadian society (Isajiw, Sev’er and Driedger, 1993; Kalbach and Kalbach, 1995). The controversy has to do in partstems partly fromwith the complexity in capturing the multidimensionality of ethnic identity, and in part, and partlywith from generalizing from narrow interpretations based on different measurements of ethnicity or ethnic identity. Nevertheless, claims about whether ethnic identity improves or retards economic performance have serious theoretical implications. If ethnic identity contributes to economic betterment, then economic improvement also provides the material incentives for strengthening the preservation of ethnic identity. Conversely, if ethnic identity produces negative economic outcomes, such outcomes serve as penalties for maintaining a distinct ethnic identity, and its future retention, is likely to be weakened.
Rather than trying to study ethnic identity as an all-embracing concept in order to determine its multidimensional effects on an equally encompassing concept of social mobility, this paper examines a more focused dimension pertaining specifically to minority language identity in Canadian society. The purpose is to determine the extent to which non-official languages are being adopted as mother tongue and home language, and to examine the economic value of minority language identity in the labour market. To the extent that non-official language identity provides positive market returns, such returns can be seen as providing incentives for its maintenance in the home setting. C; conversely, earning penalties of non-official language identity can be interpreted as disincentives that discourage its being preserved as mother tongue or home language. In this way, the market value of non-official languages provides an economic basis for strengthening or weakening minority language identity. Thus S, studying the economic returns of minority language identity consequently represents one way of assessing whether a specific component of ethnic identity helps or hurts labour market performance. In turn, the economic value of minority language identity provides a basis for understanding why minority language identity prospers or declines over time in Canadian society.
Ethnicity, Ethnic Identity and Minority Language Identity
The term “ethnicity” comes from the Greek word ethnikos, the adjective of ethnos, which means heathen nations or peoples not converted to Christianity. In the contemporary context, “ethnicity” is often used to designate the notion of a people of a similar heritage, the members of which have a sense of common origins and share some experiences of life, past and present (Cashmore, 1984:85-90). Thus, the concept of ethnicity implies an identity or a sentiment of likeness based on descent, language, religion, tradition, and other common experiences (Weber, 1968:385-398).
The internationalization of capital and labour under capitalism has greatly fractured the solidarity of ethnic identity that used to correspond more closely to nations and peoples. At the same time, increased international migration as a part of the trend of globalization has resulted in people moving across nation-states in large numbers; in doing so, s, and cultivating in destination societies have seen hybrids of ethnic identities emerge, which often reflecting the exigencies of contemporary life and only nominally on the endurance of cultural traditions (Yancey, Ericksen and Juliani, 1976). Today, modern states are likely to be polyethnic in the sense that many ethnic groups can be found within the same nation-state (Driedger, 1996:2-3; Kymlicka, 1995:11-26). As well, these states are witnessing what Krotki and Odynak (1990:415) call “the emergence of multiethnicities,” that is, the mixing of origins of people.
In an immigrant society such as Canada, there is no simple and direct correspondence between “ethnicity” and “ethnic identity.” A common ethnic label does not automatically imply a common identity because members of an ethnic origin are likely to have originated from many backgrounds and are exposed to different life experiences based on social class, gender, and other social features (Li, 1999:164-170). The end result is that there can be many identities even within the same ethnic group, with overlapping cultural and behavioural features. Thus, individual or group differences in attachment to various linguistic, social and cultural components that delineate ethnic identity may reflect less the robustness of common past traditions than the viability of present conditions that nourish or stifle such components.
Ethnic identities, as Jenkins (1994:218) puts it, “are practical accomplishments rather than static forms.” However, the conventional theoretical approach to ethnic identity tends to stress the collective internal definitions of distinctiveness at the expense of external definition and categorization (Jenkins, 1994). Ethnic identity cannot simply be simply a product of individual or group choices premised upon ascription and traditions, since unequal power relations exert substantial external pressures in the social construction of ethnic identity (Jenkins, 1994). As Gans (1997:882) points out, the retention and resurgence of ethnic identity can also be reactions to events in the larger society. Thus, the undue emphasis of ethnic identity as adherence to cultural tradition and internal solidarity has meant that external conditions and market forces are often overlooked as important sources in shaping ethnic identity.
