The deteriorating economic welfare of Canadian immigrants.

Picot, Garnett

Abstract

Both the low-income and the earnings gap between immigrants and the Canadian born are observed to have increased, particularly for recent immigrants. While the incidence of low income fell among the Canadian born during the 1990s, it increased among immigrants in all education and age groups, and from most "non-traditional" source countries. In Canada's major cities virtually all the increase in low-income rates during the 1990s was concentrated among the immigrant population.

The deterioration in immigrant economic welfare occurred despite policy changes designed to select individuals with greater potential contribution to the Canadian economy. We find that changes in immigrant characteristics accounted for roughly one-third of the increase in the gap. Decreasing economic returns to foreign work experience and education also played a large role.

Keywords: Immigrants, poverty, earnings, integration, immigration.

Resume

Cet article analyse la recente augmentation du taux de pauvrete et l'ecart salarial entre les immigrants et les Canadiens nes pendant les deux dernieres decennies. Alors que le taux de pauvrete a diminue parmi les Canadiens nes au cours des annees 1990, il a toutefois augmente parmi les differents groupes et categories d'immigrants (i.e. niveau d'education, tranches d'age etc.) et parmi les immigrants des pays source non traditionnels. L'on constate egalement que l'augmentation du taux de pauvrete dans les villes principales du Canada pendant les annees 1990 est concentree parmi la population d'immigrants.

Ainsi l'on constate une deterioration des conditions economiques des immigrants malgre les changements de la politique d'immigration au Canada basee sur la selection d'immigrant en fonction de leur contribution potentielle au developpement economique. D'une part, nous constatons que ces changements (caracteristiques de selection d'immigrants) representent environ un tiers de l'augmentation de l'ecart salarial et du taux de bas salarie. D'autre part, la diminution de retours economiques de l'experience de travail a l'etranger et le niveau d'education a egalement contribue a la deterioration des conditions economiques des immigrants

Mots cles : immigrants, pauvrete, remuneration, integration au travail

**********

Host countries, such as Canada, look to the skills and initiative of immigrants to promote economic growth. In the "knowledge-based" economy, host countries are increasingly seeking highly educated workers to drive economic growth. Immigrants, particularly the highly educated, look to the host country for opportunities to use their skills and abilities to achieve high economic standards of living. However, if immigrants are unable to convert their training to productive use, the expectations of both the host country and the arriving immigrants will not be met. Immigrant contributions to the host country, which are central to the economic justification for relatively open immigration policies, may not be fully realized. In light of these considerations, there is significant concern regarding the deteriorating economic outcomes among recent immigrants to Canada over the past two decades.

Immigrants arriving in Canada during the late 1990s were highly educated. In 2001, fully 42 percent of "recent" immigrants (those who arrived in Canada within the last five years) had a university degree, and an historically high 54 percent of these immigrants entered Canada under the "economic" admissions class. Only 31 percent were in the family class. The situation was very different twenty years earlier. In 1981, only 19 percent of recent immigrants had degrees. During the early 1980s, 37 percent of immigrants entered Canada under the "economic" class and 43 percent entered in the family class. Immigrants of the late 1990s were increasingly selected for their potential contributions to the Canadian economy.

Analysis of the employment earnings of immigrants is the most studied area relating to the economic integration of Canadian immigrants. Early findings indicated that newly arrived immigrants had lower earnings than comparable Canadian-born workers, but that this initial earnings gap narrowed significantly as immigrants adjusted to labour market conditions in the receiving society (Chiswick, 1978; Meng, 1987). Subsequent research identified an emerging trend during the early 1980s of declining earnings among successive waves of immigrants, relative to the Canadian born (Bloom and Gunderson, 1991; Abbott and Beach, 1993), and indicated that this trend continued unabated during the late 1980s (McDonald and Worswick, 1998; Baker and Benjamin, 1994; Grant 1999) and through the early 1990s (Reitz, 2001), with relatively little improvement during the late 1990s (Green and Worswick, 2002; Frenette and Morissette, 2003; Warman and Worswick, this issue). Recent studies also suggest that the gap in initial earnings increased in the 1980s and 1990s from the 1970s, and that the gap in initial earnings may not close as quickly as previously thought, even for immigrants who entered Canada during the 1970s (Hum and Simpson, 2003). Some studies note that, although there has been a large decline in entry-level earnings, the rate of growth in earnings with years in Canada is faster than among earlier cohorts (Li, 2003). Researchers concerned with the declining economic integration of immigrants have asked if this decline in the economic welfare of immigrants is associated primarily with recessions or the changing mix of immigrants by source country.

As important as studies of employment earnings are, they provide only part of the picture since they exclude the unemployed and those not in the labour market. They also exclude changes in other sources of income, such as social transfer benefits. During periods of deteriorating labour market outcomes, such studies may underestimate the decline in economic welfare among immigrants because they exclude effects related to rising unemployment or discouraged workers.

