ECN 620: Applied Economic Analysis
Applied Economic Analysis
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
ECN 620: Applied Economic Analysis
Introduction:
With today’s interconnected world, the ability to speak a second language is a vital skill
that gives you a competitive advantage over those who are unilingual. Thus, there are many
reasons for learning a second language. These reasons may be for cultural, intellectual, personal
identity, political, and economic purposes. The focus of this study is the economic reasoning for
learning a second language and how it positions you as a more valuable asset in the labour
market.
As a result of Canada’s two national languages, English and French, the language policy
poses a demand for bilingualism for economic purposes. Some of the economic incentives for
learning a second language is that it gives “an advantage for access to the job market and for
obtaining a higher income” (Savoie, 1997). This advantage shows employers you are a more
diverse job applicant which may lead to increased salary and a higher demand than unilingual
job candidates, meaning a lower chance of unemployment. Many studies have been done over
the past decades about the economic benefits and costs of Anglophone and Francophone
speakers in the workplace. Some of the findings of these papers conclude that “other languages
are useful because they serve to express various cultures” (Grenier, 1987), and “a bilingual
country will have a greater capacity to contribute to the advancement of knowledge than a
unilingual country” (Vaillancourt, 1992).
These past studies use numerous methodologies and data sets to conclude their findings.
Each past study is based on different sources and may not have the same control factors.
However, these studies show reoccurring evidence that in Quebec, bilingual male Francophone’s
have a higher economic advantage than unilingual male Anglophone’s. However, this trend is not
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 3
entirely consistent for bilingual female Francophone’s in Quebec. Additionally, it is important to
note that this advantage has a lesser effect for the other Canadian provinces outside of Quebec.
The empirical evidence from these past studies show a more significant bilingual earnings
premium for men compared to women residing in the province of Quebec, and in one case
entirely insignificant for women (Christofides and Swidinsky, 1998). Furthermore, evidence
from past studies show even less significance for the rest of Canada for both men and women,
ranging from 2% to 15%. The reason for these large ranges of earnings premiums from past
studies are a result of a high percent error. This error is due to a lack of control for various
attributes in the data, such as not taking into account the quality of ESL or FSL skills.
To add further to the research from previous works, this study will use data from the 2016
Census Public Use Microdata File. To showcase the economic implications of learning a second
language in Quebec and the rest of Canada, the sample data will be strictly observed from
individuals with attributes based on the knowledge and the use of their English and/or French
skills. Other restrictions also apply to the sample data and methodology of this research study,
which will be provided in the following section. These restrictions and information will allow the
results of this research to account for and control the attributes missing from past studies to
provide a more accurate and unbiased result.
Methodology:
In this study, data projections were found using the R-Studio statistical software, as well
as from Microsoft Excel. The 2016 Canadian census data was inputted into R-Studio to calculate
the sample distribution and average earnings for the different language groups for both males and
females. Using the R programming language within the R-Studio software to create our output, a
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 4
table was formed (Table 1), mimicking Christofides and Swidinsky’s table from past studies. To
ensure accurate results, certain data points that were deemed outliers or not-applicable were
removed from the dataset. Numerous dummy variables were created to see how various
economic incentive-affecting factors correlated with average earnings between males and
females. Examples of these dummy variables include marriage status, educational status,
language status, etc. These are the variables and restrictions that allow the results of this research
in order to account for and control the attributes missing from past studies to provide a more
accurate and unbiased result. These attributes were applied to both genders, and the data was
split into the two groups to show comparing results. With the results in an organized form, it is
easier to process the economic incentives for learning a second language. Following this step, a
log-linear regression code was executed to output the human capital earnings function for both
male and female samples (Table 2). The results recorded in this table show the coefficients and
absolute t-statistic values for the constant, along with the dummy variables. The following
section will analyze the meaning between the correlation of these variables and the average
earnings for males and females. Furthermore, this analysis will add to previous studies,
answering which characteristics and attributes give you the greatest advantage in terms of
economic incentive.
Analysis:
The sample distribution and average earnings results for different language groups are
presented in Table 1. As Table 1 shows, out of the total sample size of 30,934 males, 82.19% of
these individuals are unilingual English speakers, while only 17.81% are bilingual English and
French speakers. It is important to note there is a higher percentage of English and French
speaking work environments (BIL/MEFF) in Ontario than there are bilingual but mostly English
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 5
speaking work environments (BIL/MEFE). While bilingual speakers are the clear minority in
terms of language in Ontario, this group of individuals have a higher average wage than their
unilingual counterparts. The highest average earnings of $98,967.68 going to BIL/MEFE male
individuals. On the other hand, out of the total sample size of 26,131 females, 77.80% of these
individuals are unilingual English speakers, while only 22.20% are bilingual in English and
French. Just like the male sample, there is a higher percentage of BIL/MEFF females than there
are BIL/MEFE. It is important to note that bilingual females have an overall higher percent
differential in terms of average earnings than males. However, the average unilingual earnings
are approximately 25.8% less for unilingual females than unilingual males and approximately
16.3% less for bilingual females than bilingual males.
