The resettlement of Vietnamese refugees across Canada over three decades

ABSTRACT

Welcoming 60,000 Southeast Asian refugees in the 1979–80 period has become a celebrated part of Canada's history, but the eventual integration of these refugees into Canadian society has received insufficient attention. This study provides a comprehensive overview of Vietnamese refugees’ economic outcomes over the three decades after their arrival. This study also explores how regional contexts contributed to shaping economic outcomes. Based on analyses of multi-year census data, this study finds that adult Vietnamese refugees arrived with little human capital, but they had high employment rates, and over time they closed their initial large earnings gap with other immigrants. Childhood Vietnamese refugees out-performed other childhood immigrants and similar-aged Canadian-born individuals in educational attainment and earnings when they reached adulthood. The geographic region of residence was associated with some large variations in refugees’ socioeconomic outcomes; and regional differences in refugees’ human capital characteristics, ethnic enclave, and economic conditions played varying roles depending on the outcome measure and length of residence.

KEYWORDS:

• Canada
• economic integration
• ethnic enclaves
• Vietnamese refugees

1. Introduction

At the height of the ‘Vietnamese boat people’ crisis, major Western industrialised countries reacted differently in terms of the number of refugees admitted for permanent resettlement and how these refugees were absorbed into society (Bankston III and Zhou 2021; Barber 2021; Bösch and Su 2021; Gisselquist 2021). As a key player of this international undertaking, Canada welcomed 60,000 Southeast Asian refugees between 1979 and 1980 – the largest single influx of refugees admitted to the country in a short period. Canada accepted more refugees per capita than any other resettlement country at that time. Canadians from cities and towns of all sizes were mobilised to welcome these refugees to their communities. The generous and unprecedented welcome Canadians extended to these refugees was recognised internationally, and the United Nations High Commissioner for Refugees awarded its prestigious Nansen Medal to the ‘People of Canada’ in 1986. It is the only time the honour has been granted to an entire population.

With exceptional efforts, Canadians provided Southeast Asian refugees with an escape from persecution and the chance to establish a new life in a peaceful and prosperous society. These efforts have demonstrated convincingly the effectiveness of a model of government–public collaboration in resettling refugees, and created broad-based organisations that have played indispensable roles in Canada's responses to subsequent refugee crises (Canadian Council for Refugees 1999; Molloy et al. 2017). More than three decades have passed. Much has been written to record the events and experiences in that historical period. Numerous studies have documented various aspects of Southeast Asian refugees’ socioeconomic integration in Canada, particularly in terms of learning an official language, finding employment, health, and psychological wellbeing (e.g. Beiser 1999). Most of these studies were based on refugees’ outcomes in their initial years of resettlement and in specific regions or local communities.

This study adds to the literature by providing a comprehensive overview of Vietnamese refugees’ economic outcomes over the three decades after they began their new life in Canada. The socioeconomic outcomes of an immigrant group are influenced by a variety of factors. Segmented assimilation theory postulates that multifaceted interactions between the characteristics of an immigrant group and structural contexts in the receiving society lead to different paths of incorporation. The key characteristics of an immigrant group include financial and human capital upon arrival, and cultural values and norms. Structural contexts refer to racial stratification, economic opportunities, pre-existing ethnic communities, and the socioeconomic environment of local communities where immigrants reside (Portes and Zhou 1993; Zhou 1997; Gisselquist 2021). Relating these factors to Vietnamese refugees in Canada, this article first discusses Canada's policy and social environment for comparison with other major destination countries as discussed in other articles in this special issue (Bankston III and Zhou 2021; Barber 2021; Bösch and Su 2021). This article further examines Vietnamese refugees’ human capital deficiencies, and the contexts of regions and local communities where Vietnamese refugees were located, including regional economic conditions and the effect of ethnic enclaves. One unique contribution of this study is the in-depth exploration of how the effects of these factors in shaping refugees’ labour market outcomes changed over time and across generations.

2. National efforts of resettling Vietnamese refugees in Canada

Canada's settlement of Vietnamese ‘boat people’ occurred at a unique period in the country's immigration history. A new Immigration Act, tabled in 1976 and coming into effect in 1978, articulated, for the first time, the main objectives of Canada's immigration policy. One of the main objectives was to ‘fulfil Canada's international legal obligations with respect to refugees and to uphold its humanitarian tradition with respect to the displaced and the persecuted’. This was also the first Canadian immigration legislation to recognise refugees as a special class of immigrants. Prior to 1978, refugees were admitted on an ad hoc, case-by-case basis, and as exceptions to regular immigration procedures (Molloy et al. 2017). The new Act entrenched the definition of a Convention refugee, created a refugee determination system, and enabled the private sponsorship of refugees.

The humanitarian obligations embedded in the new Immigration Act and the provision of private sponsorship were immediately put to the test. Touched by the desperate plight of mass exodus of Southeast Asian refugees who took to the high seas in makeshift boats, many Canadians came forward to offer their help. The outpouring of public support prompted the Canadian government to quickly and substantially expand its own commitment. For each refugee individual or family admitted under private sponsorship, the government pledged to bring in another under the government-assisted programme. Response from the public was so strong that the number of private sponsorships escalated and surpassed what the government was able to match. As a result, over half of the 60,000 Southeast Asian refugees who arrived in the 1979–80 period came through the private sponsorship stream (Canadian Council for Refugees 1999; Beiser 2003; Molloy et al. 2017).

Private sponsorship played a key role in Southeast Asian refugees’ socioeconomic integration. Beiser (2003) showed that privately sponsored Southeast Asian refugees had better outcomes in terms of being gainfully employed, being able to speak an official language of the receiving country, and feeling in good health when compared to government-assisted refugees. Neuwirth and Clark (1981) suggested that private sponsorship may provide a broad range of services beyond material help, including support for social and cultural adjustment. The friendship and emotional bonds between many refugees and their private sponsors often lasted years after the formal sponsorship period (Roma 2016). Likely for these reasons, privately sponsored refugees maintained higher employment rates and earnings than government assisted refugees up to 15 years after arrival when differences in measurable sociodemographic characteristics between the two groups are taken into account (Kaida, Hou, and Stick 2019).

