# Analytical Studies Branch Research Paper SeriesChanging Immigrant Characteristics and Entry EarningsAnalytical Studies Branch Research Paper Series Changing Immigrant Characteristics and Entry Earnings

by Feng Hou and Garnett Picot
Social Analysis and Modelling Division, Statistics Canada
Research and Evaluation, Citizenship and Immigration Canada and Institute for Research in Public Policy

Release date: February 17, 2016 Correction date: (if required)

Skip to text

Text begins

## Abstract

During the 1990s and 2000s, changes in immigration selection policies significantly altered the characteristics of new immigrants to Canada across a number of dimensions, including educational attainment at landing, immigration class, source region, pre-landing Canadian work experience and geographic distribution. These changes were designed primarily to improve immigrant economic outcomes at landing. This paper examines whether immigrant entry earnings improved as a result of these changes in immigration selection and, if so, which characteristics contributed most to the improvement. Among all new immigrants and principal applicants in the economic class, entry earnings, abstracting from economic cyclical variation, changed little during the 1990s and 2000s. This stability was the result of competing influences: some that tended to increase earnings, and some that tended to reduce them (such as declined earnings returns to some characteristics). The key changes in immigrant characteristics that increased entry earnings were the rising educational attainment at landing in the 1990s and the large increase in the share of immigrants with pre-landing Canadian work experience during the 2000s. The latter characteristic also accounted for the earnings advantage of provincial nominees over skilled worker immigrants.

Keywords: immigration, entry earnings, Canadian work experience

## Executive summary

Immigration selection policies changed significantly during the 1990s and 2000s, at least in part to improve immigrant entry earnings. After the decline in both relative (to the Canadian-born) and absolute entry earnings during the 1980s and early 1990s, there was a strong desire to improve the economic outcomes of immigrants shortly after their landing. Changes in selection policies and other factors altered immigrants’ characteristics across a number of dimensions, including demographics, source region, work experience and geographic distributions. This paper examines whether immigrants’ earnings immediately after their landing improved as a result of these changes and, if so, which characteristics contributed the most to this improvement.

Among all new immigrants, abstracting from economic cyclical variation, entry earnings—defined as earnings in the first two full years after landing—remained more or less constant throughout the 1990s and 2000s. The situation was very similar for principal applicants (PAs) in the economic class. During the 1990s, rising educational attainment at landing and the increasing share of immigrants in the economic class increased entry earnings. During the 2000s, a much more complex period in terms of immigrant selection, the factors that positively influenced immigrant entry earnings included changing distribution by immigration class, notably the rise of the Provincial Nominee Program (PNP); changing source region; and, for immigrant women, rising educational attainment at landing. These factors were offset by less favourable economic conditions in destination cities and regions in the late 2000s.

However, one factor dominated all others: the rise in the share of new immigrants who had Canadian work experience, often in high-paying jobs, prior to obtaining permanent residency. Changes in this factor tended to increase entry earnings during the 2000s far more than any other variable studied. The increase in pre-landing Canadian work experience accounted for most of the positive effect of the rise of the PNP on entry earnings during the 2000s, since the increase in work experience was heavily concentrated among provincial nominees. Furthermore, differences in pre-landing Canadian work experience between provincial nominees (with more Canadian work experience) and skilled workers (SWs) (with less) accounted for virtually all of the entry earnings advantage that the provincial nominees held over the SWs during the 2000s. While other factors, such as differences in geographic distribution (more settled in the West), educational attainment at landing, unemployment in the destination regions and cities, and source region, contributed, either in a small positive or negative manner, to the entry earnings differences between provincial nominees and SWs, their contribution paled in comparison with the pre-landing Canadian work experience factor. Once adjusted for differences in pre-landing Canadian work experience, entry earnings were virtually identical between provincial nominees and SWs. These conclusions were found for all new immigrants, as well as for PAs in the economic class, and were evident for both men and women.

It is likely that the pre-landing Canadian work experience variable used here captures at least three effects. First is the effect of Canadian work experience on earnings early in immigrants’ working life after landing. Employers appear to be more willing to remunerate such experience relative to foreign work experience. Second, this variable may also reflect a selection effect. When immigrants are selected from the pool of temporary foreign workers, they come with information regarding how well they performed in their jobs in Canada. If an employer seeks to change the status of temporary foreign workers to a permanent one, it is likely because they have done well in their jobs. Hence, much of the effect on entry earnings could be because of this selection process. Third, during the 2000s, many of the workers on temporary visas who attained permanent status worked in high-paying jobs.

## 1 Introduction

Canada significantly altered the selection policies for economic immigrants during the 1990s and 2000s, at least in part to improve the economic outcomes of immigrants shortly after their landing (see Ferrer, Picot and Riddell [2014] for a review). The policy changes, combined with other pressures, altered the characteristics of new immigrants. During the 1990s, there was a significant rise in the educational attainment at landing of new immigrants and an increase in the share of immigrants in the economic class. During the 2000s, a rising number of immigrants had worked on temporary work visas and, consequently, had Canadian work experience before landing. An increasing share of immigrants entered via the Provincial Nominee Program (PNP), with fewer settling in Toronto and more settling in the West. Immigrants’ educational attainment also decreased somewhat, and the source-region composition changed significantly. This paper examines whether these changes significantly affected entry earnings during the 1990s and 2000s and, if so, which factors were the most important in increasing entry earnings.

Much of the previous economic research on immigration focused on the decline during the 1980s and early 1990s in immigrants’ earnings relative to the Canadian-born immediately after landing. The extent of this decline and possible explanations for it have been well documented (Aydemir and Skuterud 2005; Hou 2013; Picot and Sweetman 2012; Reitz 2007). But entry earnings declined in absolute terms, not just relative terms. Using census data, Frenette and Morissette (2005) found that immigrants’ average earnings during the first five years after landing declined 15% between 1980 and 1990, and declined another 13% by 1995.

This paper examines whether or not entry earnings subsequently improved. It focuses on the 2000s, when significant changes were made to immigrant selection. However, it briefly analyzes the 1980s and 1990s as well. This paper further examines the role that changes to immigrant selection and related changes in immigrant characteristics—including demographics, immigration class, source region, geographic distribution and pre-landing Canadian work experience—played in any improvement.Note 1 In particular, it considers whether the rise of the PNP during the 2000s improved entry earnings and, if so, why. Entry earnings are defined as an individual’s average annual earnings during the first two full years after becoming a permanent resident.

## 2 Changing immigration policy

Changes in immigration policy significantly alter immigrants’ traits (Beach, Green and Worswick 2011). In the early 1990s, the points system was revised to attract more highly educated immigrants, and the share of immigrants entering in the economic class was increased. These changes increased the entry earnings of immigrants selected under the points system, i.e., the principal applicants (PAs) (Picot and Hou 2009). However, PAs constitute about 20% of all immigrants, so these results tell us little about what happened to entry earnings for immigrants as a whole. In the early 2000s, the selection system was revised once more under the Immigration and Refugee Protection Act (IRPA). The points allocated to higher levels of education at landing were again increased, and English- and French-language requirements and tests were strengthened. The changes to selection policies in the IRPA also resulted in a significant decline in the concentration of immigrants entering through the economic class in particular occupations (e.g., a reduction in the large share of immigrants who are engineering and information-technology professionals), and affected the share of immigrants from particular source countries (e.g., a notable decrease in the share of immigrants from China and an increase in the share from the Philippines). The IRPA changes increased the entry earnings of PAs in the economic class (Citizenship and Immigration Canada 2010). The guiding principle behind the revisions to the points system in the 1990s and 2000s was to attract immigrants in the economic class who could respond successfully in the face of labour-market change and would perform well in the long run. This strategy is the essence of the human capital model of selection.

However, in the early 2000s, there were increasing concerns regarding the ability of the immigration system to respond to short-term occupational skill shortages. To mitigate these concerns and achieve other goals (e.g., meeting the needs of provinces in population growth and labour demand), some existing programs were expanded and new ones created. A greater share of immigrants was selected via the PNP, particularly in the West. A skilled-trades program was created, ministerial instructions (which allow Citizenship and Immigration Canada to respond to perceived occupational shortages) were implemented, and the Canadian Experience Class, which allows international students with Canadian work experience or foreign temporary workers to become landed immigrants under certain circumstances, was created. The share of new PAs who had worked in Canada on a temporary work visa prior to becoming permanent residents rose substantially during the 2000s (Hou and Bonikowska 2015). This too could affect entry earnings.

These policy changes and other economic pressures tended to alter immigrants’ demographic characteristics, admission-class distribution, source regions, pre-landing Canadian work experience, and the cities or regions in which they settled. This paper focuses on how these policy changes affected trends in immigrant entry earnings.

## 3 Data and methods

The data in this study are derived from the Longitudinal Immigration Database (IMDB). The IMDB combines immigrant landing records and annual tax records for immigrants who arrived in Canada after 1979. Immigrants who have filed at least one tax return since 1982 are included in the database. Immigrant characteristics at landing—including age, education, marital status, source country, language and immigration category (e.g., skilled worker [SW], family and refugee)—are drawn from immigrant landing records. Information on earnings and other income, current marital status and place of residence is drawn from tax records. The sample for this study includes all immigrants who obtained permanent residency status between 1981 and 2010, and had positive earnings during at least one of their first two full years in Canada. The study includes immigrants aged 20 to 54 at landing.