The log-linear regression results are depicted in Table 2. This data output will show in
more detail which variables have a stronger correlation to average earnings for both males and
females. As Table 2 shows, high school and trade/college students have the higher significance
compared to university and postgraduate students. However, high school students and
trade/college students have a negative correlation in terms of average earnings, while university
and postgraduate students have a positive correlation. This shows returns are higher for both
male and female individuals with a higher education level. In terms of language and language
skills, Table 2 shows that these variables have a significance on earnings. Male and female
individuals who are bilingual and use both English and French frequently at work earn a higher
wage than those who are BIL_MEFE and UNIL_ENGLISH. However, the least significance
goes to bilingual speakers who use mostly French at work. This is highly likely due to the fact
that Ontario is a mainly English speaking province in Canada. In terms of occupation, we see
that both male and female individuals have greater significance with respect to earnings in a
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 6
professional or management position compared to semi-professional and blue collar positions.
There is a positive correlation with earnings with a professional or management occupation and a
negative correlation with the other two. This is the same for males and females and the
coefficients for both are quite similar. With respect to industry, we see higher coefficient with
public and service industries than goods with both male and females. This is most likely due to
the fact that that the need for bilingual individuals in Ontario in these industries is higher, and
thus earns them a higher average wage. R2 is a measure of how close the data is to the fitted
regression line. Typically, a value closer to 100% means the model explains all the variability of
the data, however the value of R2 in this case is quite low. Males have a 0.1282 R-squared
value and females have a 0.1513 R-squared value. This is most likely because we are dealing
with human behaviour and psychology. As a result of the extensive pool of variable possibilities
that may affect human behaviour, it is very hard to get definite results. Finally, in terms of
marriage, married males have a higher coefficient than married females, and thus have a higher
average earning.
Conclusion:
The focus of this study is the economic reasoning for learning a second language and
how it positions you as a more valuable asset in the labour market. Specifically, this study is an
analysis of how certain characteristic traits and knowledge can earn you an economic advantage
in the workplace in Ontario, Canada. As stated in the introduction, the error in terms of large
ranges of earnings premiums from past studies is due to a lack of control for various attributes in
the data. The findings in this study prove to be more accurate and unbiased with the inclusion of
these missing attributes. These missing attributes include taking into account the quality of ESL
or FSL skills, education level, marriage status, etc. The analysis of this study is based on the
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 7
sample distribution and average earnings for different language groups, as well as the log-linear
regression using the human capital earnings function for both males and females. This data was
taken from the 2016 Census Public Use Microdata File. From Table 1, the analysis concludes
that there is a majority of unilingual individuals for both genders. Out of the sample of males and
females, on average, males earn a higher wage than females in Ontario. Furthermore, being a
bilingual speaker earns you an even higher wage. This premium is highest for individuals who
speak mostly English in the workplace (BIL/MEFE).
Next, from Table 2, the analysis concludes that a higher education level has a greater,
positive correlation in terms of average earnings. The same can be said in terms of occupation.
The more prestigious management and professional careers offer a greater, positive correlation
than semi-professional and blue collar careers. Public and service industries have more
significance than the goods industry, which is most likely due to the need of bilingualism in these
industries. The language variables, UNIL_ENGLISH, BIL_MEFE, BIL_MEFF, and
BIL_FRENCH seem to show the greatest significance related to wages. Male and female
individuals who are bilingual and use both English and French frequently at work earn a higher
wage than those who are BIL_MEFE and UNIL_ENGLISH. However, the least significance
goes to bilingual speakers who use mostly French at work. This is most likely due to the fact that
Ontario is a mostly English speaking province in Canada.
Although this study has shown major implications for the reasoning behind calculating
average earnings for males and females in Ontario, there are still other influences that contribute
to this formula. It is difficult to list all possible attributes contributing to average earnings as this
topic deals with human behaviour and psychology. Nonetheless, as a result of Ontario being
mainly English speaking, being another unilingual English speaker does not separate you from
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 8
the rest, and thus does not give you an advantage in the labour market. Therefore, bilingual
English and French speakers earn a higher average wage. Along with this, as shown in this study,
other variables such as higher education, marriage status, and occupation and industry status also
play a part in calculating average earnings. Some more than others. These attributes should be
taken into consideration when finding economic incentive in the labour market. Furthermore,
these results will most likely stay somewhat linear over time, assuming Ontario continues to be a
mainly English speaking province.