Private sponsorship also strongly affected the geographic distribution of Southeast Asian refugees across the country. Privately sponsored refugees were received by their sponsors, who were scattered over Canada's ten provinces and two territories. At the same time, each province or territory was assigned a relocation quota based on its proportional representation to the total population of Canada. These regional quotas were a key consideration of government officers in assigning the destination of government-assisted refugees. Another key consideration was whether a refugee had relatives or friends in a particular area. Quite often the need to meet the regional quotas took priority, a situation that often resulted in secondary migration to join family (Beiser 2003; Simich 2003). Overall, Southeast Asian refugees were more evenly distributed across the country than other immigrants in their initial destinations. Furthermore, many of them remained in their initial settlement communities many years after their arrival, although they tended to have a higher rate of secondary migration than other immigrants (Simich 2003; Hou 2007).

Although this paper will not evaluate the benefits of private sponsorship relative to other avenues of refugee resettlement since the data used in this paper do not contain relevant information, the fact that over one-half of Vietnamese refugees who arrived in 1979–80 were sponsored by non-government organisations and individuals can help us understand the path and process of their economic integration. Because their initial destinations to a large extent were not chosen by themselves, the examination of how the contexts of local communities affect their labour market outcomes is less subject to selection bias (at least in the initial years) than similar studies for regular immigrants.

3. Data, measures, and methods

This article draws on data from the Canadian census over three decades and uses quantitative techniques to study relationships between factors that influence economic integration. It takes advantage of the large and representative sample and standard measures of labour market outcomes, and thus is able to provide an objective portrait of refugees’ economic integration. Using a synthetic cohort approach, this article covers adult refugees’ progress in the labour market over their entire work career and examines childhood refugees’ achievements when they reached adulthood. This longitudinal perspective reflects the realities that integration takes time and the factors affecting immigrant labour market outcomes have different salience at different stages of the integration process. However, census data do not collect information on respondents’ attitudes and social behaviours, and thus could not reveal the mechanisms behind some observed relationships. The findings presented here might be complemented by more qualitative work that sheds light on the underlying processes.

This analysis focuses on Vietnamese refugees who arrived in Canada in 1979 and 1980 for three reasons. First, in these two years Canada received the largest number of the ‘boat people’. Vietnamese refugees started coming to Canada in 1975, and about 7,700 of them arrived between 1975 and 1978. This first wave of Vietnamese refugees consisted mostly of well-educated professionals, middle-class, and supporters of the US and South Vietnamese governments (Wood 1997; Beiser 1999). They were quite different from the second-wave refugees, who arrived after 1978, in their socioeconomic status and experiences of fleeing. In particular, many second-wave refugees were ethnic Chinese who, although they might have lived in Vietnam for generations, were singled out for persecution after conflict between Vietnam and China.

Second, the census did not collect information on immigrant class that is needed to directly identify refugee status. However, immigration records show that about 98 per cent of the individuals who were born in Vietnam and came to Canada over the 1979–80 period were refugees (Employment and Immigration Canada 1981, 1982). Vietnamese refugees continued to come after 1980, but in much smaller numbers, and the share of refugees among immigrants decreased rapidly, making it difficult to separate refugees from other immigrants with the census data for those who arrived after 1980. Finally, the initial outcomes of the 1979–80 arrivals can be measured in the first to second year after arrival in the 1981 census, which provides a narrowly defined starting point from which to examine their subsequent integration patterns.

For comparison purposes, this study also includes immigrants from other source regions who arrived in 1979–80, and the Canadian-born population. Comparison with other immigrants who arrived in the same period is a way to distinguish difficulties of resettling in a new country that new immigrants would generally experience from hardships that are unique to refugees. This study excludes refugees from Laos and Cambodia, who were a smaller part (about 20%) of Southeast Asian refugees arriving in the same period. Compared with Vietnamese refugees, refugees from Laos and Cambodia tended to have lower educational levels and less successful labour market outcomes.

In the 1981 census 20 per cent sample microdata file, there were 7,247 Vietnamese who arrived in 1979 and 1980, corresponding to an estimated 36,730 population. This estimated population size is somewhat smaller than the number of total arrivals from Vietnam (43,710) recorded in immigration statistics (Employment and Immigration Canada 1981, 1982). It is possible that some refugees might have emigrated (i.e. to the US and other settlement countries) at the time of the 1981 census, or refugees might be slightly under-sampled in the census.

This study uses different outcome indicators for refugees and immigrants who arrived at different life stages: childhood – arrived at age 17 or younger, and prime-age – arrived at ages 18–44. Among the Vietnamese refugee sample included in this study, about 36 per cent (2,646) were aged 0–17 at immigration, 57 per cent (4,124) aged 18–44, and only 7 per cent (477) aged 45 or over. These age groups would certainly experience different paths of socioeconomic integration in Canadian society, and should be evaluated separately. For childhood arrivals, high-school dropout rate, university completion rate (i.e. obtaining at least a bachelor's degree), and annual earnings upon reaching adulthood are used as outcome measures. For prime-age arrivals, employment rate and annual earnings are used as the outcome measures. An individual is defined as employed if he/she was engaged in gainful employment in the census reference week (i.e. the week before the census date).

Using a synthetic cohort approach, this study examines the selected outcomes over three decades following the Vietnamese refugees’ arrival in Canada. Specifically, descriptive statistics are produced in the initial years (1–2 years after arrival) in the 1981 census, 11–12 years (the 1991 census), 21–22 years (the 2001 census), and 31–32 years (the 2011 National Household Survey) for Vietnamese refugees, as well as for other immigrants who arrived in 1979–80. Similar statistics are examined as they pertain to the Canadian-born population. To achieve a better matched comparison, analyses by arrival-age group are compared to the Canadian-born population who would have been in the same age range as the refugees in a given census. For instance, refugees who arrived at ages 18–44 would have been 19–45 in 1981, and 29–55 in 1991. The Canadian-born comparison groups would be aged 19–45 in 1981 and 29–55 in 1991.