This paper first reviews the evidence relating to the decline in earnings among immigrants through the 1980s and 1990s. Then, to provide a better overview of economic welfare outcomes among immigrants, the paper considers low-income trends among recent immigrants in particular. The low-income rate is a simple, yet comprehensive measure of changing economic welfare among families at the bottom end of the income distribution. It reflects changes in unemployment, earnings, social transfers and other income sources. (1) Finally, the paper discusses potential major causes for this phenomenon, as reported in recent research studies

This paper focuses on the late 1990s, a particularly important period. With the expansion of the "economic" class as a proportion of "recent" immigrants and the very significant economic recovery in the late 1990s (as unemployment fell from 9.4 percent to 6.8 percent between 1995 and 2000), one might anticipate a marked improvement in immigrant economic outcomes. If the deterioration in the economic status of immigrants was simply related to the recessions of the early 1980s and 1990s, then there may well have been a return to better economic outcomes for immigrants by the year 2000. In this sense, the late 1990s represents a sort of watershed. This paper examines outcomes for highly educated recent immigrants, as one might expect this group to convert their training to productive use during a period of rapid economic growth, with the expansion in the information, communications and technology sector and the general expansion in the knowledge-based economy. (2) This summary paper draws on the findings from two papers in particular, Picot and Hou (2003) and Frenette and Morissette (2003).

The Declining Entry Level Earnings Among Immigrants

In spite of the increasingly economic nature of immigration during the late 1990s, earning among male adult "recent" immigrants who worked full-time full-year fell by 13% between 1980 and 2000, two years that are in roughly the same position in the business cycle (Frenette and Morissette, 2003). Among recent female immigrants earnings rose 6%. During the same period earnings among Canadian-born working full-time full years rose by 10% for men, and 11% for women.

These numbers can be put in a longitudinal perspective by focusing on particular cohorts of immigrants. These cohorts consist of immigrants who entered Canada during a five year period, say between 1975 and 1979, or 1990 and 1994. The earnings of the immigrants in each cohort are computed (as a proportion of the earnings of the Canadian-born) for the cohort after one to five years in Canada, six to ten years in Canada, and so on. In this way one can determine (a) the extent to which the earnings gap at entry between immigrants and Canadian-born is reduced with years spent in Canada, and (b) whether the earnings gap at entry is increasing.

Chart 1 suggests that the earnings gap for the cohort of the late 1970s was more than closed after twenty years in Canada. After 21-25 years adult immigrants working full-time full-year earned about 8% more than the Canadian born (both men and women). However, the earnings gap at entry has been increasing with successive cohorts. Among men, the cohort of the late 1970s earned 90% of that of the Canadian-born at entry (1) and among the early 1990s cohort, this number had fallen to 67%. Among the late 1990s cohort, however, the entry earnings gap appears to have been reduced significantly (compared to the early 90s cohort, for example), as recent immigrants earned 77% of that of the Canadian-born. Still, the gap remained much greater than in the 1970s.

However, these charts compare all Canadian-born with all immigrants. We know that these two groups are very different regarding characteristics that affect their earnings potential. In particular, immigrants have been more highly educated than Canadian-born for many years, and their education level has increased rapidly, particularly during the late 1990s. A more appropriate analysis compares recent immigrants to "like" Canadian-born, along dimensions such as age, education, visible minority status, marital status and region of employment (including major city). This is typically done within a regression format that computes the log of the ratio of the earnings of immigrants to those of the Canadian-born. The results of such an analysis by Frenette and Morissette (2003) are given in Chart 2. The log earnings ratio is an approximation of the earnings of immigrants as a proportion of those of comparable Canadian-born when the differences are small (say 10% to 20%), but when they become large (say 40% to 50%) the log (earnings ratio) overestimates the percentage difference.

This analysis provides a somewhat different picture (panel A of Chart 2). As in Chart 1, the earnings gap has been increasing significantly with each successive cohort, both at entry and after many years in Canada. However, this increase is greater after accounting for the differences between immigrants and Canadian born. Among males, the log earnings ratio at entry declined from .83 among the late 70s cohort to .55 among the early 90s cohort. Furthermore, in this analysis, one sees only a minor improvement in the earnings gap at entry between the cohorts of the early and late 1990s (log earnings ratio increased from .55 to .60). The relatively little improvement seen here compared to Chart 1 is likely because the late 90s cohort was much more highly educated. After accounting for this higher education level and the associated higher earnings expectation, little improvement in entry level earnings is observed. The earnings gap at entry remained much greater than among the cohorts of the 1970s or 1980s.

As noted earlier, the traditional "economic integration" story (sometimes called "economic assimilation") among immigrants is that they earn less at entry, but after a number of years in Canada the earnings gap with comparable Canadians narrows, or disappears (Carliner, 1981; Chiswick, 1978; Borjas, 1985; Meng, 1987). In Chart 2 the traditional pattern is observed among cohorts entering in the late 1970s, but this is the last cohort to display any signs of eliminating the earnings gap after many years in Canada. Even among this cohort, earnings do not surpass the earnings of the Canadian-born (as they did in in Chart 1), but they do almost catch-up, reaching about 97% of the earnings of "like" Canadian born after more than 20 years (Chart 2). But among the more recent cohorts, elimination of the earnings gap will be more difficult. The cohorts entering during the 1980s, even after 16 to 20 years in Canada were earning approximately 85% of their Canadian counterparts (Chart 2). For the 1990s cohort, after 6 to 10 years in Canada, the early 1990s cohort was earning roughly 70% (log earnings ratio of .7) of Canadians. It is not clear that they can "catch-up" to Canadians in 20 years. Again, the pattern is similar for women.