To determine how these changes in immigrant characteristics affect entry earnings, “adjusted” entry earnings are produced by holding the characteristics of new immigrants fixed at the levels observed during the first year of the period of interest. The adjusted earnings represent those that would have been observed had the distribution of immigrant characteristics—including demographic characteristics, source region, pre-landing Canadian work experience and geographic distribution—not changed over time, and had the earnings returns of those characteristics remained at the average value observed over the period of interest (i.e., the 2000s). Hence, for any given year, the differencebetween the actual and adjusted entry earnings represents the earnings effect associated with the changes in the distribution of new immigrants’ characteristics.

Three regression models are used to produce the adjusted earnings.

The dependent variable is annual earnings during the first two full years after becoming a permanent resident. As a check for robustness, results were also produced using the log of annual earnings as the dependent variable, and the main findings remained the same. Earnings are measured in 2011 constant dollars.

Model 1 has annual landing cohorts and immigration class. This class variable has eight categories: SW PA, provincial nominee (PN) PA, live-in caregiver PA, business class PA, economic classNote 2 spouses and dependants, family class, refugees, and others. The coefficients on these dummy variables represent the difference, observed over the period of study, in the actual average annual earnings between each class and the reference group (SW PAs in this case).

Model 2 adds all of the control variables except for pre-landing Canadian work experience. These variables include age at landing,Note 3 educational attainmentNote 4 at landing, source region,Note 5 language,Note 6 geographic distribution,Note 7 and the unemployment rate of prime-age workers in each of the 14 Canadian regions in the year when earnings were measured (i.e., the regional unemployment rate). In Model 2, the coefficients on the immigration class variable represent the difference between each class and the SW PAs (the reference group), adjusted for differences by immigration class in the control variables.

Model 3 includes all the variables in Model 2 and adds pre-landing Canadian work experience. This variable has four levels based on annual earnings in Canada in at least one year prior to becoming permanent residents: over $50,000, from$20,000 to $50,000, under$20,000, and had no pre-landing Canadian work experience.Note 8

These models are run for each of the three periods: from 1982 to 1988, from 1988 to 1999, and from 1999 to 2010. The last period includes the entry earnings years of 2000/2001 to 2011/2012, and the focus is on this period. The three periods were selected so that the period end points would roughly correspond to variations in the business cycle (see next section). The analysis is also conducted separately for two immigrant groups: first, for allnew immigrants and, second, only for PAs in the economic class. Most of the changes to the selection system, particularly during the 2000s, focused on PAs in that class. The models are run separately for immigrant men and women.

A decomposition method is used to determine which of the many variables in the analysis affect changes in entry earnings the most. The “adjusted earnings” analysis determines the total change in entry earnings over the study period that is associated with the change in all characteristics. The decomposition determines the share of this total change that is accounted for by each variable, indicating which has the largest effect on immigrant entry earnings. Following a variant of the Oaxaca decomposition (Hou 2014), the contribution of explanatory variable ${x}_{j}$ to the total “explained difference”—accounted for by all the predictors ${X}_{i}$ in the model—is computed as $\frac{\left({\overline{X}}_{j2}-{\overline{X}}_{j1}\right){\beta }_{j}}{{\sum \left({\overline{X}}_{i2}-{\overline{X}}_{i1}\right)\beta }_{i}}$, where ${\overline{X}}_{j1}$ is the mean of ${x}_{j}$ at the beginning of the period, while ${\overline{X}}_{j2}$ is the mean of ${x}_{j}$ at the end of the period, and ${\beta }_{j}$ is the coefficient of variable ${x}_{j}$ in the model with pooled data of all landing cohorts over the study period.

## 4 Results

### 4.1 Long-term trends in entry earnings

Earnings are highly cyclical for new immigrants. Compared with the comparison group,Note 9 immigrant entry earnings are more strongly influenced by the state of the Canadian economy; they typically increase during economic expansions and decline during recessions (Charts 1 and 2). Entry earnings fell during the recessions of the early 1980s and the early 1990s.

The period of the early 2000s was somewhat unique for immigrants. Compared with the recessions of the early 1980s and 1990s, the slowdown in the early 2000s was mild. The unemployment rate in Canada rose by 4 percentage points in the early 1980s and 1990s and by 1 percentage point in the early 2000s. In spite of that, entry earnings for immigrant men fell significantly during the early 2000s: between the 1999 and 2002 landing cohorts, they fell by 17% overall, and by 23% among university graduates. Among the 2002 landing cohort, entry earnings for graduate men with a bachelor’s degree were only marginally higher than those for immigrants with a high-school education. Much of this unexpected decline was related to the bust in the information technology (IT) sector and the fact that a very high percentage of new immigrant men were in occupations related to engineering or computer science (Picot and Hou 2009). However, in the later 2000s, the IT bust faded and the economy expanded. Entry earnings recovered for immigrants—highly-educated immigrants in particular—and the entry earnings gap between the high-school educated and university-educated immigrants returned to its pre-2000 level.

Long-term trends net of cyclical fluctuations are delineated by focusing on earnings in 1982, 1989, 2000 and 2007, as well as the last data point, 2012.Note 10 Since entry earnings are defined here as the average of the first two full years in Canada, the long-term changes in earnings are examined between the landing cohorts of 1981 (earnings in 1982 and 1983), 1988 (earnings in 1989 and 1990), 1999 (earnings in 2000 and 2001) and 2010 (earnings in 2011 and 2012). The earnings years for these cohorts are close to the business cycle peaks. But, even among these peak years, there remain some differences in economic conditions, as measured by the unemployment rate. Hence, in the regression models, annual unemployment rates of prime-age workers, at the regional level, are controlled for when computing adjusted earnings.

For immigrant men, entry earnings fell by 9.4% between the 1981 and 1988 landing cohorts (Chart 1 and Table 1). After the 1980s, there is little significant change in the long-term trend in entry earnings for immigrant men. However, since earnings among the comparison group rose during the 1982-to-2012 period, the earnings gapbetween new immigrant men and the comparison group increased almost continuously (excluding cyclical variation). Immigrant entry earnings for the 1981 cohort were 74% of those of the comparison group, falling to 64% for the 1988 cohort, 60% for the 1999 cohort, and 55% for the 2006 cohort, and recovering marginally to 57% for the 2010 cohort.Note 11

Long-term entry earnings of immigrant women remained remarkably constant over the entire three decades,Note 12 abstracted from business cycle changes. Certainly, during the 1990s and 2000s, there was no sign of a significant change in earnings at cyclical peaks (Chart 2 and Table 1). Again, however, since earnings among the comparison group rose significantly over the three decades, relative earnings (relative to the comparison group) of new immigrant women fell. For the 1981 cohort of new immigrant women, entry earnings were 68% of those of the comparison group. However, this fell to 53% for the 2010 landing cohort.

This analysis considers primarily the entry earnings of new immigrants. For a broader view of economic outcomes, the unemployment or non-employment outcomes of new immigrants should also be considered. This study touches only briefly on this issue, since it is not central to the questions posed here. Appendix Chart 1 shows that, for new immigrant men,Note 13 the proportion employed (i.e., with positive earnings) was somewhat lower during the 2000s (at around 83%) than it was during the 1980s (around 91%). The more significant decline was observed among new immigrant women. Their employment rate fell from a peak of 78% during the 1980s to around 65% during the 2000s (Appendix Chart 2). For both men and women, most of the decline in the employment rate occurred during the recession of the early 1990s. This rate did not recover significantly during the economic expansion since 1993. From 1999 to 2007, the employment rate increased by 4 percentage points among new immigrant men and increased by 5 percentage points among new immigrant women.