Studies of ethnic identity are further complicated by the fact that there are substantial differences in how “ethnic identity” is framed and measured. For example, the debate over Porter’s Vertical Mosaic Thesis regarding whether ethnic affiliation determines the occupational opportunity in Canada is in fact premised upon the a narrow empirical question of whether ethnic categories as indicated in Canadian censuses are related to the occupational distribution, and whether such a statistical association increases or decreases in strength over different censuses (Porter, 1965; Darroch, 1979; Lautard and Loree, 1984; Lautard and Guppy, 1999). While these studies may have implications on the theoretical debate about ethnic identity and social mobility, such implications can only be extrapolated, and not inferred.
Since “ethnic identity” encompasses many elements and since there is no consensus in adopting a universal set of dimensions in studying “ethnic identity,” there are substantial variations in theorizing the concept of ethnic identity (see Cornell and Hartmann, 1998: 153-194; Gans, 1997; Hutnik, 1986; Jenkins, 1994; Tilley, 1997), as well as in measuring itits measurement (see Isajiw, 1974; Driedger, 1996:129-151; Kalin and Berry, 1995; Statistics Canada and US Bureau of the Census, 1993). Consequently, conclusions regarding “ethnic identity” are influenced by different elements being included or excluded in mapping out the composite concept. For example, a study of ethnic identity and social mobility among four European groups in Toronto used four dimensions to measure ethnic attachments to one’s ethnic community, and found a small but significant correlation between the external/cultural dimension of ethnic identity and social mobility in several tests (Isajiw, Sev’er and Driedger, 1993). The external/cultural dimension includes items which mainly measure ethnic language use and preference of ethnic food and ethnic media, while social mobility is constructed from occupational and educational differences between the respondent and his/her father (Isajiw, Sev’er and Driedger, 1993). The authors concluded that ethnic identity can be as much a resource as a drawback to social mobility (Isajiw, Sev’er and Driedger, 1993).
Kalabach and Kalbach (1995) analyzed data from the 1981 and 1991 Census to see if socio-economic status is related to what they call “ethnic connectedness” or “ethnic identity,” as measured by the proportion of an ethno-religious group reporting the use of an ethnic language at home. They concluded that “individuals in the more traditional ethno-religious groups, who exhibit their greater ethnic commitment or connectedness through greater use of their ethnic language in the home, tend to report lower levels of educational and economic status attainment than those who are less ethnically connected…” and that among immigrants in particular, those who are more ethnically connected tend to be more disadvantaged (Kalbach and Kalbach, 1995: 31).
In another study, Pendakur and Pendakur (forthcoming) used the detailed microdata file of the 1991 Census to examine the returns of official languages and non-official languages, and found that while official language ability brings positive market returns, knowledge of non-official languages rarely improves and in fact penalizes labour market outcomes. It is difficult to extrapolate from the study whether knowledge of non-official languages can be equated with ethnic identity, since, as the authors pointed out, individuals may be endowed with a non-official language as mother tongue, or may acquire such language ability in later socialization. Nevertheless, the study by Pendakur and Pendakur (forthcoming) casts serious doubts on the general conclusion regarding the positive effects of ethnic identity on market performance.
The present analysis assesses the economic value of minority language identity using the 1996 Census in order to better understand the retention or demise of minority languages. Minority language identity is defined as the adoption of a non-official language as mother tongue or home language. This is not to say that those who adopt a non-official language as mother tongue or home language would necessarily have a strong sense of belonging to one’s ethnic group. However, it is clear that the retention of a minority language as mother tongue or home language constitutes an added component in the construction of ethnic minority identity, and that people who retain such a mother tongue or home language have a stronger linguistic capacity to link themselves with their ethnic community than others who do not retain such language. Since the adoption of a minority language as mother tongue or home language represents an ethnic endowment and an ethnic language choice respectively, it can be described as a phenomenon of minority language identity.