However, it is important to note that the greater the earnings gap at entry, the faster the improvement in earnings (Chart 2). Hence, the rate of improvement in earnings (i.e. the slope of the line in Chart 2) was much greater among the 1990s cohorts than the 1980s cohorts. In spite of this, however, earnings remained lower among the more recent cohorts than among earlier cohorts, no matter how long they had been in Canada.

Earnings Declines at Entry were as Great Among the University Educated

The bottom panels of Charts 1 and 2 display similar results for immigrants with bachelors degrees. Among men working full-time full-year, immigrants entering during the late 70s earned about 82% of their Canadian born counterparts (log earnings ratio of .82, controlling for estimated years of work experience, and geographical region of employment). By the 1990s, the log earnings ratio had fallen to around .5. Some minor improvement was observed between the early and late 1990s, but relative entry level earnings remained well below that of the cohorts of the 1970s or 1980s. Once again, the pattern is similar for women. "Recent" immigrant university graduates were increasingly unable to convert their education and experience into earnings in the way that earlier cohorts had. As with all immigrants, however, the larger the earnings gap at entry, the faster the growth in earnings (with years spent in Canada).

The Deteriorating Low-Income Position of Recent Immigrants

Despite the increasing proportion of immigrants who enter Canada in the "economic" class, average earnings among new immigrants have declined over the last two decades. Reviews of several studies, for example that by Hum and Simpson in this issue, confirm that this trend is widely observed. Moreover, a similar decline was observed for both men and women with university degrees relative to the Canadian born. While recent entry cohorts with lower-entry earnings appear to have somewhat higher rates of earnings increases over time, there is no evidence to suggest that they will catch up to either the Canadian born or immigrant cohorts of earlier decades.

Earnings represent the major income source for most families. Therefore, it is not surprising that the decline in relative entry-level earnings for immigrants is reflected in the low-income rate. The proportion of recent immigrants with family incomes below the low-income cutoff rose from 24.6 percent in 1980, to 31.3 percent in 1990, and to 35.8 percent in 2000. (3) In contrast, the low-income rate among the Canadian born fell from 17.2 percent in 1980 to 14.3 percent in 2000. As these years are roughly at business cycle peaks, they are reasonable indicators of longer-term trends. These statistics reveal that it was not a general deterioration in economic conditions affecting all Canadians that was responsible for the rising low-income rates during this period. Recent immigrants had low-income rates 1.4 times that of the Canadian born in 1980, and 2.5 times higher by 2000. There was improvement between 1995 and 2000 in the low-income rate among recent immigrants (falling from 47.0 percent to 35.8 percent). However, most of this change did not represent a longer-term trend in improved outcomes, but was associated with the very strong economic recovery portion of that business cycle. The low-income rate of recent immigrants, relative to the Canadian born, improved only marginally (falling from 2.7 to 2.5 times higher over this five-year period). This deterioration in low-income rates over the past 20 years was not restricted to recent immigrants but was observed across all immigrant groups, no matter how long they had resided in Canada (Table 1). The only exception is immigrants who had lived in Canada for more than 20 years.

While it may have been possible that the deterioration in family economic welfare among recent immigrants was concentrated among particular groups, Picot and Hou (2003) concluded that this was by and large not the case. The increasing low-income rate was observed among all education groups, all age groups, all family types, as well as those whose home language was French, English and/or "other." The only dimension along which some differences were observed was source regions. The low-income rate actually fell between 1980 and 2000 among recent immigrants from the source regions of Western Europe, Southeast Asia, the United States, and the Caribbean (even after controlling for differences in education, home language, family type, age across cohorts). However, low-income rates were rising among the source regions from which three-quarters of all recent immigrants originated in 2001. This is illustrated in Table 2. Low-income rates were rising for recent immigrants from most parts of Asia, Northern, Eastern and Southern Europe as well as Africa. These results suggest that some factor (e.g., recognition of work experience or schooling, language skills, school quality, etc.) associated with source regions may play a significant role in the declining economic outcomes of immigrants. We will return to this issue later.

Given the widespread nature of the rise in low-income, Picot and Hou (2003) conclude that less than one-half, and quite possibly much less than one-half, of the rise in the low-income rate among recent immigrants is associated with their changing characteristics. They could not be more precise because of limitations of the analytical techniques that exist to determine this effect.

Just as the earnings results reported above focus on university graduates, so too did the low-income analysis. Although recent immigrants with a university degree had lower levels of low-income than those with a high school diploma, the rate of increase in low-income during the 1990s was higher among university-educated immigrants. Among recent adult immigrants, with high school graduation as the highest education level, the low-income rate rose by 25 percent. Among recent university-educated immigrants, it rose from 19.1 percent in 1990 to 27.5 percent in 2000, representing an increase of 44 percent (see Table 3). Furthermore, there seems to be no major difference by degree type. Among engineering and applied science degree holders aged 24 to 44 years, low-income rates rose from 14.7 percent in 1990 to 24.2 percent in 2000. This is in spite of the technology boom that was occurring over the decade. In 1990, recent immigrant degree holders had low-income rates 4.6 times than their Canadian-born counterparts; by 2000 immigrant degree holders had low-income rates 7.0 times higher. Similar patterns are observed for other disciplines. For example, among recent prime-age immigrants with teaching, social sciences and commerce degrees, low-income rates rose from 18.2 percent in 1990 to 27.7 percent in 2000.