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 9
References
Christofides, L. N., & Swidinsky, R. (1998). Bilingualism and Earnings: A Study Based
on 1971, 1981 and 1991 Census Data. In New Canadian Perspectives: Economic
Approaches to Language and Bilingualism, Ottawa: Official Languages Branch,
Department of Canadian Heritage. 123-186.
Christofides, L. N., & Swidinsky, R. (2010). The Economic Returns to the Knowledge and
Use of a Second Official Language: English in Quebec and French in the Rest-of-
Canada. Canadian Public Policy, 36(2), 4–10. doi:10.3138/cpp.36.2.137
Grenier, G. (1987). Earnings by Language Group in Quebec in 1980 and Emigration from
Quebec between 1976 and 1981. The Canadian Journal of Economics / Revue
Canadienne d’Economique, 20(4), 774-791. doi:10.2307/135415
Savoie, G. (1997). The Comparative Advantages of Bilingualism on the Job Market:
Survey of Studies. In New Canadian Perspectives: Official Languages and the Economy.
Ottawa: Official Languages Branch, Department of Canadian Heritage. 65-88.
Vaillancourt, F. (1992). An Economic Perspective on Language and Public Policy in
Canada and the United States. In Immigration, Language, and Ethnicity, Canada and the
United States (ed. Barry R. Chiswick), Washington, D.C.: The AEI Press. 179-228.
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 10
Appendix
Table 1
Sample Distribution and Average Earnings for Different Language Groups in ON, Canada
SAMPLE
SIZE
% OF TOTAL
SAMPLE
% OF
BILINGUAL
ANNUAL
EARNINGS ($)
%
DIFFERENTI
AL
MEN
TOTAL
SAMPLE
30934 100.00% - 83685.00 -
UNILINGUA
L
25426 82.19% - 81436.29 -
BILINGUAL 5508 17.81% 100.00% 70114.29 -13.90%
BIL/MEFE 2248 7.27% 40.81% 98967.68 21.53%
BIL/MEFF 3138 10.14% 56.97% 91509.50 12.37%
BIL/FRENC
H
122 0.39% 2.21% 78673.08 -3.39%
WOMEN
TOTAL
SAMPLE
26131 100.00% - 62843.00 -
UNILINGUA
L
20330 77.80% - 60443.64 -
BILINGUAL 5801 22.20% 100.00% 58717.56 -2.86%
BIL/MEFE 2503 9.58% 43.15% 76323.36 26.27%
BIL/MEFF 3010 11.52% 51.89% 67862.88 12.27%
BIL/FRENC
H
288 1.10% 4.96% 72907.58 20.62%
ECN 620 RESEARCH PROJECT 2 FINAL DRAFT 11
Table 2
Log-Linear Regression Results using Human Capital Earnings Function
MEN WOMEN
Coefficient |t| stat. Coefficient |t| stat.
CONSTANT: 11.11489 66.245 11.047607 116.471
MARITAL STATUS:
MARRIED 0.46797 37.828 0.268026 23.709
SCHOOLING:
HIGHSCHOOL -0.48951 21.303 -0.462721 24.684
TRADE/COLLEGE -0.31561 14.376 -0.410334 24.197
UNIVERSITY 0.11749 5.353 0.127398 7.617
POSTGRAD 0.16891 8.987 0.15681 8.359
LANGUAGE:
UNIL_ENGLISH 0.15613 0.938 0.074767 0.829
BIL_MEFE 0.15981 0.968 0.036901 0.417
BIL_MEFF 0.1747 1.046 0.017125 0.188
BIL_FRENCH -0.03558 0.325 -0.051859 0.773
OCCUPATION:
MANAGEMENT 0.19543 13.424 0.197737 7.394
PROFESSIONAL 0.13839 8.26 0.137539 4.803
SEMIPROF -0.19862 11.925 -0.225837 8.164
BLUECOLLAR -0.19942 13.598 -0.23058 8.534
INDUSTRY:
PUBLIC 0.13351 8.447 0.091395 7.196
GOODS 0.07322 5.602 0.03076 2.359
SERVICE 0.12924 8.218 0.07501 4.28
ADJ. R SQUARED 0.1282 0.1513
# OF
OBSERVATIONS
30934 26131