Landing cohort 1981 1988 1999 2006 38,200 34,600 34,300 33,400 34,300 18,900 21,500 20,100 20,700 21,800 47,800 43,100 41,400 41,100 42,100 25,200 30,000 29,000 29,200 27,900 0.74 0.64 0.60 0.55 0.57 0.68 0.70 0.60 0.52 0.53 0.93 0.80 0.73 0.68 0.70 0.90 0.98 0.81 0.73 0.68 Notes: New immigrants include those who were aged 20 to 54 at landing and who had positive earnings in at least one of the first two full years in Canada. Entry earnings are defined as the average annual earnings during the first two full years in Canada, rounded to the nearest $100. For cohort years 1988, 1999, 2006, and 2010, the comparison group includes the Canadian-born plus immigrants who have been in Canada for 10 years or more (9 years or more for the 1988 cohort). For the 1981 cohort, the comparison group includes the Canadian-born plus immigrants in Canada for two years or more. Hence, entry earnings as a percentage of the earnings of the comparison group are somewhat overestimated in 1981 compared with other years. Source: Statistics Canada, Longitudinal Immigration Database. Description for Chart 1 Data table for Chart 1 Table summary This table displays the results of Data table for Chart 1 Comparison group, All new immigrants and Principal applicants in the economic class (appearing as column headers). Landing year Comparison group All new immigrants Principal applicants in the economic class 1981 51,430 38,154 47,846 1982 50,627 34,744 45,694 1983 51,324 29,947 43,423 1984 51,761 29,613 41,931 1985 51,643 30,479 42,790 1986 52,126 31,891 45,152 1987 53,516 35,548 47,021 1988 53,951 34,585 43,126 1989 52,855 30,107 38,575 1990 50,813 25,500 32,555 1991 51,036 24,271 33,308 1992 50,686 25,082 35,962 1993 51,757 26,306 36,237 1994 51,960 27,063 35,846 1995 52,175 27,640 35,314 1996 53,411 28,516 35,578 1997 54,696 30,522 36,872 1998 55,571 32,915 40,102 1999 56,837 34,340 41,382 2000 57,006 32,645 38,184 2001 56,902 29,387 32,803 2002 56,698 28,649 32,102 2003 57,555 28,708 32,455 2004 58,506 31,442 37,615 2005 59,840 31,322 36,694 2006 60,713 33,354 41,128 2007 61,024 34,177 42,952 2008 59,382 33,671 42,511 2009 59,999 33,306 41,877 2010 60,191 34,292 42,066 Description for Chart 2 Data table for Chart 2 Table summary This table displays the results of Data table for Chart 2 Comparison group, All new immigrants and Principal applicants in the economic class (appearing as column headers). Landing year Comparison group All new immigrants Principal applicants in the economic class 1981 28,040 18,940 25,169 1982 28,105 18,005 25,355 1983 28,377 16,752 21,024 1984 28,572 17,220 20,329 1985 28,845 17,711 21,670 1986 29,266 18,969 24,261 1987 30,135 20,909 28,377 1988 30,687 21,544 30,014 1989 31,173 19,888 27,848 1990 30,957 18,228 25,329 1991 31,991 17,581 23,160 1992 32,037 17,447 23,237 1993 32,578 17,186 21,593 1994 32,572 16,933 21,985 1995 32,711 16,945 22,144 1996 33,187 17,474 23,024 1997 34,187 18,491 25,505 1998 34,949 19,277 27,031 1999 35,952 20,058 29,026 2000 36,170 19,604 28,888 2001 36,608 18,230 25,421 2002 36,679 17,904 24,815 2003 37,371 18,163 25,398 2004 37,877 19,098 26,749 2005 38,883 19,497 26,848 2006 39,861 20,699 29,244 2007 40,479 21,466 30,121 2008 41,083 21,578 29,191 2009 41,238 21,437 28,612 2010 41,304 21,779 27,922 ### 4.2 Changing immigrant characteristics #### 4.2.1 During the 1980s Between the 1981 and 1988 landing cohorts, immigrant characteristics changed significantly. Most of these changes tended to reduce entry earnings, particularly among men. The share of new immigrants from Northern and Western Europe fell from 23.5% to 9.0% for men and from 19.6% to 8.7% for women. Meanwhile, the proportion of immigrants from East Asia (mainly China) increased (Table 2). Related in part to the source-region changes, language abilities in English and French declined. Among new immigrant men, the immigration class composition changed as well. The share of immigrants who were PAs in the economic class decreased, and the presence of refugees increased. All of these changes would tend to put downward pressure on entry earnings. #### 4.2.2 During the 1990s Significant changes in immigration selection occurred in the 1990s. The rise in new immigrants’ educational attainment was the most significant change in characteristics that was likely to affect entry earnings. This increase was related, at least in part, to the changes in the immigrant selection system in the early 1990s. Among new immigrant men in the study sample, the proportion of those with a university degree rose from 24.2% (for the 1988 cohort) to 52.9% (for the 1999 cohort). Among women, this proportion rose from 17.9% to 42.4%. These changes would tend to increase entry earnings (Table 2). There were also significant changes in immigration class, which reversed much of the change that occurred during the 1980s. Among men, the share of immigrants in the economic class (including SW PAs, business class PAs, and economic class spouses and dependants) rose from 52.6% (for the 1988 cohort) to 63.2% (for the 1999 cohort) (Table 2). This would tend to increase entry earnings, as economic immigrants earn more than other classes. During the same period, the proportion of refugees declined from 25.4% to 13.0%. A change in economic class was less evident among new immigrant women (Table 2). The share of new immigrants with some pre-landing Canadian work experience also increased, from 9.3% to 16.3% for men and from 11.1% to 14.6% for women. This would tend to increase entry earnings since pre-landing Canadian work experience is one of the most important predictors of immigrant entry earnings (Bonikowska, Hou and Picot 2015). Other noticeable changes included an increase in the share of immigrants from East Asia and South Asia. For men, this share rose from 26.8% to 44.6%, and, for women, from 31.9% to 42.7%. In the 1990s, the share of immigrants from Europe continued to decline from 28.7% to 22.6% for men between the 1988-to-1999 landing cohorts. These source-region changes would tend to reduce entry earnings, because, on average, immigrants from Europe earn more in the labour market than those from developing countries. #### 4.2.3 During the 2000s As noted above, the revisions to the immigration selection policy were more complex during the 2000s, resulting in large changes in the demographic characteristics, work experience, source region and geographic distribution of immigrants entering Canada. Immigration class was considerably altered. Between the 1999 and 2010 landing cohorts, the share ofimmigrant men whoentered as SW PAs declined from 51.1% to 33.4% (Table 3), while the share of immigrants who entered as PN PAs increased from 0.3% to 12.5%. Previous studies show that PN PAs earn more than their federal skilled worker counterparts during their first few years in Canada. This advantage diminishes after a number of years (Citizenship and Immigration Canada 2010). The share of immigrant men who were economic classspouses also rose, from 8.8% to 19.2%. This would tend to reduce entry earnings. The proportion of immigrant men with pre-landing Canadian work experience increased from 16.3% to 28.9% (Table 2), and most of this increase consisted of immigrants who had high-paying jobs before landing. This increase was likely related to the fact that provincial nominees are more likely than SWs to be selected by employers and to have previously worked in Canada. In terms of source-country composition, the proportion of immigrants from Southern and Eastern Europe and East Asia decreased while it increased from Southeast Asia and the Caribbean and Central and South America (Table 2). There was a significant shift away from Toronto as the first destination (the share declined from 44.9% to 30.5% over the period) towards Alberta (the share increased from 7.3% to 14.9%) and Manitoba (1.8% to 6.0%), as well as Montréal (12.4% to 16.0%). As with men, the immigration class under which women