CAPITULO III. DE LOS SUBSIDIOS
EL CUMPLIMIENTO Y LA PRESTACION DEL SERVICIO
We follow Hatton and Leigh (2011)’s econometric specification which is based on defining the supply and demand of ethnic labour for a given ethnic origin group. There is no clear and complete definition in the literature on what constitutes ethnic labour. Bonacich (1972) comes close in providing a definition. She understands the term ‘ethnicity’ as referring to groups defined socially as sharing a common ancestry in which membership is therefore inherited or ascribed, whether or not members are currently physically or culturally distinctive. Bonacich (1972) argues that labour markets include ethnic lines or groups and those differences exist not only across ethnic groups but also within groups. Individuals within a group vary markedly in skills, and these differences may lead to overall differences in average skills across different groups or the labour market. However, theoretically, we expect ‘like’ inputs or ethnic groups to be more suitable substitutes than ‘unlike’ ethnic groups. For example, white Western European immigrants are similar to Canadians/or share more similarities to native Canadians than black and Hispanic immigrants.
My work is partly related to this literature on ethnic labour but not completely. Workers are classified in different ethnic groups not because they play a different role in production but because their wage depends on natives' attitudes towards different origin groups.
We assume that ethnic origin is related to country of origin but it is not the same. People born in the same country might have different ethnic backgrounds. For example, from the 2006 PUMF of the Canadian census, from the immigrant sample, 5% were born in China. Immigrants from China reported ethnic origins other than Chinese (such as Canadian, English and German, amongst others).
In our study when we refer to 'ethnic group' what we truly mean is 'country of birth'. Precisely, the terms 'ethnic group', 'ethnic origin' or 'ethnic community' and 'place of birth' are used interchangeably throughout this study to mean country of birth. Pseudo- person groups are defined in terms of country of birth rather than ethnicity due to the stable and fixed definition of the place of birth of respondent variable in the census.
The Canadian census has collected information on the ethnicity of persons since 1861.However, this variable is not stable over time. The question wording and format, instructions and data processing on ethnic origin has changed over the years. For example in the 2006 census, respondents were asked: "What were the ethnic or cultural origins of this person's ancestors?" and in 2001, 1996 and 1991, the question was different: "To which ethnic or cultural group(s) did this person's ancestors belong?" However, in 1996, the format of the ethnic origin question was changed. The 1991 census question included 15 mark-in categories and two write-in spaces. The 2001 and 1996 questions did not include any mark- in categories. Respondents were required to write in their ethnic origin(s) in four write-in spaces. In 1996, the ethnic origin question gave 24 examples of ethnic origins whereas in 2001, there were 25 examples. Thus, historical comparisons of ethnic origins across censuses have limitations and should be made with caution. However, the place of birth of respondent variable has been relatively stable over the censuses. Furthermore, ethnicity is a self-perceived matter and respondents' knowledge of the ethnic and cultural history of their ancestors can influence the type of response given at the time of the census. New knowledge may be acquired and changed over time meaning that the same respondent may change their ethnic origin from one census to another. This means that two respondents with the same ethnic ancestry could report different ethnic origins. For example, a respondent may identify as 'East Indian' as an ethnic origin while another respondent with a similar ancestral background may report 'Punjabi' or 'South Asian'. Also, the social environment in which questions on ethnicity have been asked can influence the type of response one gives. The place of birth of respondent variable does not suffer from this, as it is objective and fixed.
Another issue is related to persons reporting more than one ethnic origin. Marriages, common-law unions between people from different cultural and ethnic groups and knowledge of family history may prompt individuals to identify with more than once ethnicity. From the 2006 PUMF immigrant sample, around 17% reported multiple ethnicities. Also, the census categories multiple responses in bundles. It does not identify the exact ethnic origins one picks in multiple response cases. All these issues make the construction of pseudo-person groups problematic. For example, if a Polish immigrant in Canada identifies as being of 'Canadian' and 'Polish' ethnicity, which pseudo-person group does this individual fit in?
One way of dealing with this is to remove individuals with multiple ethnic origin responses from the study. However, this means we lose observations and may bias the results. Using the place of birth variable, we do not have to deal with the issue of multiple responses as a person can only be born in one country.
For all the reasons mentioned above, we chose to look at place of birth of respondents as opposed to ethnic origin. Of course, it would be interesting to incorporate ethnicity into studies on immigrant assimilation, especially when looking at second- generation immigrants-something we do not do here.