Data and Method
The analysis uses the 1996 Census to see the extent to which non-official languages are being adopted as mother tongue and home language among immigrants and native-born Canadians of various ethnic origins. The market value of minority languages is then assessed in terms of the returns of non-official languages as mother tongue and home language, while controlling for variations in other individual and labour market features.
This analysis is based on the Public Use Microdata File on Individuals of the 1996 Census of Canada, which is a 2.8 per cent probability sample of the population enumerated in the census. The file contains 792,448 records of individuals.1 For the purpose of this analysis, only permanent residents of Canada, including landed immigrants and native-born Canadians, are included.2 A further restriction is applied to the analysis of the market value of mother tongue and home language. Those who were less than 15 years of age and those who did not work in 1995, as well as those with no wages, salaries or self-employment income are excluded. The resulting file has 401,653 cases, including 215,839 men and 185,814 women, at least 15 years of age, who worked in 1995 and had an earning from employment or self-employment.
The dependent variable is "annual earnings from employment and self-employment,” which is the sum of gross wages and salaries, and net self-employment income before paying individual income taxes. Statistics Canada applies upper and lower limits to individual earnings to ensure confidentiality.3 Wages and salaries are always positive, but net self-employment income can have a negative value. Earnings from employment and self-employment are used here to indicate labour market outcomes, and some individuals had earnings from both sources. Actual earnings are retained for easy interpretation.
The independent variables measuring individual variations in human capital and work-related features include: years of schooling, experience estimated by subtracting from age the years of schooling and the six years before schooling began, experience squared, knowledge of the official languages, mother tongue or home language, the number of weeks worked in 1995 (1 to 52), the nature of work in terms of whether the weeks worked were full-time or part-time,4 occupation (14 categories),5 and the industry of work (14 categories).6 In addition, a variable— “years since landing in Canada”— is used as a proxy of Canadian experience for immigrants. The variable is measured as the number of years since an immigrant has immigrated to Canada, and native-born Canadians are coded as 0. The "years of schooling" is constructed from several variables. For individuals with post-secondary education, the variable “years of schooling” is the sum of years of university or non-university education, whichever is higher, and 12 years of elementary and secondary grades. For those with secondary school graduation certificate, it is coded as 12. For those with less than secondary school graduation certificate, the highest grade coded is 11 even though higher completed grades may have been reported. Individuals with only "grade 5-8" education are coded as having an average of 6.5 years of schooling, and those with "less than grade 5,” an average of 2 years of schooling. In addition, two other variables measuring the characteristics of the local market are used in the analysis; they pertain to the unemployment rate and the percentage of immigrant population in the CMA as calculated from the 1996 census microdata file.
The inclusion of occupation and industry of work as independent variables reduces the magnitude of income differences that can be attributed to language ability and language identity. Indeed,,since language capacity and identity can be seen as exogenous factors that determine the type of industry in which a person works and the type of job a person holds. In turn, the job type and industry of work affect the level of earnings. Thus, earning differences, associated with language characteristics after variations in industry of work and job type, are taken into account and may be seen as direct net effects of language characteristics on earnings, as opposed to their indirect effects on earnings which operate via the influence of job type and industry of work. In other words, the inclusion of occupation and industry creates a more restrictive condition for assessing the impact of language characteristics on earnings.
Multiple Classification Analysis (Andrews et al., 1976) is used to analyze the gross and net differences in earnings which can be attributed to language abilities, including mother tongue or home language and knowledge of the official languages. The statistical procedure is essentially a least- squares solution which treats the dependent variable as a linear combination of a set of categorical and interval variables. For each interval variable in the equation, Multiple Classification Analysis calculates the unstandardized multiple regression coefficient; for categorical variables, it produces a regression coefficient for each category and expresses it as a deviation from the grand mean of the dependent variable. The gross deviations measure the effects when variations in other independent variables have not been adjusted; the net deviations are effects when inter-group variations in other independent variables have been taken into account.