entered changed significantly. The share of all immigrant women entering as provincial nominees rose from 0.1% to about 6.3% between the 1999 and 2010 landing cohorts. The share of those who were live-in caregivers rose significantly (from 5.0% to 11.8%), and the share in the family class declined from 35.1% to 24.5% (Table 3). Interestingly, educational attainment at landing among women rose, unlike among men, where a slight decline was observed (Table 2). Landing year All new immigrants Principal applicants in the economic class 1981 1988 1999 2010 1981 52.3 40.6 51.5 46.6 94.6 90.7 94.8 96.1 3.0 4.2 2.9 1.9 5.4 9.3 5.3 3.9 3.5 7.8 8.8 19.2 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 25.4 21.2 23.2 20.1 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 15.8 25.4 13.0 12.2 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 0.1 1.0 0.7 0.0 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 23.0 18.4 9.7 9.3 15.5 8.2 2.8 2.8 27.8 24.4 24.4 22.2 27.6 22.1 24.0 22.2 21.1 22.6 25.8 24.4 24.9 25.5 29.1 27.3 12.0 16.6 19.0 18.9 15.3 21.2 22.1 21.4 7.2 9.6 12.0 12.8 9.4 13.5 14.0 13.7 4.6 5.1 6.1 8.2 4.9 6.8 6.4 9.0 4.3 3.2 3.0 4.2 2.5 2.8 1.7 3.7 18.7 17.6 7.0 8.2 13.0 12.6 1.6 4.4 29.6 32.2 22.4 20.2 24.3 23.6 7.4 8.0 28.7 25.9 17.8 21.7 33.7 30.5 15.0 20.9 16.9 18.6 37.6 32.4 20.2 24.9 52.1 38.3 6.1 5.6 15.3 17.6 8.8 8.5 24.0 28.4 23.5 9.0 6.6 7.2 35.7 14.5 8.1 9.2 16.6 19.7 16.0 7.9 11.3 13.1 16.0 8.6 5.4 7.3 9.4 15.2 6.8 7.0 7.9 17.4 11.0 17.7 22.4 11.1 13.7 28.1 31.8 12.9 8.3 9.1 22.2 20.6 5.6 7.2 20.0 20.7 13.9 13.7 6.3 15.9 5.1 9.2 3.5 14.1 3.2 8.2 7.0 7.3 4.2 9.2 6.7 6.9 10.6 12.0 7.7 11.7 9.8 8.5 4.6 7.8 1.7 1.1 0.8 1.4 1.7 0.9 0.4 1.4 5.8 2.3 1.6 1.9 6.2 2.3 1.0 1.0 1.8 1.0 1.5 1.7 2.4 1.6 1.9 2.3 1.9 1.2 1.9 3.9 2.5 1.9 2.7 6.1 25.6 40.0 55.7 58.0 26.2 52.1 66.0 56.8 4.3 2.6 4.0 4.3 5.2 2.7 2.9 3.1 3.1 3.6 4.6 11.2 3.5 5.0 6.1 16.0 32.8 36.4 22.7 11.6 20.8 18.7 12.7 8.3 30.5 15.3 9.6 9.4 39.4 17.9 7.8 7.4 8.9 5.8 5.5 6.4 9.0 5.8 5.5 6.3 Note ...: not applicable 1.4 2.4 7.8 Note ...: not applicable 2.4 3.6 13.0 Note ...: not applicable 3.4 5.4 12.9 Note ...: not applicable 3.4 4.2 16.5 Note ...: not applicable 4.5 8.5 8.2 Note ...: not applicable 3.3 5.1 7.2 0.3 0.2 0.1 0.3 0.3 0.2 0.1 0.3 0.1 0.1 0.1 0.2 0.1 0.1 0.0 0.2 1.1 0.7 0.6 0.8 1.2 0.8 0.5 0.8 0.5 0.3 0.2 0.6 0.6 0.3 0.2 0.8 2.1 1.5 1.8 3.3 2.1 1.5 1.7 3.9 16.9 15.4 13.9 9.2 18.7 13.3 14.6 7.5 4.2 2.8 1.8 6.0 3.0 1.8 1.6 7.9 1.9 0.7 0.7 3.2 1.5 0.6 0.6 4.4 14.6 7.5 7.3 14.9 15.4 6.7 6.4 16.0 4.0 2.2 2.3 2.9 3.8 1.9 1.6 2.6 0.3 0.3 0.1 0.3 0.3 0.3 0.1 0.3 13.2 10.4 12.4 16.0 14.9 11.6 11.3 18.8 30.8 46.3 44.9 30.5 28.6 47.9 47.5 25.0 10.2 11.7 13.9 12.0 9.4 13.0 13.7 11.5 ... not applicable Note 1. Provincial Nominee Programs were introduced in the late 1990s. Return to note 1 of table 2-11 referrer Note 2. Northwest Territories, Nunavut and Yukon. Return to note 2 of table 2-12 referrer Source: Statistics Canada, Longitudinal Immigration Database. Landing year All new immigrants Principal applicants in the economic class 1981 1988 1999 2010 1981 22.6 26.8 24.6 38.6 98.1 97.7 97.5 99.1 0.4 0.6 0.6 0.4 1.9 2.3 2.5 0.9 30.6 27.5 30.2 27.1 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 32.4 29.4 35.1 24.5 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 13.7 14.6 8.9 9.4 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 0.2 1.0 0.6 0.0 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 29.0 20.7 16.2 12.5 29.8 11.9 5.2 3.1 26.9 24.8 26.8 25.6 32.1 30.1 31.2 26.3 18.1 22.6 23.0 23.2 19.0 27.9 29.0 28.0 10.3 15.8 16.3 16.6 9.6 17.5 17.8 19.9 6.3 8.6 10.0 11.7 5.2 8.1 10.6 12.4 4.9 4.4 5.0 6.9 2.8 3.3 4.8 7.4 4.5 3.2 2.8 3.5 1.5 1.2 1.5 3.0 26.0 20.1 9.9 7.4 17.7 6.6 1.6 2.0 39.9 36.6 27.3 18.0 37.0 28.8 11.8 4.4 20.4 25.5 20.5 19.6 27.7 40.4 20.9 20.3 11.2 15.1 33.3 41.0 13.6 20.4 49.1 53.9 2.5 2.8 9.1 14.0 4.0 3.8 16.6 19.4 19.6 8.7 5.9 5.7 17.6 10.8 9.8 6.0 14.8 16.2 16.3 7.6 5.6 4.1 15.4 6.6 4.7 5.4 7.1 11.1 5.4 5.7 5.1 8.4 13.4 21.4 25.6 13.2 17.1 32.1 28.5 12.4 7.2 10.5 17.1 17.3 4.2 4.0 5.8 11.1 15.4 15.7 11.1 26.3 14.1 22.5 22.3 43.4 2.1 4.4 4.6 4.3 2.3 3.7 3.6 3.1 13.6 13.0 9.3 11.5 26.2 13.3 7.4 7.0 1.6 1.3 0.8 1.2 1.7 1.8 0.7 1.2 7.7 3.5 2.2 1.9 5.8 2.0 1.4 0.7 1.7 0.7 1.3 1.5 3.7 0.7 2.4 1.4 2.1 1.3 1.8 3.1 3.3 2.3 4.4 4.8 21.3 38.6 46.3 61.7 26.2 62.1 65.0 68.2 4.9 2.0 3.5 4.2 14.0 1.8 2.3 2.5 2.5 2.9 3.9 8.9 4.1 4.6 7.0 12.8 36.6 36.9 32.6 12.8 20.3 7.6 8.2 5.4 31.0 17.6 10.6 7.7 28.4 20.8 10.7 4.9 8.9 5.7 5.5 6.4 9.3 5.6 5.6 6.4 Note ...: not applicable 0.3 0.7 3.3 Note ...: not applicable 0.7 1.9 6.1 Note ...: not applicable 2.0 4.2 14.5 Note ...: not applicable 4.1 11.7 26.5 Note ...: not applicable 8.8 9.7 13.5 Note ...: not applicable 18.5 18.4 18.4 0.3 0.1 0.1 0.2 0.5 0.1 0.1 0.2 0.1 0.1 0.1 0.2 0.1 0.0 0.0 0.2 0.8 0.6 0.5 0.7 1.1 0.5 0.5 0.6 0.4 0.2 0.2 0.5 0.3 0.1 0.2 0.3 1.4 1.0 1.4 2.6 1.5 0.6 1.8 2.7 14.4 13.3 12.8 8.9 9.6 8.0 11.1 6.9 4.1 2.2 1.8 6.0 5.1 1.2 1.4 6.4 1.6 0.7 0.8 3.3 1.0 0.3 0.9 3.1 17.0 7.8 8.3 15.8 13.1 7.1 8.0 15.5 4.9 2.3 2.9 3.4 3.8 1.3 2.1 3.0 0.3 0.2 0.2 0.3 0.3 0.2 0.3 0.4 13.4 10.0 11.1 13.8 25.3 9.9 13.4 16.2 29.3 48.8 43.5 31.2 27.0 55.9 39.6 30.3 12.0 12.7 16.4 13.2 11.5 14.7 20.8 14.2 ... not applicable Note 1. Provincial Nominee Programs were introduced in the late 1990s. Return to note 1 of table 2-21 referrer Note 2. Northwest Territories, Nunavut and Yukon. Return to note 2 of table 2-22 referrer Source: Statistics Canada, Longitudinal Immigration Database. All new immigrants Principal applicants in the economic class 1999 2010 0.3 12.5 0.5 25.8 0.1 0.7 0.2 1.5 51.1 33.4 94.1 68.9 2.9 1.9 5.3 3.9 8.8 19.2 Note ...: not applicable Note ...: not applicable 23.2 20.1 Note ...: not applicable Note ...: not applicable 13.0 12.2 Note ...: not applicable Note ...: not applicable 0.7 0.0 Note ...: not applicable Note ...: not applicable 0.1 6.3 0.3 16.1 5.0 11.8 19.7 30.2 19.6 20.5 77.5 52.7 0.6 0.4 2.5 0.9 30.2 27.1 Note ...: not applicable Note ...: not applicable 35.1 24.5 Note ...: not applicable Note ...: not applicable 8.9 9.4 Note ...: not applicable Note ...: not applicable 0.6 0.0 Note ...: not applicable Note ...: not applicable ... not applicable Source: Statistics Canada, Longitudinal Immigration Database. #### 4.2.4 Changing characteristics among principal applicants during the 2000s For the 2000s, a separate analysis was conducted for PAs, as many of the policy changes were targeted at this group. There were significant changes in their characteristics. Some of the changes are similar to those observed among all new immigrants, particularly in terms of educational attainment, source region and geographic distribution. The economic conditions in the regions and cities that immigrants entered were less favourable at the end of the decade, negatively affecting entry earnings. The weighted average unemployment rate in these destinations was 6.4% in 2011/2012, compared with 5.5% in 2000/2001. Changes in immigration class were much more pronounced among PAs in the economic class than they were among all immigrants. For example, among men, the share of immigrants entering under the PNP rose from 0.5% to 25.8%, and the share of SWs fell accordingly, from 94.1% to 68.9%, between the 1999 and 2010 landing cohorts. Similar changes were observed among women (Table 3). The share of male PAs who had pre-landing Canadian work experience rose from 12.9% among those landing in 1999 to 36.7% among the 2010 landing cohort (32.0% to 51.0% for women) (Table 2). Among men, the share of immigrants with a high-paying job (paying more than$50,000) increased from 3.6% to 13.0%. As noted earlier, this increase was most evident in the PNP, where, for men, the number of PAs with pre-landing Canadian work experience rose from 10.6% to 61.5%, and the share of PAs with high-paying jobs prior to entry rose from 1.8% to 23.2% (Chart 3). Similar but less dramatic changes were observed among female PAs (Chart 4).