To examine possible variations in the selected outcomes by geographic region, descriptive statistics are produced for six separate regions: Toronto, Montreal, Vancouver, second-tier cities (including the metropolitan areas of Ottawa, Calgary, Edmonton, Winnipeg, Hamilton, and Quebec City), small metropolitan areas, and small urban or rural areas. The first three largest metropolitan areas are the gateways of contemporary immigration to Canada.

For adult arrivals, multiple regression models are constructed to account for observed regional differences in employment rates and annual earnings. The main predictors include age, marital status, education, official language ability, regional ethnic concentration, and regional labour market conditions.

Regional ethnic concentration is defined as the share of individuals with Chinese ethnic origin in a region for Vietnamese refugees who reported Chinese ethnic ancestry, and the share of individuals with Vietnamese ethnic origin in a region for refugees who reported Vietnamese ethnic origin. In deriving this measure, a region is defined as a census metropolitan area (an urban area with a population over 100,000), or census agglomeration (a small urban area with a population over 10,000) for urban areas, or census division (county or other provincially legislated regional districts) for rural areas. In 1981, there were 171 such regions where Vietnamese refugees were located. Regional labour market conditions are measured by employment rates among prime-aged (ages 25–54) Canadian-born men in a region by four education levels (less than high-school graduation, high-school graduation, some post-secondary education, and with a university degree), and average annual earnings among prime-aged Canadian-born men by four educational levels. The regional employment rate is used in the models predicting employment among refugees, while regional average earnings are used in the models predicting earnings.

A regression decomposition technique is used to evaluate the relative role of each predictor in accounting for the observed regional differences in outcomes (Hou 2014).

4. Results

4.1. Prime-aged refugees

4.1.1. Improvement in language ability and education

Since prime-aged (18–44 years) arrivals constituted the majority of Vietnamese refugees, their labour market outcomes would determine the group's overall economic wellbeing, at least in the first couple of decades. Their success in the labour market depends to a large extent on their human capital characteristics, particularly the ability to speak an official language and educational level. Table 1 shows the percentages speaking English or French and high-school graduation rates for prime-aged Vietnamese refugees and other immigrants by length of stay in Canada, and for similar-aged Canadian-born individuals.

Table 1. Human capital improvement over time among prime-age Vietnamese refugees and other immigrants who arrived in 1979–1980.

	Speaking an official language				High school graduation
	1981	1991	2001	2011	1981	1991	2001	2011
Vietnamese refugees	percent				percent
Overall	75.4	88.4	88.1	84.0	38.2	43.1	45.8	61.2
Montreal	83.6	91.2	91.0	88.3	59.9	61.7	69.3	74.4
Toronto	68.8	86.0	84.9	80.6	39.9	41.6	42.3	59.5
Vancouver	70.3	87.1	83.6	80.1	39.7	41.8	42.0	60.8
Second-tier cities	69.8	88.7	90.3	85.8	34.6	39.1	45.4	57.8
Small metropolitan areas	82.7	93.4	94.3	91.1	37.0	41.0	42.3	60.2
Small urban or rural areas	83.3	91.1	95.6	91.4	31.4	39.2	36.6	61.7
Immigrants from other countries
Overall	90.4	95.8	96.1	95.3	68.8	74.1	76.1	82.9
Montreal	93.0	96.6	96.8	97.7	66.0	69.5	70.5	75.2
Toronto	88.1	95.0	95.5	93.7	68.5	73.2	75.7	82.2
Vancouver	87.4	92.8	93.1	90.4	69.8	74.2	74.3	82.4
Second-tier cities	92.0	96.7	96.2	97.0	70.2	76.3	77.9	85.3
Small metropolitan areas	92.1	97.3	98.6	97.5	72.2	76.7	80.5	86.6
Small urban or rural areas	92.2	98.1	98.5	99.2	67.6	75.5	78.0	87.1
Similar-aged Canadian-born
Overall	…	…	…	…	62.8	69.5	71.7	80.3
Montreal	…	…	…	…	70.6	73.3	76.3	81.3
Toronto	…	…	…	…	69.3	78.4	81.2	88.5
Vancouver	…	…	…	…	68.9	78.6	80.8	89.8
Second-tier cities	…	…	…	…	67.9	75.5	77.7	85.7
Small metropolitan areas	…	…	…	…	64.6	72.5	74.6	82.9
Small urban or rural areas	…	…	…	…	56.0	62.1	65.1	74.4

Note: … essentially all Canadian-born individuals speak an official language.

Data sources: Author's calculation based on data from the 1981, 1991, and 2001 census and 2011 National Household Survey.

Compared with other immigrants, Vietnamese refugees were less likely to speak English or French both in the initial years after arrival and in the long run. About 1–2 years after arrival (in 1981), about one-quarter of Vietnamese refugees could not speak English or French. Note that at the time of arrival very few refugees could speak an official language. According to the Immigrant Landing File (ILF), about 90 per cent of prime-aged Vietnamese refugees who arrived in 1980 self-declared as not speaking an official language, compared with about one-quarter among other immigrants who arrived in the same year (author estimate from the ILF). Assuming the census and ILF data sources are comparable, the majority of Vietnamese refugees acquired some ability to speak an official language within 1–2 years after arrival. This remarkable improvement was made possible likely because the federal government offered second-language classes to refugees and other new immigrants. A survey on Southeast Asian refugees showed that over three-quarters of adult Southeast Asian refugees had taken government-funded language classes, close to 10 per cent had participated in formal Canadian education, and about 10 per cent had used private English tutoring within 1–2 years after arrival (Hou and Beiser 2006). In spite of their extraordinary progress, about 12 per cent of adult Vietnamese refugees remained unable to speak an official language after 20 years in Canada, compared with 4 per cent for other immigrants.