Similar patterns are observed regarding low-income rates. Chart 3 shows the predicted low-income rate at entry for cohorts entering between the early 1970s and late 1990s, as well as the change in this rate with years since immigration (based on a regression model that accounts for differences among cohorts in education, age, marital status, home language and family type). Rates do fall with years since immigration, and perhaps more importantly, the higher the low-income rate at entry, the faster the decline (improvement) in the rates. Hence, there are signs that there is some "catch-up" among the 1980s and 1990s cohorts with earlier cohorts regarding low-income rates. However, in no way is the gap between these groups erased. The low-income rate among immigrants entering during the 1980s and early 1990s remains well above that of earlier immigrants, no matter how long they have been in Canada.

Rising Low-Income in the Largest Cities is Heavily Concentrated Among Immigrants

Overall, a dichotomy appears to have developed, with low-income rates rising among immigrants and declining among the Canadian born. The low-income rate among the Canadian born fell from 17.2 percent to 14.3 percent between 1980 and 2000, and low-income rates for groups traditionally associated with low-income also fell. During this period, for example, the low-income rate among Canadian-born seniors fell by 12.5 percentage points; among lone-parent families by 16 percentage points; and it remained virtually unchanged among Canadian-born young families. In contrast, as previously noted, the low-income rate rose among immigrants, except among those resident in Canada for more than 20 years.

Combined with the fact that immigrants, particularly recent immigrants, constitute a rising share of the population in Canada's major cities, this evidence suggests that any change in the low-income rate in these cities may be heavily concentrated among immigrants. Picot and Hou (2003) confirm that this is the case. In Canada's three largest cities, where the vast majority of recent immigrants have settled, the low-income rate increased between 1990 and 2000. It rose 1.9 percentage points in Toronto; 4.1 points in Vancouver; and 0.3 points in Montreal. This was in spite of the fact that economic conditions in these cities were similar or much stronger in 2000, compared to 1990. The unemployment rate declined from 10.5 to 7.7 percent in Montreal, from 7.5 to 5.9 percent in Vancouver, and rose marginally from 5.2 to 5.5 percent in Toronto.

In this environment, the increase in the low-income rate was concentrated among the immigrant population (see Table 4). Two factors contribute to the upward pressure on the low-income rate in cities. First, immigrants, particularly recent immigrants, constitute an increasing share of the population and they have traditionally had higher low-income rates than the population in general. The second factor is the rise in the low-income rate among immigrants. Combined, these factors contributed to an increase in Toronto's low-income rate by 2.8 percentage points between 1990 and 2000 (Table 4). Most of this increase was associated with immigrants who have been in Canada ten years or less. This tendency was offset by a decline in the low-income rate for the Canadian born, which reduced the city's low-income rate by 0.9 percentage points. The end result was an increase in Toronto's low-income rate of 1.9 percentage points. A similar story holds for the other two cities. In Vancouver the changing low-income position among immigrants (and their increasing population share) tended to push up that city's low-income rate by 4.7 percentage points. This was partially offset by a decline in the low-income rate among the Canadian-born population, which reduced the rate by 1.7 points, resulting in a rise of 3.1 points for the city overall.

For Canada as a whole, there was little change in the low-income rate during the 1990s. This resulted from two offsetting trends. Changes among the Canadian born tended to push down the low-income rate by 1.0 percentage points, while changes among immigrants tended to increase it by 1.0 points. However, one should recall that we are discussing changes in low-income rates. If one concentrates on the share of the total low income population, immigrants still hold a minority, although increasing, position. In 2000, immigrants constituted 28.5 percent of the low-income population. This is up from 20 percent from 1980 and much higher than the immigrant share of the total population.

Why the Decline in Relative Entry-Level Earnings and the Rise in Low-Income Among Recent Immigrants?

Researchers have proposed a number of possible explanations for the cause of the decline in entry-level earnings in particular. This section provides a summary review of the more prominent potential explanations. A short outline, however, cannot do justice to many of these possible explanations, and interested readers are advised to read the papers referenced for fuller explanations.

1) The Changing Characteristics of Immigrants Entering Canada

It is well known that immigrants are entering Canada from very different countries now than in previous decades. Between 1981 and 2001, for example, the proportion of immigrants that came to Canada from the United States, Northern Europe, Southern Europe, the Caribbean, South and Central America, and Southeast Asia fell from 65 to 28 percent. In contrast, the proportion from such regions as Eastern Europe, South Asia (India and Pakistan), East Asia (China, Korea and Japan), Western Asia (Iraq, Iran and Afghanistan) and Africa increased from 35 percent in 1981 to 72 percent in 2001. Immigrants from these source regions may have lower earnings on entry, even if they arrive with comparable levels of education and experience. Their human capital may be less transferable initially, due to potential difficulties related to language, culture, education quality, and possibly discrimination.

That fewer immigrants have a home language that is either English or French may negatively affect entry-level earnings (although unfortunately we do not have good measures of language ability at entry). Studies by Baker and Benjamin (1994), Frenette and Morissette (2003), and Aydemir and Skuterud (2004) suggest that perhaps one-third of the decline in entry-level earnings is associated with the changing characteristics of immigrants, particularly the shift in source regions and home language.

2) Other Factors Associated with the Shift to the Newer Source Regions: Education Quality, Language Skills and Discrimination

There are other effects, largely unmeasurable, such as a potentially lower quality education system in some of the countries from which immigrants now arrive, that may be associated with the change in the source regions. Sweetman (2003) has attempted to focus on this issue by using test scores in international literacy and numeracy surveys to proxy a country's quality of education. He finds some support for the notion that lower school quality in some of the newer source regions contributes to the decline in entry-level earnings.

A decline in language skills among newer immigrants may be another potential contributor. Language and communication skills are related to productivity, and hence, the wages of workers. Unfortunately, good quality data on change in the language skills (in either official language) of recent immigrants through time does not exist. It has, therefore, been difficult for researchers to assess the impact of this variable. Even if conversational skills appear adequate, it may be that immigrants' ability to work in either official language, when performing more complex tasks, has declined among entering immigrants. A recent paper by Ferrer, Green and Riddell (2003) asks whether the decline in returns to foreign experience (discussed later) may be related to increased relative literacy or numeracy issues among older immigrants. Using detailed data with test scores on literacy and numeracy, they find little evidence to support this notion though they do find a direct effect of literacy and numeracy.

Increased discrimination is also a possibility, as the number of immigrants who are visible minorities has increased significantly. Here again, reliable data to estimate the effect of this factor are scarce. Some of these issues (e.g., language, school quality, discrimination) may be more prevalent among immigrants from the more recent source countries, such as Eastern Europe, Asia and Africa. The increase in immigrants from these countries has increased the importance of these issues for immigrants as a whole, resulting in a decline in entry-level earnings. If this were so, then the effects would be picked up in the type of statistical analysis reported above, which indicates that roughly one-third of the earnings decline is related to the changing composition of immigrants by source country and other factors. If the importance of these issues is changing for particular source countries (e.g., declining school quality relative to the Canadian system, increasing discrimination for people from particular countries, deteriorating language skills from immigrants within country), or if the importance of these issues is increasing in the Canadian labour market (e.g., similar language deficiencies are more detrimental to immigrants now than formerly, because the demand for superior language and communication skills is increasing), then they would not be captured in the "one-third" mentioned above. However, assessing the effect of such changes is difficult given the data available.

3) Returns to Years of Schooling and Credentialism

Returns to years of schooling have been declining among immigrants from the newer source countries in particular (Aydemir and Skuterud, 2003). A year of university education, for example, when converted to earnings in the labour market, was worth less to recent immigrants in the late 1990s than during the 1960s. This has occurred within source regions and not simply as a result of the changing composition of immigrants by source region. Hence, we have seen the earnings of university-educated immigrants decline and their low-income rate increase, similar to immigrants with lower education levels.

Researchers studying credentialism have posed the question in a slightly different manner. After accounting for years of schooling, they ask what is the value of a university credential (i.e., a degree) to an immigrant. This is referred to as the "sheepskin" effect. Is the earnings advantage of having a university degree, relative to not having it, changing? Hypotheses relating to credentials have been expressed for a number of years. However, to contribute significantly to the decline in entry-level earnings, this effect would have had to become more important in recent years.

Recent evidence presented in two papers (Ferrer and Riddell, 2003; Aydemir and Skuterud, 2004) suggests that this is not happening. This research focuses on the "sheepskin" effect, that is, the value of a university degree itself (not the number of years of schooling required to obtain the degree but the simple fact of having this credential). The research does not distinguish between degrees in licenced fields, such as medicine and engineering, and unlicenced fields such as history and economics. Nonetheless, the results are instructive. Both studies find that having a degree increases immigrant earnings significantly and that this effect is stronger for immigrants than for the Canadian born. However, researchers also conclude that the advantage of this credential for immigrants has changed little since the early 1980s. This does not mean that the earnings of immigrants with degrees have not declined. As noted earlier, earnings have fallen, as the returns to years of schooling have declined. Rather, it means that the relative advantage of having a university degree, as opposed to not having a degree, has not changed significantly. This suggests that the "credentialism" issue across all fields has not worsened, although it may or may not have worsened in specific fields.

4) Declining Returns to Foreign Labour Market Experience

Another factor taking on increasing importance in research is "declining returns to foreign experience." This refers to the concentration of the earnings decline at entry among older immigrants. Human capital consists largely of education/ training and the skills developed through work experience. One typically expects some return to human capital when entering employment. A number of recent studies indicate that the foreign work experience of entering immigrants is increasingly discounted in the Canadian labour market (Schaafsma and Sweetman, 2001; Green and Worswick, 2002; Frenette and Morissette, 2003; Aydemir and Skuterud, 2003). Put another way, older immigrants entering Canada in the late 70s or early 80s earned significantly more than their younger counterparts, presumably because of their experience (controlling for education, etc.). Foreign work experience appears to be more heavily discounted, and immigrants who land at an older age seem to have an increasing earnings disadvantage.

Schaafsma and Sweetman (2001) observe that age at immigration matters. Using Census data from 1986 to 1996, they find that immigrants who arrived in their 30s or later experience, on average, low (or zero) returns to foreign labour market experience, and low returns to foreign education. Green and Worswick (2002) observe that returns to foreign experience have fallen for successive entering immigrant groups, thus making this a possible explanation for declining entry-level earnings. They also observe this effect among both family and economic classes, indicating that this factor is not driven by a shift in immigrant classes. Green and Worswick (2002) conclude that declining returns to experience is a major, and perhaps the most important, factor associated with the decline in earnings among recent immigrants. Studies by Aydemir and Skuterud (2003), and Frenette and Morissette (2003), provide support for this notion. They report that earnings did not decline among very young "recent" immigrant workers (aged 25 to 29) between 1980 and 2000, but that the decline in earnings was concentrated among male immigrant workers over aged 30.

Green and Worswick (2002), when first observing this effect, suggest that it may be related to two factors: (1) the shift in the composition of immigrants those from source countries from which it is more difficult to evaluate the benefits of the foreign experience, and (2) a decline in the returns to foreign experience within source regions. They conclude that the decline in returns to foreign experience is concentrated among immigrants from the newer source regions and is not evident among immigrants from the traditional source regions, such as Northwestern Europe and English-speaking countries. Aydemir and Skuterud (2003) also find that the decline in returns to foreign labour market experience is observed only among immigrants from "non-traditional" source regions, most notably Eastern Europe, Asia, and Africa. They also find that foreign experience has become almost worthless in the Canadian labour market for immigrants from these regions.

What accounts for the decline in returns to foreign experience within source regions? There is little hard evidence on this point. As noted above, Ferrer, Green and Riddell (2003) find, though the estimates are imprecise, that relative literacy competencies between younger and older workers accounted for little of the effect. Others have discussed the possibility of an emerging technological divide. The experience acquired with the technologies used in the Eastern European and Asian countries may not be directly applicable to the Canadian workplace, and hence, may explain the discounted returns to experience. This effect could be increasing in importance, as technologies change more rapidly in more advanced countries like Canada. With a rapidly rising supply of highly educated workers in Canada (see point 7 below), employers may have less incentive to evaluate the foreign work experience (and years of schooling) of immigrants, particularly those who are highly educated (or they may evaluate it and find it to be not as readily applicable to the Canadian context as previously). Thus, the experience and schooling are both somewhat discounted, but this is only speculative at this stage.

Aydemir and Skuterud (2003) conclude that, among recent immigrants, the declining returns to years of schooling, and in particular, the declines in returns to foreign experience account for roughly two-thirds of the decline in entry-level earnings reported earlier.

5) Deteriorating Labour Market Outcomes for New Labour Market Entrants in General, of Which Immigrants are a Part

Labour market outcomes for young labour market entrants have been deteriorating in Canada through the 1980s and 1990s. Earnings of young men, in particular, have been falling through these decades (Picot, 1998; Beaudry and Green, 2000). Entering immigrants are themselves new labour market entrants, and it may be that whatever is causing the decline in earnings of the young in general (and that is not well understood) may be also affecting the earnings of "recent" immigrants. Green and Worswick (2002) find that this may account for 40 percent of the decline in entry-level earnings for recent immigrant men. They also find, however, that this effect was concentrated in the 1980s and was less important in the 1990s. Frenette and Morissette (2003) and Aydemir and Skuterude (2003) came to the same conclusion that the general decline in labour market outcomes for young workers is detrimental to entering immigrants as well.

6) Fluctuations in Macro-Economic Conditions

Canada experienced two severe recessions during the early 1980s and 1990s. A major decline in entry-level earnings was first observed in the early 1980s. The position of recent immigrants declined significantly again in the early 1990s, when immigration levels remained high. This has prompted a number of researchers to speculate that changing macro-economic conditions have played a major role in earnings trends. Bloom and Gunderson (1991) and McDonald and Worswick (1998) noted the effect of the business cycle on immigrant earnings, and the latter concluded that immigrants are more negatively affected by recessions than the Canadian born. In a paper on unemployment rate differentials between immigrants and the Canadian born, McDonald and Worswick (1997) conclude that the unemployment gap is much larger during recessions, and much less, or near zero during expansions. Reitz (2001) concludes that fluctuations in macro-economic conditions do not fully explain the general downward trend in immigrants' economic performance. Aydemir (2002), focusing on declines in employment and participation rate (not earnings) among immigrants, concludes that macro-conditions (as measured by the unemployment rate) may account for up to 50 percent of the decline in the participation rate and less of the decline in the employment rate among immigrants during the recession in the early 1990s. There was a substantial deterioration in economic outcomes (participation and employment rates) even after accounting for macro-economic conditions. During the last half of the 1990s, economic conditions improved dramatically; by 2000 unemployment fell to levels not seen in over 20 years. The relative earnings of immigrants, to comparable Canadian-born workers, marginally improved but remained far below the levels observed in the 1970s or even in the 1980s. (This was noted above.) Hence, it seems unlikely that macro-economic conditions account for all or even most of the long-term decline in the relative earnings of recent immigrants.

7) Strong Competition from the Increasingly Highly Educated Canadian-Born

The supply of highly educated workers in Canada has been increasing at a very rapid pace. For example, the number of women in the labour force with a university degree has quadrupled in just 20 years. The number of men has more than doubled. Studies on the "wage premium" associated with a university degree (i.e., the earnings of university graduates compared to high school graduates) have generally observed that this premium has increased significantly in the United States but little in Canada over the past twenty years (Burbidge et al., 2002). (4)

In the face of an apparently increasing demand for more highly educated workers, why have the relative wages of university graduates risen in the United States but not in Canada? Two high profile papers argue that this is related to differences in the supply of highly educated workers in the two countries. The supply of highly educated workers (relative to the high school educated) has been increasing much more rapidly in Canada than in the United States (Freeman and Needels, 1993; Murphy, Romer and Riddell, 1998). Reitz (2001) argues that, in spite of the rising educational levels of immigrants, their relative economic advantage has declined as a result of the more rapidly rising levels of education among Canadian-born workers. He also suggests that immigrants have not benefited to the same extent as the Canadian born from increases in education.