Description for Chart 3
Data table for Chart 3
Table summary
This table displays the results of Data table for Chart 3 With jobs that paid less than $20,000, With jobs that paid from$20,000 to $50,000 and With jobs that paid more than$50,000 (appearing as column headers).
Landing year and class With jobs that paid less than $20,000 With jobs that paid from$20,000 to $50,000 With jobs that paid more than$50,000
1999 SW 5.2 4.4 3.9
1999 PN 5.3 3.5 1.8
2001 SW 3.4 3.1 3.3
2001 PN 3.3 5.9 5.6
2003 SW 3.3 2.6 4.1
2003 PN 3.1 4.5 10.8
2005 SW 4.3 3.4 4.5
2005 PN 4.7 10.6 16
2007 SW 8.7 7.7 8.7
2007 PN 6.6 20.2 18
2009 SW 8.2 12.1 11.4
2009 PN 6.4 29.4 25.1
2010 SW 7.8 11.0 10.7
2010 PN 4.5 33.8 23.2

Description for Chart 4
Data table for Chart 4
Table summary
This table displays the results of Data table for Chart 4 With jobs that paid less than $20,000, With jobs that paid from$20,000 to $50,000 and With jobs that paid more than$50,000 (appearing as column headers).
Landing year and class With jobs that paid less than $20,000 With jobs that paid from$20,000 to $50,000 With jobs that paid more than$50,000
2003 SW 5.5 3.6 2.9
2003 PN 9.7 12.8 14.3
2005 SW 6.8 4.8 3.3
2005 PN 5.1 18.0 20.0
2007 SW 12.1 9.6 5.4
2007 PN 7.8 20.7 10.4
2009 SW 11.0 15.4 6.5
2009 PN 7.0 30.6 15.8
2010 SW 7.6 11.7 6.3
2010 PN 3.7 30.2 15.8
Control immigrant class only Control all except pre-landing Canadian work experience Add pre-landing Canadian work experience Note ...: not applicable Note ...: not applicable Note ...: not applicable 12,100 13,900 0 -5,800 -4,900 -6,700 -16,900 -8,800 -5,500 -12,800 -8,400 -5,800 -9,800 -8,300 -5,300 -14,600 -8,900 -9,600 -10,200 -5,600 -15,400 Note ...: not applicable Note ...: not applicable 71,800 Note ...: not applicable Note ...: not applicable 9,900 Note ...: not applicable Note ...: not applicable 800 Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable Note ...: not applicable 10,000 10,500 1,100 -3,900 -3,900 -8,700 -13,000 -7,900 -5,500 -10,700 -8,300 -6,600 -11,000 -8,100 -6,100 -12,000 -7,800 -7,700 -8,500 -4,400 -8,400 Note ...: not applicable Note ...: not applicable 61,700 Note ...: not applicable Note ...: not applicable 11,000 Note ...: not applicable Note ...: not applicable 3,500 Note ...: not applicable Note ...: not applicable Note ...: not applicable ... not applicable Note: All numbers are rounded to the nearest $100. Source: Statistics Canada, Longitudinal Immigration Database. ### 4.3 The effect of changes in immigrant selection on entry earnings during the 2000s Given the significant changes in immigration class composition, the analysis in this section will first examine the overall effect of changes in immigration class on the entry earnings trend. It will further examine whether changes in immigrant characteristics, particularly pre-landing Canadian work experience, account for the effect of changes in immigration class and whether they contribute to the entry earnings trends. #### 4.3.1 The role of the Provincial Nominee Program and pre-landing Canadian work experience in increasing the entry earnings of immigrant men The role that changing immigration class composition plays is examined by running the three models and the related decompositions outlined in the “Data and methods” section (Section 3). The coefficients for the regression models, using both log annual earnings and annual earnings as the dependent variable, are shown in Appendix Table 1. The results reported in the paper use annual earnings as the dependent variable. Since the focus is on the role of immigration class and pre-landing Canadian work experience, the coefficients for these two factors are shown in Table 4. In Model 1, the PN PAs’ entry earnings were, on average,$12,100 more than the SW PAs’ entry earnings during the 2000s. This is the difference in actual, unadjusted average entry earnings over all landing cohorts between 1999 and 2010. Earlier work also found entry earnings to be higher among provincial nominees than among federal skilled workers (Pandey and Townsend 2013). In Model 2, after controlling for all other variablesNote 14 except pre-landing Canadian work experience, this difference increased marginally to $13,900.Note 15 However, a large change occurs with Model 3, when pre-landing Canadian work experience is added. The difference in entry earnings between the PN PAs and the SW PAs goes from$13,900 in Model 2 to $0 in Model 3. Virtually all the difference in entry earnings can be attributed to the fact that, during the 2000s, the share of new PAs with pre-landing Canadian work experience was much higher among provincial nominees than it was among SWs. Among the 2010 landing cohort, 61.5% of PN PAs had some pre-landing Canadian work experience, compared with 29.5% of SWs (Chart 3). Perhaps more importantly, a much higher percentage had experience in high-paying jobs. Among male PN PAs, 23.2% had pre-landing Canadian work experience in a job that paid more than$50,000, compared with 10.7% of SW PAs. Only 4.5% of the PNs had a job that paid less than $20,000, compared with 7.8% of SWs. Therefore, the main reason for the higher entry earnings among PN PAs compared with SW PAs in the 2000s was the larger share of PN PAs that had high-paying jobs in Canada before landing. Other differences, such as the region in which the immigrants were working (and the shift to the West), mattered much less. Detailed decomposition further quantifies the contribution of the immigration class and pre-landing Canadian work experience to the trends in entry earnings. The decomposition was conducted for each of the three models, and focused on the change that occurs from 2000/2001 (the 1999 landing cohort) to 2011/2012 (the 2010 landing cohort). In the Model 1 decomposition, the change in the immigration class variable (the only independent variable) tends to increase immigrant earnings by about$800 (i.e., 100% of the change associated with the compositional changes) (Table 5). In Model 2, with all control variables except pre-landing Canadian work experience added, the change in the immigration class accounts for an even larger change in immigrant entry earnings—around $1,200 (i.e., 59% of$2,100).

In Model 3, after the pre-landing Canadian work experience variable is added, changes in immigration class have a small negative effect, if any, on entry earnings (Table 5). The pre-landing Canadian work experience becomes dominant, accounting for 92% of the $5,000 rise in entry earnings associated with changes in all of the variables. These results were replicated using the log of annual earnings (rather than annual earnings) as the dependent variable, and are shown in Appendix Table 2. The key findings remain the same.Note 16 Hence, virtually all of the positive effect that changes in the immigration class had on entry earnings was associated with a rise in the share of immigrants who had pre-landing Canadian work experience in relatively high-paying jobs. And, as noted earlier, this increase was most evident in the PNP. Other variables that had a significant effect on entry earnings included the shift in source regions, which tended to increase entry earnings, and the fact that economic conditions were worse (as measured by the unemployment rate) in 2011/2012 than in 2000/2001 in the regions and cities in which immigrants settled. The weighted average unemployment rate in the settlement regions and cities rose from 5.5% to 6.4% between 2000/2001 and 2011/2012. To summarize, there was no change in the entry earnings amongimmigrant men between the 1999 and 2010 cohorts, yet the changes in immigrant characteristics—including demographic characteristics, source region, geographic distribution and pre-landing Canadian work experience—tended to increase entry earnings by$5,000. Of the changes, the rise in the share of landed immigrants with pre-landing Canadian work experience accounted for most of the effect. Many immigrants with pre-landing Canadian work experience would have been temporary foreign workers, and earlier research suggests that they outperform skilled immigrants without any pre-landing Canadian work experience (Hou and Bonikowska 2015; Sweetman and Warman 2014). Pre-landing Canadian work experience also largely explains why PN entry earnings were higher than those of the SW immigrants throughout the 2000s.

#### 4.3.2 Results for new immigrant women during the 2000s

As with men, an analysis of the regression coefficients in the three models indicates that the pre-landing Canadian work experience variable accounted for virtually all of the higher entry earnings among new immigrant women entering through the PNP during the 2000s, in comparison with SWs. In the regressions, the coefficient on the PNP was around $10,000 in both Models 1 and 2 (Table 4). This means that immigrant women entering via the PNP had entry earnings of about$10,000 more than SWs—whether controlling for differences in characteristics (Model 2, except for pre-landing Canadian work experience) or not (Model 1). But when pre-landing Canadian work experience is added in Model 3, female PN PAs earn only about $1,100 more than SWs. Thus, almost all the difference in entry earnings between the two classes can be attributed to the difference in the share of immigrants with pre-landing Canadian work experience, particularly in high-paying jobs. For landing year 2010, 49.7% of new female PN PAs in the study sample had some pre-landing Canadian work experience, whereas only 25.6% of the SW immigrants did. Furthermore, 15.8% of new female PN PAs had a job that paid more than$50,000 prior to landing, compared with 6.3% of SW immigrants (Chart 4).