There were also large regional variations in the share of Vietnamese refugees who could not speak an official language. Toronto and Vancouver had the highest share of Vietnamese refugees who could not speak an official language, both in the initial years and after 2–3 decades of resettlement. It is possible that refugees who could not speak an official language were more likely to be attracted to the co-ethnic communities in Toronto and Vancouver both initially and in subsequent migration. The existence of large co-ethnic communities could also provide a socioeconomic environment and leisure activities that allow the use of a mother tongue, and reduce the need and incentives to learn the language of the new society (Espenshade and Fu 1997). Conversely, Vietnamese refugees had much higher rates of speaking an official language in small metropolitan areas and small urban or rural areas, particularly 2–3 decades after arrival. This is likely attributable to the lack of co-ethnic communities and outmigration. It could also be due to a greater degree of welcome in smaller communities leading to increase in the number of social contacts from the resident population.

Vietnamese refugees had a relatively high rate of speaking an official language in Montreal, both in the initial years and in the long run, compared with the pattern in the other two gateway centres. Montreal was the destination for over one-half of the first wave of Vietnamese refugees who arrived between 1975 and 1978. These first-wave refugees tended to be well-educated and privileged individuals before fleeing Vietnam. Many second-wave refugees who were settled in Montreal were likely their relatives who had higher socioeconomic status than other second-wave refugees who were settled in other parts of Canada. Furthermore, both the first wave and the 1979–80 cohort of Vietnamese refugees who settled in Montreal were mostly of Vietnamese ethnic origin and there was no pre-existing large-scale Vietnamese ethnic community and ethnic economy for them to rely on. In comparison, both the first wave and the 1979–80 cohort of Vietnamese refugees who settled in Toronto and Vancouver consist of a large proportion of individuals with Chinese ethnic origins who could rely on the large ethnic community for social and economic activities without the ability to speak the official language.

In addition to the ability to speak the official language of the receiving country, education is another key human capital characteristic that affects the economic outcomes of refugees and immigrants. The right-hand panel of Table 1 shows that Vietnamese refugees had much lower rates of high-school graduation than other immigrants and the similar-aged Canadian-born population both in the initial years and in the long run. While their high-school graduation rate did increase over time, the increase was small – only about 8 percentage points over two decades – given their low initial level. After 20 years in Canada, over one-half of the adult Vietnamese refugees had education below high-school graduation.

Similar to the regional differences observed in the ability to speak an official language, Vietnamese refugees who settled in Montreal had a much higher rate of high-school graduation than their counterparts in Toronto and Vancouver. As discussed above, it is possible that better-educated Vietnamese were settled in Montreal partly because many of them were related to the first wave of Vietnamese refugees who disproportionately settled in Montreal. More strikingly, the high-school graduation rate rose by about 10 percentage points over the first two decades among refugees in Montreal where refugees’ initial rate of high-school graduation was over 20 percentage points higher than in other areas. In contrast, little improvement was observed in Toronto and Vancouver.

4.1.2. Employment

In spite of their deficiencies in human capital characteristics, Vietnamese refugees were eager to find jobs, and within a year or two after arrival their employment rates surpassed those of other immigrants. The left panel of Table 2 presents the employment rates for prime-age Vietnamese refugees and other immigrants who arrived in 1979–80. In 1981, about 75 per cent of adult Vietnamese refugees who arrived in 1979–80 were already employed, compared with rates of 70 per cent among other adult immigrants and 73 per cent in the similar-aged Canadian-born population. Vietnamese refugees’ high level of employment within two years of arrival was likely due to two reasons. First, many privately sponsored refugees received help from their sponsors in finding employment. Second, Vietnamese refugees had a strong desire to become self-sufficient and also obligations to make and save money in order to sponsor family members who were left behind in refugee camps or in Vietnam. They took on any available jobs, often part-time and with little pay (Beiser 1999; Canadian Council for Refugees 1999).

Table 2. Employment rates and earnings of prime-age Vietnamese refugees and other immigrants who arrived in 1979–1980.

	Employment rates				Median annual earnings
	1981	1991	2001	2011	1980	1990	2000	2010
Vietnamese refugees	percent				2010 constant dollars
Overall	75.1	75.9	77.6	64.8	11,900	30,700	36,600	37,400
Montreal	64.8	71.5	71.7	59.7	11,800	26,700	33,600	30,500
Toronto	79.1	75.4	78.3	62.3	13,200	33,900	39,100	40,100
Vancouver	81.2	73.5	74.0	64.3	12,500	26,700	29,300	33,000
Second-tier cities	79.1	77.8	81.8	68.4	13,200	30,100	34,300	38,000
Small metropolitan areas	71.7	79.0	79.2	73.0	9,500	29,700	36,600	39,200
Small urban or rural areas	70.0	79.6	75.4	69.8	10,600	26,700	31,500	34,300
Immigrants from other countries
Overall	69.9	79.5	78.7	63.3	17,300	36,100	39,100	41,000
Montreal	62.2	71.3	70.6	57.1	15,800	29,700	31,800	33,400
Toronto	73.2	80.9	79.6	63.2	17,200	39,000	42,700	43,500
Vancouver	69.1	80.8	79.5	65.9	16,300	35,700	36,600	40,000
Second-tier cities	75.6	81.3	81.9	69.8	19,100	35,700	36,600	46,000
Small metropolitan areas	66.9	79.6	79.0	62.6	17,400	37,100	42,700	40,400
Small urban or rural areas	64.8	79.8	78.2	57.6	18,100	31,800	36,600	38,300
Similar-aged Canadian-born
Overall	73.0	77.8	72.7	56.8	29,400	38,600	41,300	41,600
Montreal	71.4	76.3	72.8	56.5	31,800	40,100	42,700	41,600
Toronto	81.4	82.8	79.7	66.4	31,700	47,600	54,800	54,800
Vancouver	79.9	81.0	77.6	63.0	34,200	43,500	48,800	50,000
Second-tier cities	77.9	81.4	76.4	61.0	31,800	42,800	46,400	50,000
Small metropolitan areas	73.6	78.8	73.1	56.5	29,100	39,500	42,700	42,700
Small urban or rural areas	69.0	75.2	69.7	53.0	26,600	33,400	36,400	35,000

Sources: Author's calculation based on data from the 1981, 1991, and 2001 census and 2011 NHS.