This fits with other findings. Although the "sheepskin" effect (the fact of having a degree) has not diminished among immigrants, the returns to years of education have fallen among "recent" immigrants. Labour market outcomes among university educated immigrants have also been declining faster than among others (controlling for age), as noted above. In a very competitive labour market for new entrants, including highly educated new entrants, it may be that foreign experience and years of schooling from the "new" source countries are increasingly discounted. Language and other factors associated with the new source regions place "recent immigrants" at a relative competitive disadvantage.

8) Changes in Social Transfer Usage, and Other Income Sources

The above-mentioned reasons may influence the earnings of immigrants but low-income rates may be affected by other factors as well. Changes in the availability or use of the social transfer system (social assistance, employment insurance, etc.) by immigrants can obviously affect low-income rates. Income inequality may also have changed among the immigrant population. Most of the earnings analysis relates to mean or median earnings. Rising levels of family income inequality, unemployment and possibly the discouraged worker effect could result in increasing low-income rates, above and beyond that expected, based on changes in mean family earnings. Little recent research has been done on these issues.

Conclusion

As the economic welfare of recent immigrants deteriorated through the 1980s and 1990s, it seems unlikely that many "recent" immigrants will achieve earnings or low-income rate levels comparable to the Canadian born, or even those of earlier immigrant cohorts (Frenette and Morissette, 2003). Possible reasons for this decline are numerous, and they are not, of course, necessarily independent of one another. For example, increased competition from domestic labour sources, combined with a shift in the source regions of immigrants, may together result in a reduced ability of immigrants to convert their education and experience into earnings. Declining returns to foreign experience, decreasing language ability (which may be the case but is largely unmeasured) and changes in discrimination may themselves be related to changes in source countries.

Current research suggests that some (perhaps as much as one-third) of the decline in aggregate earnings among entry-level immigrants is related to the shift in language skills and source regions. Rising levels of education should have offset this decline, to some extent, but apparently did not. Declining returns to foreign experience appears to top the list as the most likely significant factor contributing to rising low-income among immigrants. This deterioration seems more heavily concentrated among older workers and is affected by declining returns to years of schooling. As this review of the literature suggests, more research is necessary to resolve current debates about the many possible explanations for the deteriorating economic welfare of Canadian immigrants.

Table 1. Low-income rates by immigration status, Canada, 1980-2000

 Total Non- All
 population immigrants immigrants

1980 0.171 0.172 0.170
1985 0.187 0.185 0.193
1990 0.155 0.151 0.171
1995 0.191 0.176 0.247
2000 0.156 0.143 0.202

 Low-income rate *
 Years of residence in Canada

 All
 <= 5 6-10 11-15 16-20 > 20 immigrants

1980 0.246 0.187 0.144 0.147 0.167 1.0
1985 0.342 0.260 0.198 0.159 0.165 1.0
1990 0.313 0.242 0.190 0.152 0.126 1.1
1995 0.470 0.353 0.272 0.221 0.155 1.4
2000 0.358 0.283 0.227 0.191 0.133 1.4

 Low-income rates relative to Canadian-born
 Years of residence in Canada

 <= 5 6-10 11-15 16-20 > 20

1980 1.4 1.1 0.8 0.9 1.0
1985 1.8 1.4 1.1 0.9 0.9
1990 2.1 1.6 1.3 1.0 0.8
1995 2.7 2.0 1.6 1.3 0.9
2000 2.5 2.0 1.6 1.3 0.9

* After transfer, before taxes; based on the 1992 LICO updated for
change in CPI only Data sources the 1981 to 2001 Census 20% sample
data; Source: Picot and Hou, 2003

Table 2: Percentage Change in the Low-Income Rate, 1980-2000, by Source
Region

 Controlling for age, education,
 Raw data language, family type

Source region
 Southeast Asia -35% -25%
 Western Europe -24% -16%
 U.S. -18% -12%
 Caribbean -2% -5%
 South & Central America -2% +1%
 Northern Europe +14% 24%
 Eastern Europe +22% 34%
 Western Asia +52% 64%
 Southern Europe +61% 115%
 East Asia +68% 90%
 South Asia +82% 86%
 Africa +121% 94%

Source: Picot and Hou, 2003

Table 3: Low-Income Rates Among Recent Immigrants Aged 25 to 65, by
Education Level

 change in rate,
Highest education level 1990 2000 1990-2000

Less than high school 34.3 38.4 13.1%
High school graduate 31.0 38.8 25.2%
Some post-secondary 26.4 33.7 27.6%
University degree 19.1 27.5 44.0%

Source: Picot and Hou. 2003

Notes

(1) The proportion of the population with family incomes below Statistics Canada Low-Income Cutoff (LICO) is fixed at the 1992 level and adjusted only for changes in the CPI. Hence, over the period of study, this is a "fixed", not a relative, low-income measure.