Decomposition analysis shows that the change in pre-landing Canadian work experience was the most significant factor influencing entry earnings among immigrant women, like it was for men. Average entry earnings increased by $1,700 for women between the 1999 and 2010 landing cohorts. When controlling for all independent variables except pre-landing Canadian work experience (see Model 2 in Table 5), changes in these control variables tended to raise entry earnings by$2,800, with changing immigration class accounting for the majority of the increase (54%). However, when the pre-landing Canadian work experience variable is added (Model 3), it dominates all others, accounting for 66% of the $4,300 change in earnings associated with changes in characteristics. Most significantly, the contribution of changes in immigration class becomes quite small (failing to 8%), implying that most of the positive effect of immigration class on entry earnings between 2000/2001 and 2011/2012 came through pre-landing Canadian work experience. #### 4.3.3 Change in entry earnings among principal applicants in the economic class during the 2000s The majority of the changes to immigration policy during the 2000s were directed at PAs in the economic class—those who were largely selected based on the points system. In this study, PAs in the economic class include those who entered as provincial nominees, live-in caregivers, business class immigrants, and SWs, or through the Canadian Experience Class.Note 17 Again, the analysis starts with men who landed between the ages of 25 and 54 and were employed at some time during their first two full years in Canada. Entry earnings of male PAs in the economic class fell dramatically during the early 2000s, declining from an average of$41,300 for the 1999 cohort to $32,100 for the 2002 cohort (Chart 1). This decline was related, in part, to the high-tech bust of the period and the fact that a large proportion of PAs in the economic class were in computer science or engineering occupations (Picot and Hou 2009). Earnings recovered to$41,100 for the 2006 cohort and remained at roughly that level up to the 2010 cohort.

Actual average entry earnings for male PAs changed little between the 1999 and 2010 landing cohorts (they increased by $700). The decomposition suggests that changes to all control variables (Model 3) over the 11 years tended to increase entry earnings by$8,900 (Table 5, right column). In particular, the rising share of PAs with pre-landing Canadian work experience accounted for the majority (94%) of this increase associated with changing composition.Note 18 Changing source region also contributed positively to entry earnings (21% of the accounted-for increase). Meanwhile, lower levels of educational attainment and increased settlement in regions or cities with higher unemployment rates by the 2010 landing cohort tended to reduce entry earnings (Table 5, right column).

Pre-landing Canadian work experience also explained the entry earnings advantage that the PN PAs held over their SW counterparts during the 2000s. With all the independent variables (except pre-landing Canadian work experience) in the regression, a male PN PA earned $3,900 more than his SW PA counterpart. When the pre-landing Canadian work experience variable is added, this falls to$0. All the difference in entry earnings between PN PAs and SW PAs is accounted for by differences in pre-landing Canadian work experience.

The results for femalePAs are very similar. Actual entry earnings decreased marginally (by $1,100) during the 2000s, but changes in immigrant characteristics tended to drive up entry earnings by$3,800. And, the increase in the share of immigrants with pre-landing Canadian work experience was responsible in large part for this rise in earnings associated with all factors included in this analysis (Table 5). These factors also explain the contribution that changing immigration class (i.e., the rise of the PNP) made to rising entry earnings (Table 5), and why female PN PAs had higher entry earnings than SW PAs during the 2000s.Note 19

All new immigrants Principal applicants in the economic class Model 1 Model 2 Model 3 Model 1 0 0 0 700 700 700 800 2,100 5,000 3,000 3,900 8,900 Note ...: not applicable 0.01 -0.01 Note ...: not applicable 0.03 0.01 1.00 0.59 -0.05 1.00 0.94 -0.02 Note ...: not applicable 0.00 -0.01 Note ...: not applicable -0.17 -0.06 Note ...: not applicable 0.66 0.23 Note ...: not applicable 0.60 0.21 Note ...: not applicable 0.02 0.03 Note ...: not applicable -0.16 0.01 Note ...: not applicable -0.08 0.02 Note ...: not applicable -0.22 -0.01 Note ...: not applicable -0.19 -0.14 Note ...: not applicable -0.02 -0.07 Note ...: not applicable Note ...: not applicable 0.92 Note ...: not applicable Note ...: not applicable 0.94 1,700 1,700 1,700 -1,100 -1,100 -1,100 1,900 2,800 4,300 1,400 1,000 3,800 Note ...: not applicable 0.00 -0.01 Note ...: not applicable -0.07 -0.02 1.00 0.54 0.08 1.00 1.53 -0.17 Note ...: not applicable 0.25 0.13 Note ...: not applicable 0.58 0.12 Note ...: not applicable 0.20 0.14 Note ...: not applicable 0.02 0.19 Note ...: not applicable 0.13 0.07 Note ...: not applicable -0.49 -0.07 Note ...: not applicable -0.05 -0.01 Note ...: not applicable -0.40 -0.01 Note ...: not applicable -0.07 -0.06 Note ...: not applicable -0.19 -0.10 Note ...: not applicable Note ...: not applicable 0.66 Note ...: not applicable Note ...: not applicable 1.06 ... not applicable Notes: Model 1 adjusts for immigration class; Model 2 adjusts for immigration class plus demographic characteristics and regional unemployment rate; Model 3 adjusts for immigration class plus demographic characteristics and regional unemployment rate as well as pre-landing Canadian work experience. Total change and change caused by changes in characteristics are rounded to the nearest $100. Source: Statistics Canada, Longitudinal Immigration Database. ### 4.4 The effect of changes in immigrant selection on entry earnings during the 1990s The 1990s differed from the 2000s in terms of changes to immigrant selection and immigrant characteristics. The significant changes in the 1990s included large increases in the share of SW immigrants and in educational attainment at landing. Between the approximate business cycle peaks of 1989/1990 (the 1988 landing cohort) and 2000/2001, the actual entry earnings of immigrant men remained constant at$34,500. But changing characteristics tended to increase their entry earnings by $3,500 (Table 6). Rising educational attainment at landing of immigrants was by far the most significant factor putting upward pressure on entry earnings, accounting for 69% of the accounted-for rise in earnings. This factor tended to increase average entry earnings by$2,400 (69% times $3,500). Changing immigration class (a rising share of immigrants in the economic class) and the rising share of immigrants with pre-landing Canadian work experience also tended to increase entry earnings, by$1,100 and $700 respectively. But there were changes that offset part of these gains, notably changing source regions, which tended to reduce entry earnings by$700. Results were similar when log annual earnings were used as the dependent variable, rather than annual earnings.Note 20

Changing characteristics had less effect on the entry earnings ofimmigrant women during the 1990s; they improved average entry earnings by only $1,200, compared with$3,500 for men. The factors influencing entry earnings were the same for women as for men, except for changing immigration class, which had virtually no effect among women. As with men, rising educational attainment at landing was the dominant factor pushing up women’s entry earnings,Note 21 followed by an increase in the share of women with pre-landing Canadian work experience. These increases were partially offset by declining earnings related to changing source regions and changing language (Table 6).

A similar decomposition analysis was conducted for PAs in the economic class (Table 6). Qualitatively, the results were similar. Rising educational attainment at landing was the main factor putting upward pressure on entry earnings, along with the rising share of PAs in the economic class who had pre-landing Canadian work experience. These gains were partially offset by losses related to changing source regions and language and, for women, changes in destination cities or regions (Table 6). Immigration class changes did not play much of a role, since among PAs in the economic class there were only minor changes in the distribution of the business class and federal SWs within this major class during the 1990s.

### 4.5 The effect of changes in immigrant selection on entry earnings during the 1980s

Compositional changes during the 1980s, described earlier, tended to decrease, not increase, entry earnings, particularly among men. Model 3 could not be run for this period, since the data to determine pre-landing Canadian work experience were not available. The results reported here are based on Model 2, which includes all of the explanatory variables except pre-landing Canadian work experience.

There are three major points. First, among men, the changes in source region, language ability, immigration class and age (a proxy for foreign work experience) accounted for the entire decline of $3,600 in entry earnings among men between the 1981 and 1988 landing cohortsNote 22 (Table 6). Not surprisingly, changing source region was the most significant factor, followed by immigration class and language. Second, entry earnings did not decline among new immigrant women over this period; they actually rose by about$2,600 (Table 6). This observation is not generally known, but it is consistent with earlier research (Hou 2013). Using census data, Frenette and Morissette (2005) also observed an increase during the 1980s in women’s entry earnings, defined as average earnings in the first five years after immigration. Third, the included explanatory variables accounted for little—only $500—of the$2,600 increase in women’s entry earnings over the period (Table 6). Other unknown factors were driving up women’s entry earnings in spite of the shift in source regions and language skills, which would have had a negative effect on entry earnings for both men and women.

Between 1981 and 1988 landing cohorts Between 1988 and 1999 landing cohorts All new immigrants Principal applicants in the economic class -3,600 -4,700 -200 -1,700 -3,700 -4,800 3,500 3,500 -0.14 -0.40 0.03 0.03 0.41 0.17 0.33 0.09 -0.01 -0.14 0.69 1.21 0.88 1.24 -0.21 -0.44 0.37 0.43 -0.10 -0.26 -0.16 -0.13 -0.02 0.02 -0.36 -0.17 0.06 0.10 Note ...: not applicable Note ...: not applicable 0.22 0.24 2,600 4,800 -1,500 -1,000 500 2,400 1,200 2,700 0.32 0.24 0.02 0.00 0.57 -0.01 -0.09 -0.07 0.79 0.48 1.13 1.10 -0.64 0.08 -0.24 -0.07 -1.01 0.10 -0.13 -0.12 1.77 0.79 -0.10 -0.19 -0.81 -0.68 0.05 0.00 Note ...: not applicable Note ...: not applicable 0.35 0.35 ... not applicable Note: Total change and changes caused by changes in characteristics are rounded to the nearest \$100. Source: Statistics Canada, Longitudinal Immigration Database.