By 10–11 years after arrival (i.e. in 1991), however, Vietnamese refugees lost their lead over other immigrants in employment rates as the latter group had a faster rate of growth. The Canadian economy was in recession in 1991, and the less educated were hit particularly hard. Vietnamese refugees were likely affected more strongly because most of them did not finish high school. After 20 years (in 2001) and 30 years (in 2011) in Canada, Vietnamese refugees had similar employment rates as other immigrants, and both had higher rates than the similar-aged Canadian-born population.

A rather counter-intuitive regional variation in employment rates can be seen in Table 2 (left panel). Although Vietnamese refugees had a much higher proportion speaking an official language and a much higher high-school graduation rate in Montreal than in Toronto and Vancouver (as in Table 1), their employment rate was much lower in Montreal, particularly in 1981. In 1981, the employment rates of Vietnamese refugees differed by 14.3 percentage points between Montreal and Toronto, and 16.4 percentage points between Montreal and Vancouver. These regional differences more or less persisted over the entire three decades. There are two likely explanations for these regional differences. The first is an ethnic enclave effect, particularly in the initial years. Proportionately more Vietnamese refugees with Chinese ethnic origin lived in Vancouver and Toronto, where the existing ethnic economy and ethnic network could help newcomers find employment. The second is the impact of local labour market conditions. Given that other immigrants and the Canadian-born population had similar large differences in employment rates between Montreal and the other two major gateway centres, local labour market conditions could be the underlying determinant of the observed regional differences for refugees and other population groups.

To evaluate the relative role of ethnic enclave, regional labour market conditions, and individual-level characteristics in accounting for the observed large regional differences in employment rates among Vietnamese refugees, linear probability regression models were estimated for 1981 and 1991 separately, as in the left panel of Table 3. For each year, Model 1 contains only the dummy variables for regions, with Toronto as the common reference. The coefficients simply replicate the observed differences in employment rates between Toronto and each of the other regions, as shown in Table 2. For instance, the coefficient −0.143 associated with Montreal in Model 1 for 1981 indicates that the employment rate of Vietnamese refugees in Montreal was 14.3 percentage points lower than that in Toronto, and this difference is statistically significant. Model 2 adds in all the covariates. The changes in the coefficients associated with the dummy variables for regions from Model 1 and Model 2 represent the portion of the observed differences that can be accounted for by the added covariates. For instance, the coefficient associated with Montreal changed to −0.028 and was not significant in Model 2, implying that most of the observed 14.3 percentage point difference in employment rates between Montreal and Toronto was accounted for by the added covariates. In the same year, the large gaps in employment rates observed for small metropolitan areas and small urban or rural areas were also mostly accounted for by the added covariates.

Table 3. OLS regression models accounting for regional differences in employment rates and earnings among prime-aged Vietnamese refugees who arrived in 1979–1980.

	Employment rates				Annual earnings
	1981		1991		1981		1991
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Intercept	0.791***	−1.342***	0.754***	−0.674	9.269***	0.479	10.265***	2.4543
Region (ref: Toronto)
Montreal	−0.143***	−0.024	−0.039	0.002	−0.058	0.140*	−0.199***	−0.127
Vancouver	0.021	0.000	−.019	−0.013	−0.066	−0.115	−0.271***	−0.139**
Second-tier cities	0.000	0.022	0.023	0.011	0.033	0.063	−0.135***	−0.059
Small metropolitan areas	−0.075***	−0.026	0.035	0.041	−0.253***	−0.019	−0.101	0.037
Small urban or rural areas	−0.091***	−0.025	0.042	0.031	−0.115*	0.053	−0.216***	−0.099
Men (ref: women)		0.148***		0.128***		0.242***		0.298***
Age		0.037***		0.014		0.062***		0.020
Age Squared/100		−0.050***		−0.018		−0.091**		−0.032
Education (Ref: university)
Less than high school graduation		0.148***		0.011		0.266**		−0.196
High school graduation		0.097**		−0.036		0.162		−0.182
Some post-secondary education		0.088*		−0.012		0.186*		−0.261*
Speak official language (ref: not)		0.066***		0.147***		0.104**		0.097*
Marital status (ref: married)
Single		0.087***		−0.061***		−0.038		−0.096**
Divorce, separated or widowed		0.053		−0.145***		−0.378***		−0.012
% of own-ethnic members in a region		0.021***		0.004		0.022*		0.000
Regional employment rates		0.013***		0.012***		…		…
Weeks worked		…		…		0.037***		0.024***
Full time (ref: part time)		…		…		0.320***		0.321***
Regional average earnings		…		…		0.545***		0.558*
Sample size	4124	4124	3829	3829	2684	2684	3244	3244
Model R squared	0.015	0.077	0.003	0.075	0.01	0.499	0.013	0.288

Note: … not included. * significant at p < 0.05, ** p < 0.01, ***p < 0.001.

Sources: Author's calculation based on data from the 1981, 1991 census.

Model 2 for 1981 in Table 3 also shows that higher employment rates were associated with men (relative to women), increased age (at least in the younger age range), less than university education, speaking an official language, and not being married. The effect of living in an ethnic enclave is statistically significant and large. A one percentage point increase in the share of own-ethnic group members in a region was associated with a 2.1 percentage point increase in the employment rate. The effect of regional employment rates was also significant and substantially large.