(2) The "high-knowledge" sector was the fastest growing sector during the late 90s; employment grew by 28.6 percent in this sector, compared to 6.4 percent for the economy as a whole.

(3) Statistics Canada's after transfer, before tax LICO (Low-Income Cutoff) is used, since taxes paid are not available in the census data.

(4) The wage premium has been increasing in recent years among the young in Canada, however.

References

Abbott, Michael G. and Charles M. Beach. 1993. Immigrant earnings differentials and birth-year-effects for men in Canada: post-war-1972. Canadian Journal of Economics. 25: 505-524.

Aydemir, Abdurrahman. 2002. Effects of Business Cycles on the Labour Market Assimilation of Immigrants. Analytical Studies Research Paper Series No. 203, Statistics Canada, Ottawa.

Aydemir, Abdurrahman and Mikal Skuterud. 2003. Explaining the Deteriorating Entry Earnings of Canada's Immigrant Cohorts: 1966-2000. Analytical Studies Research Paper Series, Statistics Canada, Ottawa. Forthcoming.

Baker, Michael and Dwayne Benjamin. 1994. The performance of immigrants in the Canadian labour market. Journal of Labor Economics 12: 369-405.

Beaudry, Paul and David Green. 2000. Cohort patterns in Canadian earnings: assessing the role of skill premia in inequality trends. Canadian Journal of Economics 33: 907-936.

Bloom, David E. and Morley Gunderson. 1991. An analysis of the earnings of Canadian immigrants. In Immigration, Trade and the Labour Market, eds. John M. Abowd and Richard B. Freeman. Chicago: The University of Chicago Press.

Borjas, George J. 1985. Assimilation, changes in cohort quality, and the earnings of immigrants. Journal of Labor Economics 3: 463-489.

Burbidge, J.B., L.S. Magee, and A.L. Robb. 2002. The Education Premium in Canada and the United States. Canadian Public Policy, 28 (2): 203-217.

Carliner, Geoffrey. 1981. Wage differences by language group and the market for language skills in Canada. Journal of Human Resources 16: 384-399.

Chiswick, Barry R. 1978. The effect of Americanization on the earnings of foreign-born men. Journal of Political Economy 86: 897-921.

Ferrer, Ana, David Green and Craig Riddell. 2003. The Effect of Literacy on Immigrant Earnings. University of British Columbia, Department of Economics.

Ferrer, Ana and Craig. Riddell. 2003. Education, Credentials and Immigrant Earnings. University of British Columbia, Department of Economics.

Freeman, R.B. and K. Needels. 1993. Skill Differentials in Canada in an Era of Rising Labor Market Inequality. In Small Differences That Matter, ed. D. Card and R.B. Freeman. Chicago: University of Chicago Press.

Frenette, Marc and Rene Morissette. 2003. Will they ever converge? Earnings of immigrant and Canadian-born workers over the last two decades. Analytical Studies Research Paper Series, No. 215, Statistics Canada, Ottawa.

Grant, Mary L. 1999. Evidence of new immigrant assimilation in Canada. Canadian Journal of Economics. 32: 930-955.

Green, David A. and Christopher Worswick. 2002. Earnings of Immigrant Men in Canada: The Roles of Labour Market Entry Effects and Returns to Foreign Experience. University of British Columbia, Department of Economics, prepared for Citizenship and Immigration Canada.

Hum, Derek and Wayne Simpson. 2003. Reinterpreting the Performance of Immigrant Wages From Panel Data. University of Manitoba, Department of Economics.

Li, Peter S. 2003. Initial Earnings and Catch-up Capacity of Immigrants. Canadian Public Policy 29 (3): 319-337.

McDonald, James Ted and Christopher Worswick. 1997. Unemployment Incidence of Immigrant Men in Canada. Canadian Public Policy 23 (4): 353-373.

McDonald, James Ted and Christopher Worswick. 1998. The earnings of immigrant men in Canada: Job tenure, cohort, and macroeconomic conditions. Industrial and Labour Relations Review 51: 465-482.

Meng, Ronald. 1987. The earnings of Canadian immigrant and native-born males. Applied Economics 19: 1107-1119.

Murphy, Kevin M., Paul Romer and Craig Riddell. 1998. Wages, skills and technology in the United States and Canada. NBER working paper series, 6638. National Bureau of Economic Research.

Picot, Garnett. 1998. What is Happening to Earnings Inequality and Youth Wages in the 1990s? Analytical Studies Research Paper Series, No. 116, Statistics Canada, Ottawa.

Picot, Garnett and Feng Hou. 2003. The rise in low-income among immigrants in Canada. Analytical Studies Branch Research Paper Series, 11F0019MIE2003198, Analytical studies. Ottawa: Statistics Canada.

Reitz, Jeffrey. 2001. Immigrant success in the knowledge economy: institutional changes and the immigrant experience in Canada, 1970-1995. Journal of Social Issues 57: 579-613

Schaafsma, J. and A. Sweetman. 2001. Immigrant Earnings: age at immigration matters. Canadian Journal of Economics 34 (4): 1066-99.

Sweetman, Arthur. 2003. Immigrant Source Country Education Quality and Canadian Labour Market Outcomes. Queen's University: School of Policy Studies.

Warman, Casey R., and Christopher Worswick. 2004. Immigrant Earnings Performance In Canadian Cities: 1981 Through 2001. This issue.