## 5 Conclusion and discussion

Among new immigrants, abstracting from economic cyclical variation, earnings immediately after landing remained more or less constant throughout the 1990s and 2000s. There were some minor improvements in the late 2000s. The trend for landing principal applicants (PAs) in the economic class was very similar. However, despite this apparent stability, during the 2000s in particular, there were significant pressures on earnings—primarily positive—related to changing immigrant selection and characteristics. During the 1990s, rising educational attainment at landing and the increasing share of immigrants in the economic class increased entry earnings. During the 2000s, a much more complex period in terms of immigrant selection, the factors that positively influenced immigrant entry earnings included changes in the distribution by immigration class, notably the rise of the Provincial Nominee Program (PNP); changing source region; and, for immigrant women, rising educational attainment at landing. The PNP effect was primarily reflecting a rise in the share of immigrants with pre-landing Canadian work experience. These positive effects were offset by some negative pressures related to less favourable economic conditions in destination cities and regions in 2011/2012 than in 2000/2001.

The following discussion might assist in interpreting these earnings pressures. In this analysis, the effect of any characteristic on entry earnings—say, a rising share of immigrants with a bachelor’s degree—assumes that the earnings value of a bachelor’s degreeNote 23 for a new immigrant remained constant at the average value observed over a given study period (e.g., the 2000s). Increasing the share of immigrants with a degree would put upward pressure on entry earnings, since a degree holder has higher earnings than, say, a new immigrant with a high-school education.

Despite the overall upward pressure that changing characteristics put on entry earnings (holding the earnings value of any characteristic constant), entry earnings did not in fact rise over the two decades. That means that one of two things may have been happening. There may have been downward pressure on entry earnings in general for reasons not included in this research, such as an increase in the supply of immigrants (Hou and Picot 2014), a general deterioration in earnings in labour-market entry jobs (Green and Worswick 2010) and other possible reasons (Reitz 2007). Or, more specific events, rather than general events, may have been occurring. Notably, the earnings potential associated with some specific characteristics (such as a university degree, a particular source region, pre-landing Canadian work experience, or particular language skills) may have declined. Entry earnings would have risen had one of these possibilities not occurred. Assessing what exactly was occurring is left for future research.

One factor was more important than all others during the 2000s. The rise in the share of new immigrants who had pre-landing Canadian work experience, often in high-paying jobs, tended to put far more upward pressure on entry earnings than any other variable studied. The increase in pre-landing Canadian work experience accounted for most of the positive effect of the rise of the PNP on entry earnings during the 2000s, since it was heavily concentrated among provincial nominees. Furthermore, differences in pre-landing Canadian work experience between provincial nominees (with more Canadian work experience) and SWs (with less) accounted for virtually all of the entry earnings advantage that the provincial nominees held over SW immigrants during the 2000s. While other factors, such as differences in geographic distribution (more in the West), educational attainment at landing, unemployment in the destination regions and cities, and source region, contributed, either in a small positive or negative manner, to the entry earnings differences between provincial nominees and SWs, their contribution paled in comparison with the pre-landing Canadian work experience factor. Once differences in pre-landing Canadian work experience were taken into account, entry earnings were virtually identical between provincial nominees and SWs. These conclusions apply to all new immigrants, as well as to PAs in the economic class, and are evident for both men and women.

It is likely that the pre-landing Canadian work experience variable used here captures at least three effects. First is the possible effect of pre-landing Canadian work experience on earnings early in immigrants’ working life after landing. Employers may be willing to remunerate such experience, in contrast to foreign work experience, which receives little remuneration. But all, or perhaps most, of the effect associated with this variable is not likely related to the remuneration of Canadian work experience. Second, this variable may also reflect a selection effect (Hao 2013; Hou and Bonikowska 2015; Sweetman and Warman 2014). When immigrants are selected from the pool of temporary foreign workers, they come with information regarding how well they performed in their jobs in Canada. If an employer seeks to change the status of temporary foreign workers to a permanent one, it is likely because they have done well in their jobs. Hence, much of the effect on entry earnings could be because of this selection process. Third, during the 2000s, many of the workers on temporary visas who attained permanent status worked in high-paying jobs. If one increases the share of PAs who are in high-paying jobs, then, naturally, average entry earnings will rise. Whether this trend will continue in the future is unknown. If the PAs whose status changes from temporary foreign worker to permanent resident are in increasingly lower-paying, lower-skilled jobs, then entry earnings may fall, despite the PAs having pre-landing Canadian work experience.

## 6 Appendix tables and charts

Description for Appendix Chart 1
Canadian-born men and established immigrant men New immigrant men 93.6 93.5 93.2 91.7 93.5 90.9 93.3 90.9 92.3 90.9 93.1 92.2 93.2 92 92.9 90.1 91.4 88 90 82.7 89.1 79.2 87.9 74.5 87.6 75 87.6 73.5 87.5 73.4 88 74.1 88.2 76 88.6 79.6 88.5 81.3 90.1 82 89.7 80.7 89.8 80.3 89.8 82.1 90.4 83.5 89.9 83.1 90 84.3 89.8 85 88.9 83 88.5 83 88.8 83.8 Source(s): Statistics Canada, Longitudinal Immigration Database.