Regression decomposition (Hou 2014) results show that differences in regional economic conditions, as measured by the employment rate of prime-aged Canadian-born men, accounted for 59 per cent of the gap in employment rates of Vietnamese refugees in Montreal relative to Toronto, while the effect of living in an ethnic enclave accounted for another 22 per cent (the detailed decomposition results are available on request). For the gap in employment rates between Montreal and Vancouver, the ethnic enclave variable played a relatively larger role (about 54 per cent) than regional employment rates (about 31 per cent). For the differences between smaller metropolitan areas and Toronto, and between small urban or rural areas and Toronto, regional employment rates accounted for about 60 per cent, while ethnic enclave accounted for about 40 per cent.

In 1991, the regional variation in the employment rates of Vietnamese refugees became smaller relative to 10 years earlier as employment rates rose considerably in Montreal, small metropolitan areas, and small urban or rural areas, but decreased in Toronto and Vancouver. This was likely because the recession in 1991 affected Toronto and Vancouver more severely than other places since similar changes were also observed among the Canadian-born population. It could also relate to selective migration of less successful immigrants to Toronto and Vancouver.

In the models predicting employment of Vietnamese refugees in 1991, sex, official language ability, marital status (signs changed), and regional employment rates remained significant explanatory variables. The effect of ethnic enclave became not significant, likely suggesting its diminished role in helping longer-term refugees. The full model accounted for the entire 3.9 percentage point gap in employment rate between Montreal and Toronto, and regional employment rates alone accounted for 70 per cent of the difference.

4.1.3. Annual earnings

While finding a job is the first major step towards self-sufficiency of refugees, the quality of the job, particularly the earnings levels, determines the level of economic wellbeing gained from engaging in the labour market. In this regard, Vietnamese refugees did poorly relative to other immigrants and the Canadian-born population, mostly in the first 10 years. Table 2, right panel, presents median annual earnings among individuals who earned at least CA$500 of employment income (in 2010 constant dollars) for prime-aged Vietnamese refugees, other immigrants, and the similar-aged Canadian-born population. In 1980, Vietnamese refugees earned about 30 per cent less than other immigrants.

This large gap in 1981 likely resulted from at least three possible sources: (1) some refugees were still receiving financial support from either the government or private sponsors, and many refugees were still in language training. Thus they worked fewer hours than other immigrants or the Canadian-born population. (2) More of them were located in small metropolitan areas and small urban or rural areas where earnings were relatively low also for the Canadian-born population. (3) They had lower language ability and education levels. A regression analysis showed that the three factors could account for about half of the earnings gap between Vietnamese refugees and other immigrants in 1981, with regional economic conditions (as measured by average earnings of prime-aged Canadian-born men by education) and weeks worked playing more important roles than group differences in human capital factors.

The gap of Vietnamese refugees in annual earnings relative to other immigrants narrowed to about 15 per cent 10 years after arrival in 1991. The gap continued to narrow to 6 per cent 20 years after arrival in 2001. Thus, although Vietnamese refugees started with a large initial earnings gap relative to other immigrants, their earnings growth rate was much faster, and the gap became small after 20 years in Canada.

In terms of the regional variation, Vietnamese refugees in 1980 had much lower median annual earnings in small metropolitan areas and small urban or rural areas than in Toronto (Table 4). These gaps, particularly the one with small metropolitan areas, were large and statistically significant, as shown in regression Model 1 in Table 3 (right panel) for log earnings of Vietnamese refugees. Model 1 contains dummy variables for geographic regions, with the Toronto metropolitan area as the common reference. To examine factors that may account for these regional differences, Model 2 adds sex, age, education, official language, marital status, ethnic enclave, weeks worked in the year, full-time versus part-time status, as well as regional economic conditions as measured by the log of average annual earnings of prime-aged Canadian-born men by education.

Table 4. Educational attainment and earnings among Vietnamese childhood refugees and other childhood immigrants who arrived in 1979–1980.

	High school dropout rate		University completion rate		Median annual earnings
	2001	2011	2001	2011	2000	2010
Vietnamese refugees	percent		percent		2010 constant dollars
Overall	18.3	13.7	28.3	35.7	34,200	48,700
Montreal	8.1	13.4	42.9	42.4	36,600	47,300
Toronto	20.0	11.0	31.4	37.6	39,100	50,000
Vancouver	22.8	15.1	15.5	25.8	26,900	44,900
Second-tier cities	17.3	17.9	25.2	38.9	30,500	53,600
Small metropolitan areas	17.9	9.1	26.5	28.5	26,900	44,400
Small urban or rural areas	22.3	18.7	26.3	17.2	33,900	39,600
Immigrants from other countries
Overall	15.0	8.7	24.7	31.5	30,500	46,800
Montreal	12.9	8.9	23.6	31.5	25,600	37,600
Toronto	13.2	7.8	28.9	33.9	36,600	50,300
Vancouver	10.9	6.7	27.6	32.6	32,700	47,700
Second-tier cities	15.6	7.2	25.3	34.4	30,500	53,200
Small metropolitan areas	13.9	7.5	20.4	30.5	26,900	44,000
Small urban or rural areas	23.8	16.2	15.4	21.2	26,900	41,200
Similar-aged Canadian-born
Overall	18.1	10.3	18.9	25.6	28,100	44,300
Montreal	13.8	9.3	23.5	30.6	29,300	43,100
Toronto	11.8	6.0	30.9	39.9	36,600	53,300
Vancouver	13.6	6.0	23.9	33.6	31,800	48,300
Second-tier cities	15.2	7.9	23.3	31.7	29,300	50,500
Small metropolitan areas	15.5	8.3	19.5	24.7	26,900	44,000
Small urban or rural areas	24.3	15.1	11.0	15.2	24,700	38,300

Sources: Author's calculation based on data from the 2001 census and 2011 NHS.

Model 2 for 1981 shows that speaking an official language was associated with 10 per cent higher earnings, while higher level of education was not associated with higher earnings. Weeks worked, full-time status, local concentration of ethnic group members, and regional average earnings were strong predictors of annual earnings of Vietnamese refugees. When all the covariates are included, the earnings gaps associated with small metropolitan areas and small urban or rural areas relative to Toronto became not significant. Further decomposition results suggest that fewer weeks worked and regional economic conditions were the key factors accounting for the observed gaps.

Ten years later (in 1990), large differences emerged among three gateway centres. Vietnamese refugees had much higher earnings in Toronto than in Montreal and Vancouver (Table 2). Similar regional differences, although smaller between Toronto and Vancouver, were also observed for other immigrants and the Canadian-born population. This suggests that the regional differences observed among Vietnamese refugees were at least partly driven by regional economic conditions. The regression models for log earnings in 1990 show that much of the observed earnings gaps in Vancouver and Montreal relative to Toronto are accounted for by the included covariates. Further decomposition analysis indicates that weeks worked and regional economic conditions played the major roles. Model 2 for 1990 also shows that ethnic enclave is no longer a significant predictor of earnings among Vietnamese refugees, while weeks worked, full-time status, and regional average earnings are strongly associated with refugees’ earnings.

Similar regional differences in earnings of Vietnamese refugees persisted 20 (in 2000) and 30 (in 2010) years after their arrival, with Vietnamese in Toronto having higher earnings than in Montreal and Vancouver. Similar regional differences are observed among other immigrants and the Canadian-born population in the corresponding years.

4.2. Childhood refugees

Many refugees risked their lives to flee persecution, violence, and hardship, not just for a safe place to survive for themselves, but mostly to find an environment for their children to have a brighter future. In this regard, the socioeconomic outcomes of childhood refugees are important indicators of long-term integration of Vietnamese refugees in Canada.

Table 4 presents the high-school dropout rates, university completion rates, and median annual earnings of refugees and other immigrants who arrived at age 17 or younger. The outcomes are measured when they reached ages 20–37 in 2001 and 30–47 in 2011. These outcomes are also presented for similar-aged Canadian-born persons in the respective years.

The high-school dropout rates and university completion rates of childhood Vietnamese refugees relative to other immigrants and the Canadian-born population reveal a bifurcation phenomenon that has been observed among Vietnamese youth in the US (Bankston and Zhou 1997). On the one hand, childhood Vietnamese refugees were more likely to have dropped out of high school than other immigrants or the similar-aged Canadian-born population, particularly in 2011. On the other hand, despite their parents’ generally low levels of education, childhood Vietnamese refugees who finished high school were more likely to complete university than other immigrants and the Canadian-born population. For instance, in 2011, close to 36 per cent of childhood Vietnamese refugees had finished a university degree at age 30–47, compared with 32 per cent among other childhood immigrants and 26 per cent among the similar-aged Canadian-born population.

Corresponding to their higher university completion rates, childhood Vietnamese refugees had higher median earnings than other childhood immigrants and the similar-aged Canadian-born population, particularly in 2001. However, at the group level, other childhood immigrants and the Canadian-born population had a faster earnings growth in the following decade. As a result, childhood Vietnamese's lead in median earnings over other immigrants and the Canadian-born population became much smaller by 2011.

There were regional variations in education and annual earnings patterns. Childhood Vietnamese refugees had the highest university completion rate and lowest high-school dropout rate in Montreal in both 2001 and 2011. The advantage in Montreal was consistent with a much higher educational level and official language ability among Vietnamese refugee parents in Montreal relative to other regions, as shown previously in Table 1. Previous Canadian studies suggest that parents’ education and language ability are key determinants, more important than family income, of the educational attainment among the children of immigrants and refugees (e.g. Hou and Bonikowska 2017). Childhood Vietnamese refugees in Vancouver had the highest high-school dropout rate and lowest university completion rate in 2001, when they were age 20–37, and had considerably lower educational levels than their counterparts in Montreal and Toronto when they reached age 30–47 in 2011. It is not clear why there was a large difference between Vancouver and Toronto. Vietnamese refugee parents in Vancouver and Toronto had similar official language ability and educational levels. Vancouver had a larger Chinese community than Toronto, so the effect of ethnic enclave could be a possible explanation. However, childhood Vietnamese refugees with Chinese and Vietnamese ethnic origins had similar university completion rates in Vancouver. This suggests that the existing Chinese enclave was not directly associated with a disadvantage for those with Chinese ethnic origin.

Although childhood Vietnamese refugees in Montreal had an advantage in educational attainment relative to their counterparts in Toronto, their annual earnings were lower in Montreal than in Toronto in both 2000 and 2010. An even larger difference in annual earnings between Toronto and Montreal existed among other childhood immigrants and similar-aged Canadian natives, implying that Toronto had an advantage in the wage structure of the economy. It seems that, for childhood Vietnamese refugees, Toronto's large advantage in economic conditions trumped Montreal's moderate advantage in educational attainment. However, childhood Vietnamese refugees had higher earnings in Montreal than in Vancouver, while the opposite held for other childhood immigrants and the Canadian-born population. This suggests that, for childhood Vietnamese refugees, Vancouver's moderate advantage in economic conditions was not enough to offset the group's large gap in educational attainment and the possible negative effect of ethnic concentration.

5. Conclusion and discussion

Welcoming 60,000 Southeast Asian refugees in the 1979–80 period has become a celebrated part of Canada's history, but the eventual integration of these refugees into Canadian society has received insufficient attention. This study provides a comprehensive overview of Vietnamese refugees’ economic outcomes over the three decades after their arrival. This study also explores how regional contexts contributed to shaping economic outcomes.

Adult Vietnamese refugees arrived with little human capital. Only a small proportion spoke one of Canada's official languages, and most had not finished high school. Remarkably, within 1–2 years after arrival, the majority of adult Vietnamese refugees acquired some official language fluency. Clearly, they took full advantage of the language training opportunities offered by the government. Some also upgraded their education, although their high-school graduation rate remained much lower than other similar-aged immigrants and the Canadian-born population. In spite of deficiencies in human capital, adult Vietnamese refugees were successful enough in finding jobs that, within a year or two after arrival, they were more likely than other immigrants to be employed. In the initial years, adult Vietnamese refugees worked primarily in low-quality jobs, as reflected in large earnings gaps between the refugees, other immigrants, and similar-aged Canadian-born workers. However, they experienced faster earnings growth, and their earnings gap became small after 20 years in Canada.

The refugees’ low educational attainment was an apparent source of continuing vulnerability. For example, during the recession in the early 1990s, the refugees were more likely to be unemployed than their better-educated immigrant and native-born counterparts.

By 20–30 years after arrival, with the economy restored to more normal functioning, the refugees were as likely to be employed as other immigrants. Both refugees and immigrants had higher employment rates than native-born Canadians of the same age.

Vietnamese refugees with a university education faced a unique challenge in their initial years of resettlement: they had lower employment rates and earnings than their less well-educated counterparts. This finding is consistent with the observation from a different survey on Southeast Asian refugees (Beiser and Hou 2001). Hartog and Zorlu (2009) also find that higher educations acquired in the home country generally do not pay off for refugees during the first five years in the Dutch labour market. Similar experiences are shared by some other refugee groups (e.g. refugees from Somalia, Afghanistan, and Pakistan) who arrived in Canada more recently than Vietnamese refugees (Picot, Zhang, and Hou 2019). These results are very different from the strong positive effect of education on labour market outcomes for other refugee groups, immigrants in general, and non-immigrants. One possible explanation for the little returns to education among Vietnamese refugees and other refugee groups in the initial years of resettlement relates to the complementarity between language and educational qualification (Berman, Lang, and Siniver 2003). Previous studies have shown that immigrants without proficiency in the official language of the receiving country benefit little from their home-country education (Bonikowska, Hou, and Picot 2015). However, as refugees and immigrants improve their language ability, the values of home-country education tend to increase with the length of residence in the receiving country (Picot, Hou, and Qiu 2016), as also observed among Vietnamese refugees.

Childhood Vietnamese refugees achieved greater success than their adult counterparts. In spite of the hardships of the refugee experience and limited family resources, childhood Vietnamese refugees who went on to post-secondary education were much more likely to complete a university degree than other childhood immigrants or their Canadian-born counterparts. However, they also had a higher high-school dropout rate. These results parallel findings from the US, where Vietnamese American youth were over-represented among both high achievers and the disadvantaged (e.g. Bankston and Zhou 1997).

Probably because of their high educational attainment, childhood Vietnamese refugees had much higher annual earnings than other childhood immigrants and the similar-aged Canadian-born population in the early stage of their working careers (ages 20–37). However, their lead in earnings narrowed during the following decade. Future studies should examine whether the trend continues into the later stage of their work career; and if so, whether that may be due to differences in occupational distributions or to structural factors limiting upward mobility.

The geographic region of residence was associated with some large differences in refugees’ socioeconomic outcomes, but the patterns varied by the outcome measure and length of residence. Regional differences in refugees’ human capital characteristics, ethnic enclave, and general economic conditions played different roles. Over time, the 1979–80 cohort of adult Vietnamese refugees in Montreal achieved a much higher rate of speaking an official language and higher educational level than their counterparts in Toronto and Vancouver. Many refugees who initially settled in Montreal were likely the relatives of first-wave Vietnamese refugees who mostly were the privileged population back in Vietnam. Conversely, the existing ethnic enclaves in Vancouver and Toronto were more likely to receive and retain refugees with limited human capital. The regional differences in educational attainment among adult refugees were passed on to, or even amplified among, childhood refugees, as manifested by a very large difference in university completion rates between Vancouver and Montreal. The large regional differences in educational attainment among childhood refugees likely reflect the predominant influence of parents’ educational attainment on their children's educational outcomes, and the role of ethnic enclave in reducing the need or motivation to pursue higher education.

Despite a comparative deficit in human capital, adult refugees in Toronto had consistently higher employment rates and annual earnings than those in Montreal, probably because of a prevailing economic advantage enjoyed by Toronto. The working careers of refugees who arrived as children paralleled those observed among refugee adults, suggesting that large regional differences in general economic conditions may trump moderate regional differences in refugees’ human capital characteristics. These findings are consistent with previous reports that immigrants experience little disadvantage when there is strong labour demand, but face large disadvantages under poor economic conditions (e.g. Hou 2013). Like-ethnic concentration had positive effects on adult refugees’ employment rates and earnings, but the effect was significant only in the initial years after arrival.

Integration into the labour market and making an economic contribution to the receiving society are as important to refugees as to other immigrants. Despite limitations in human capital at arrival, Vietnamese refugees who entered the country as adults achieved high levels of employment and closed initial earnings gaps with other immigrants during the 30 years after coming to Canada. Refugees who came to Canada as children out-performed their immigrant and native-born similar-age counterparts.

Canada admits refugees because of a sense of moral obligation and because it belongs to a group of nations that have committed themselves to save the oppressed and vulnerable. Refugees are not expected to make economic contributions as much as economic immigrants. Nevertheless, the evidence suggests that Vietnamese refugees did not become an economic burden; rather, over time, they contributed to Canada's economy much like other immigrants who arrived in the same period. Overall, the experience of Vietnamese refugees in Canada provides an illustrious showcase that a highly vulnerable yet exceedingly striving immigrant group successfully integrated into a welcoming host society over time and across generations.

Acknowledgements

The author would like to thank Morton Beiser, Rachel Gisselquist, Carla Valle Painter, Garnett Picot, participants at the UNU-WIDER workshop on Forced Migration and Inequality, and the journal reviewers for advice and comments on an earlier version of this paper. Any errors are the responsibility of the author.

Disclosure statement

No potential conflict of interest was reported by the author(s).