Description for Appendix Chart 2
Canadian-born women and established immigrant women New immigrant women 67.8 69.6 68.4 69.4 70 71 72.3 74 73 75.5 74.8 77.3 76.4 78.4 77.3 77.3 78.5 75 76.7 68.9 76.3 65.1 74.8 58.8 74.8 58.7 75.4 55.8 75.7 53.8 76.8 53.6 77.6 55.2 78.4 59.4 79.6 62.1 81.8 62.9 81.7 62 82.3 61.6 82.7 62.8 83 64.2 83.2 65.4 83.5 66.7 83.6 67.1 83 64.8 82.5 64.3 82.6 65.3 Source(s): Statistics Canada, Longitudinal Immigration Database.
Annual earnings (in 2011 constant dollars) as the outcome Log annual earnings as the outcome Model 1 Model 3 Model 1 Model 3 estimate standard error 39,349 105 58,418 319 10.098 0.003 10.423 0.011 815 152 5,089 135 0.008 0.005 0.076 0.005 -1,274 144 3,975 124 -0.050 0.005 0.064 0.004 -4,696 141 650 122 -0.163 0.004 -0.037 0.004 -5,240 146 -409 126 -0.178 0.005 -0.065 0.004 -4,386 147 22 127 -0.150 0.005 -0.048 0.004 -1,510 144 953 127 -0.078 0.005 -0.035 0.004 -1,420 139 1,245 125 -0.048 0.004 -0.014 0.004 601 141 2,099 126 0.013 0.004 0.022 0.004 1,059 143 2,646 122 0.003 0.005 0.052 0.004 109 142 1,496 127 -0.042 0.004 0.035 0.004 -691 140 -384 120 -0.041 0.004 -0.010 0.004 12,082 148 27 150 0.396 0.005 -0.006 0.005 -5,762 579 -6,654 503 0.184 0.018 -0.222 0.017 -16,870 213 -5,500 192 -0.528 0.007 -0.164 0.007 -12,787 97 -5,774 89 -0.317 0.003 -0.213 0.003 -9,836 79 -5,287 82 -0.118 0.002 -0.089 0.003 -14,574 92 -9,606 98 -0.358 0.003 -0.306 0.003 -10,159 870 -15,436 744 -0.203 0.027 -0.403 0.025 Note ...: not applicable Note ...: not applicable 682 155 Note ...: not applicable Note ...: not applicable 0.126 0.005 Note ...: not applicable Note ...: not applicable 2,100 144 Note ...: not applicable Note ...: not applicable 0.276 0.005 Note ...: not applicable Note ...: not applicable 2,204 143 Note ...: not applicable Note ...: not applicable 0.261 0.005 Note ...: not applicable Note ...: not applicable 2,112 146 Note ...: not applicable Note ...: not applicable 0.214 0.005 Note ...: not applicable Note ...: not applicable 1,955 151 Note ...: not applicable Note ...: not applicable 0.167 0.005 Note ...: not applicable Note ...: not applicable 1,326 161 Note ...: not applicable Note ...: not applicable 0.100 0.006 Note ...: not applicable Note ...: not applicable -8,339 126 Note ...: not applicable Note ...: not applicable -0.097 0.004 Note ...: not applicable Note ...: not applicable -9,115 99 Note ...: not applicable Note ...: not applicable -0.124 0.003 Note ...: not applicable Note ...: not applicable -8,897 95 Note ...: not applicable Note ...: not applicable -0.090 0.003 Note ...: not applicable Note ...: not applicable -3,783 79 Note ...: not applicable Note ...: not applicable -0.033 0.003 Note ...: not applicable Note ...: not applicable -403 220 Note ...: not applicable Note ...: not applicable 0.134 0.008 Note ...: not applicable Note ...: not applicable -12,029 230 Note ...: not applicable Note ...: not applicable -0.091 0.008 Note ...: not applicable Note ...: not applicable -12,112 223 Note ...: not applicable Note ...: not applicable -0.235 0.008 Note ...: not applicable Note ...: not applicable -19,802 231 Note ...: not applicable Note ...: not applicable -0.468 0.008 Note ...: not applicable Note ...: not applicable -14,462 223 Note ...: not applicable Note ...: not applicable -0.157 0.008 Note ...: not applicable Note ...: not applicable -11,100 234 Note ...: not applicable Note ...: not applicable 0.080 0.008 Note ...: not applicable Note ...: not applicable -16,763 237 Note ...: not applicable Note ...: not applicable -0.392 0.008 Note ...: not applicable Note ...: not applicable -11,004 219 Note ...: not applicable Note ...: not applicable -0.093 0.007 Note ...: not applicable Note ...: not applicable -4,751 308 Note ...: not applicable Note ...: not applicable 0.075 0.011 ... not applicable Note: Model 3 also controls for geographic distribution across provinces and major metropolitan areas. Source: Statistics Canada, Longitudinal Immigration Database.
Annual earnings (in 2011 constant dollars) as the outcome Log annual earnings as the outcome Model 1 Model 3 Model 1 Model 3 estimate standard error Note ...: not applicable Note ...: not applicable -10,048 269 Note ...: not applicable Note ...: not applicable -0.098 0.009 Note ...: not applicable Note ...: not applicable -4,637 189 Note ...: not applicable Note ...: not applicable 0.040 0.006 Note ...: not applicable Note ...: not applicable -6,408 118 Note ...: not applicable Note ...: not applicable -0.148 0.004 Note ...: not applicable Note ...: not applicable -9,786 186 Note ...: not applicable Note ...: not applicable -0.296 0.006 Note ...: not applicable Note ...: not applicable -7,164 150 Note ...: not applicable Note ...: not applicable -0.157 0.005 Note ...: not applicable Note ...: not applicable -8,114 133 Note ...: not applicable Note ...: not applicable -0.245 0.005 Note ...: not applicable Note ...: not applicable -776 31 Note ...: not applicable Note ...: not applicable -0.045 0.001 Note ...: not applicable Note ...: not applicable 71,759 135 Note ...: not applicable Note ...: not applicable 1.224 0.005 Note ...: not applicable Note ...: not applicable 9,922 102 Note ...: not applicable Note ...: not applicable 0.428 0.003 Note ...: not applicable Note ...: not applicable 805 101 Note ...: not applicable Note ...: not applicable 0.037 0.003 ... not applicable Note: Model 3 also controls for geographic distribution across provinces and major metropolitan areas. Source: Statistics Canada, Longitudinal Immigration Database.
All new immigrants Principal applicants in the economic class Controlling for immigration class only Controlling for all covariates 0.022 0.022 0.033 0.033 0.030 0.098 0.109 0.163 Note ...: not applicable -0.07 Note ...: not applicable -0.05 1.00 -0.15 1.00 0.00 Note ...: not applicable 0.00 Note ...: not applicable -0.03 Note ...: not applicable 0.54 Note ...: not applicable 0.43 Note ...: not applicable 0.13 Note ...: not applicable 0.05 Note ...: not applicable -0.05 Note ...: not applicable -0.16 Note ...: not applicable -0.41 Note ...: not applicable -0.22 Note ...: not applicable 1.01 Note ...: not applicable 0.98 0.094 0.094 -0.037 -0.037 0.097 0.189 0.093 0.137 Note ...: not applicable -0.02 Note ...: not applicable -0.08 1.00 0.09 1.00 -0.20 Note ...: not applicable 0.08 Note ...: not applicable 0.09 Note ...: not applicable 0.26 Note ...: not applicable 0.52 Note ...: not applicable 0.13 Note ...: not applicable -0.05 Note ...: not applicable -0.02 Note ...: not applicable -0.02 Note ...: not applicable -0.12 Note ...: not applicable -0.17 Note ...: not applicable 0.59 Note ...: not applicable 0.91 ... not applicable Source: Statistics Canada, Longitudinal Immigration Database.
Betweeen 1981 and 1988 landing cohorts Between 1988 and 1999 landing cohorts All new immigrants Principal applicants in the economic class -0.061 -0.090 -0.134 -0.199 -0.052 -0.101 0.091 0.073 -0.27 -0.31 0.06 0.10 0.98 0.23 0.56 0.16 0.02 -0.10 0.49 1.11 1.33 1.06 -0.31 -0.45 0.80 0.55 -0.09 -0.32 -0.39 -0.21 -0.03 0.03 -1.46 -0.23 0.10 0.20 Note ...: not applicable Note ...: not applicable 0.22 0.18 0.179 0.271 -0.239 -0.228 0.067 0.165 0.032 0.072 0.08 0.09 0.03 -0.04 0.23 -0.01 0.00 -0.10 0.18 0.23 1.25 1.25 -0.01 0.18 -0.49 -0.06 -0.46 0.15 -0.30 -0.18 0.75 0.57 -0.21 -0.38 0.24 -0.20 0.15 0.01 Note ...: not applicable Note ...: not applicable 0.58 0.50 ... not applicable Source: Statistics Canada, Longitudinal Immigration Database.

## References

Aydemir, A., and M. Skuterud. 2005. “Explaining the deteriorating entry earnings of Canada’s immigrant cohorts: 1966–2000.” Canadian Journal of Economics 38 (2): 641–672.

Beach, C., A. Green, and C. Worswick. 2011. Toward Improving Canada’s Skilled Immigration Policy: An Evaluation Approach. C.D. Howe Institute Policy Studies.

Bonikowska, A., F. Hou, and G. Picot. 2015. Which Human Capital Characteristics Best Predict the Earnings of Economic Immigrants? Analytical Studies Branch Research Paper Series, no. 368. Statistics Canada Catalogue no. 11F0019M. Ottawa: Statistics Canada.

Citizenship and Immigration Canada. 2010. Evaluation of the Federal Skilled Worker Program. Report no. Ci4-54/2010E-PDF. Available at: http://www.cic.gc.ca/english/pdf/research-stats/FSW2010.pdf.

Ferrer, A., G. Picot, and C. Riddell. 2014. “New directions in immigration policy: Canada’s evolving approach to the selection of economic immigrants.” The International Migration Review 48 (3): 846–867.

Frenette, M., and R. Morissette. 2005. "Will they ever converge? Earnings of immigrant and Canadian-born workers over the last two decades." The International Migration Review 39 (1): 228–257.

Green, D., and C. Worswick. 2010. “Entry earnings of immigrant men in Canada: The roles of labour market entry effects and returns to foreign experience.” In Canadian Immigration: Economic Evidence for a Dynamic Policy Environment, ed. T. McDonald, E. Ruddick, A. Sweetman, and C. Worswick, p. 77–110. Montréal: Queen’s Policy Studies Series.

Hao, L. 2013. “Admission-group salary differentials in the United States: The significance of the labour-market institutional selection of high-skilled workers.” Journal of Ethnic and Migration Studies 39 (8): 1337–1360.

Hou, F. 2013. “Immigrant entry earnings over the past quarter century: The roles of changing characteristics and returns to skills.” Canadian Studies in Population 40 (3-4): 149–163.

Hou, F. 2014. “A general approach to effect decomposition.” Social Science Quarterly 95 (3): 894–904.

Hou, F., and A. Bonikowska. 2015. The Earnings Advantage of Landed Immigrants Who Were Previously Temporary Residents in Canada. Analytical Studies Branch Research Paper Series, no. 370. Statistics Canada Catalogue no. 11F0019M. Ottawa: Statistics Canada.

Hou, F., and G. Picot. 2014. “Annual levels of immigration and immigrant entry earnings in Canada.” Canadian Public Policy 40 (2): 166–181. doi:10.3138/2013_017.

Pandey, M., and J. Townsend. 2013. “Provincial nominee programs: An evaluation of earnings and settlement rates of nominees.” Canadian Public Policy 39 (4): 603–618.

Picot, G., and F. Hou. 2009. Immigrant Characteristics, the IT Bust, and Their Effect on Entry Earnings of Immigrants. Analytical Studies Branch Research Paper Series, no. 315. Statistics Canada Catalogue no. 11F0019M. Ottawa: Statistics Canada.

Picot, G. and A. Sweetman. 2012. Making it in Canada: Immigration outcomes and policies. IRPP Study, vol. no. 29. Montréal: Institute for Research on Public Policy.

Reitz, J. 2007. “Immigrant employment success in Canada, Part II: Understanding the decline.” Journal of International Migration and Integration 8 (1): 37–62.

Sweetman, A., and C. Warman. 2014. “Former temporary foreign workers and international students as sources of permanent immigration.” Canadian Public Policy 40 (4): 392–407.

Is something not working? Is there information outdated? Can't find what you're looking for?

Please contact us and let us know how we can help you.

Privacy notice

Date modified: