Journal of International Migration and Integration

, Volume 14, Issue 4, pp 689–710

Human Capital and the Economic Assimilation of Recent Immigrants in Hong Kong

Authors

    • The Chinese University of Hong Kong
  • Suet-ling Pong
    • The Chinese University of Hong Kong and Penn State University
Article

DOI: 10.1007/s12134-012-0262-4

Cite this article as:
Ou, D. & Pong, S. Int. Migration & Integration (2013) 14: 689. doi:10.1007/s12134-012-0262-4

Abstract

Using data from the 1996, 2001, and 2006 years of the Hong Kong Population Census, this paper reported the nativity earnings gap among a synthetic cohort of immigrant and native male Chinese employees in Hong Kong. Consistent with previous research, we found earnings divergence for all workers. However, this earnings divergence masked a reverse trend for low-skilled workers. Over time, low-skilled immigrant workers gained earnings assimilation with low-skilled native workers, but high-skilled immigrant workers did not gain assimilation with high-skilled native workers. A decomposition analysis suggested that the relative skill prices cannot explain the overtime change in the relative mean-earnings gaps by nativity. Further separating pre- and postmigration education of immigrants did not improve the explanatory power of the relative skill prices. Our results for Hong Kong are consistent with the findings from recent research on the economic assimilation of low-skilled immigrants in other countries.

Keywords

ImmigrantsEarnings differentialsReturns to skillsHong Kong

Introduction

Previous studies (Lam and Liu 2002a, b) found that economic assimilation—the convergence of earnings between immigrant and native workers—was absent in Hong Kong during the decade of 1981–1991, as immigrant workers earned increasingly less than did native workers. This earnings divergence was very different from other major immigrant countries, such as the United States, where immigrant workers tend to improve their economic position over time relative to native workers (Borjas 1985, 1995; Chiswick 1978). The situation is even more puzzling because the lack of economic assimilation in Hong Kong is found among Chinese immigrants from mainland China, who are largely of the same racial, ethnic, and cultural heritages as Hong Kong natives. The human capital of Chinese immigrants, as measured by their educational attainment, has not only been on the rise since the decade studied, but the quality of mainland Chinese students’ education has also improved substantially, as shown by the outstanding achievement of Shanghai students in the 2009 Programme for International Students’ Assessment. Studies on secondary-school students have found that immigrant students in secondary schools tend to attain higher test scores than do native students in all subjects except the English language (Pong and Tsang 2010), further suggesting that the education quality in mainland China may not be inferior to that in Hong Kong. Thus, it is important to understand whether the previous findings on earnings divergences by nativity were specific to a historical period or represent a general trend.

Most studies on immigrant workers’ assimilation have treated immigrants as one type of labor and have not examined the heterogeneity within this subgroup. On the one hand, many immigrant workers are low-skilled; they cross borders to seek higher pay for their labor. On the other hand, immigration policies such as the 1990 Immigration Act in the United States, which gave preference for economic immigrants based on skills and qualifications, encouraged high-skilled individuals to immigrate. The patterns of assimilation likely differ between high- and low-skilled workers. Because the concept of economic assimilation entails a comparison between immigrants and natives, a better approach is to examine the nativity gap within skill groups in which immigrants and natives can be regarded as perfect substitutes (Card 2005). In the US, high-skilled immigrant workers tend to experience occupational downgrading (Akresh 2006), and some of them face the glass ceiling, unable to enjoy the same remuneration for their work as do high-skilled native workers (Zeng and Xie 2004). By contrast, low-skilled immigrant workers in the United States have been found to obtain significant gains over time in earnings relative to low-skilled native workers (Hall and Farkas 2008). Differential returns to skills would mean differential achievement in the labor market for high- and low-skilled immigrant workers, relative to their native counterparts. In this paper, we present evidence of systematic variations in economic assimilation by skill levels, which can be masked by the overall immigrant–native earnings differentials.

This study had three main objectives. The first concerned earnings inequality by immigrant status in Hong Kong. We examined whether the earnings divergence found in 1981–1991 between Hong Kong natives and immigrants from mainland China (Lam and Liu 2002a) remained in more recent years. The second objective was to investigate potential heterogeneity of the nativity earnings gap by skill level. We explore whether high- and low-skilled workers exhibited different assimilation patterns. Third, through a decomposition analysis, we analyzed how returns to immigrants’ pre- and postmigration human capital contributed to the recent trends in the nativity earnings gap.

Hong Kong Context

As one of the world’s financial centers, Hong Kong enjoys a real GDP growth rate of about 6.8 % (Bureau of East Asian and Pacific Affairs, 2011). Hong Kong’s economic success is due in large part to the human resources of its immigrant population. According to data from recent years of the Hong Kong Census, about one third of its population was born in mainland China in 1996, 2001, and 2006 (Hong Kong Census and Statistics Department 2006). Hong Kong has a continuous population inflow from the mainland. Illegal immigrants started to arrive after a quota system was in place in 1950. They continued coming through the 1960s. First adopting a lassie faire policy toward illegal immigration, the Hong Kong government later allowed illegal immigrants to become permanent residents by way of the “reach-base policy,” which was implemented in 1974. Only until illegal immigration reached sky-high levels in late 1970s, during political turmoil in the mainland, did Hong Kong take serious measures to curb it. In 1980, the reach-base policy was abolished. In 1983, with consent from the Chinese government, Hong Kong sealed its border with China, enforced deportation of illegal immigrants, issued Hong Kong identity cards, and imposed fines on businesses who hired illegal workers (Lam and Liu 1998).

Since 1983, Hong Kong has been admitting legal immigrants daily from China, and the status of Chinese immigrants has changed from primarily illegal to primarily legal. After Hong Kong’s handover to China in 1997, immigrants from the mainland to Hong Kong are no longer “underground” people of a different country. Many speak the same language and enjoy the same government benefits as native Chinese. In 2003, the Admission Scheme for Mainland Talents and Professionals was implemented primarily to encourage high-skilled individuals from the mainland to work in Hong Kong. Most mainland immigrants who were recruited under this policy were university-educated professionals or accomplished athletes, artists, and musicians. The policy objective was to fill needed professional positions to enhance Hong Kong’s status and competitiveness with other countries (Hong Kong Immigration Department 2011). This policy has boosted the overall education levels among immigrant workers.

It is worth noting that Lam and Liu’s (2002a) finding of earnings divergence between immigrant and native workers was based on an artificial cohort of adult workers who reported in 1981 that they had resided in Hong Kong within the preceding 5 years. The sample represented a specific cohort of immigrants who most likely entered Hong Kong illegally. Thus, Lam and Liu’s finding of economic divergence was based largely on Chinese immigrants who entered Hong Kong without legal documents. It is necessary to investigate a more recent period of time to determine whether Lam and Liu’s findings also apply to the more typical immigrants today from mainland China through legal means. In this paper, we examine the nativity earnings gap between immigrant and native Chinese male workers using newer Hong Kong census data from 1996, 2001, and 2006. We examined the economic assimilation of male employees of different skill levels as defined by their educational attainment, focusing on both high- and low-skilled employees. Our results provide evidence of an earnings convergence among low-skilled Chinese immigrants in Hong Kong, which is consistent with the convergence of the nativity gap in the returns to education among low-skilled workers. By contrast, the nativity earnings gap enlarged over time among high-skilled workers, despite the fact that relative skill prices were actually in favor of highly educated Chinese immigrants. These results point to differential patterns of economic assimilation for high- and low-skilled workers.

Data and Summary Statistics

We used data from the 1996, 2001, and 2006 years of the Hong Kong Population Census 20% samples.1 Consistent with previous literature on immigrants’ assimilation in Hong Kong, the samples were restricted to Chinese men who were employees and had nonzero earnings at the time of interview. It is worth noting that the Chinese ethnicity made up 95 % of Hong Kong’s population in 2006.2 We constructed an artificial cohort of immigrants based on their place of birth, age, and duration of residence in Hong Kong. The immigrant cohort includes Chinese immigrants who were aged 15 to 50 years in 1996 and were new to Hong Kong, residing in Hong Kong for no more than 5 years (0–5) at the time of interview. This cohort was observed in three time points that spanned 15 years: 1996, 2001, and 2006. The individuals were 20 to 55 years old in 2001, and they had resided in Hong Kong for 5 or more years but no more than 10 (5–10) years in 2001. Similarly, in 2006, they were 25 to 60 years old and had resided in Hong Kong for ten or more years but no more than 15 (10–15) years. The age restriction of 15 to 50 years in the base year was appropriate for our study of employees because 15 years was the minimum age required to work legally in Hong Kong,3 and the upper age limit eliminated the potential bias of attrition due to retirement at the final time point of observation.4 We included data from natives of the same age as a comparison group.

Table 1 shows the percentage distribution of education levels by immigrant status in the base year of 1996. The sample was divided into three skill groups according to their highest levels of completed education.5Low-skilled workers were defined as those who had attained no more than lower secondary school. Middle-skilled workers were defined as those who had completed upper secondary school. High-skilled workers are defined as those who had attained some postsecondary education or possessed credentials such as a higher diploma, first degree, or postgraduate degree. Unlike samples of immigrant workers in previous studies (e.g., Capps et al. 2003), the cohort of immigrants in our study clearly had more education than natives. About 46 % of immigrants were high-skilled workers, whereas barely 30 % of native workers were in the high-skilled category. In a descriptive analysis not reported here, we found that 30 % of immigrant workers had a first degree or postgraduate degree. The corresponding figure for native workers was less than 15 %.
Table 1

Percentage distribution of education levels (1996) and average years of education of Chinese males aged 15–50 years by immigrant status (1996–2006)

Education level in 1996

Immigrants

Natives

All

     

Low-skilled

41.07

37.19

37.30

     

Middle-skilled

20.84

34.21

33.86

     

High-skilled

38.09

28.59

28.86

     

Total

100

100

100

     

N

4760

174,116

178,926

     

Years of education

 

All

Low-skilled

High-skilled

Middle-skilled

 

Immigrants

Natives

Immigrants

Natives

Immigrants

Natives

Immigrants

Natives

1996

11.19

10.72

7.57

7.29

15.54

15.08

10.80

10.89

2001

10.62

11.04

7.74

7.23

14.83

15.10

10.84

10.91

2006

10.82

11.48

7.60

7.31

14.92

15.35

10.83

10.91

High-skilled is defined as having an education of Form 6 or Form 7 or above. Middle-skilled is defined as having at least lower secondary education but no more than a high school degree. Low-skilled is defined as having a lower secondary education or below. N is the total number of observations. Source: Self-help Tabulation Service data sets of the Hong Kong Population Censuses/By-censuses

When education was defined by the number of years in school, immigrant workers as a whole had on average 11 years of education. Overall, native workers had about half a year less education than immigrant workers in 1996. Native workers caught up and were slightly more educated than immigrant workers in 2001 and 2006. This overall trend by nativity appears to reflect the trend among high-skilled workers. One possible reason is sample selection by labor force participation among these workers. However, our exploratory analysis suggests that labor force participation rates were about the same for both nativity groups in each census year (see Appendix Table 1). A more plausible reason is the different exit and entry rates of immigrant and native workers. Hong Kong has been considered a stepping stone for highly educated immigrants from mainland China. Some high-skilled immigrant workers in the 1996 cohort may have left Hong Kong for other countries in subsequent years, hence the decreasing trend of immigrants’ educational attainment. With respect to native workers, an exodus of highly educated Hong Kong natives occurred before 1997, the year Hong Kong reunified with China. After the handover, many of these native workers returned as the political situation of Hong Kong became more stable. This may explain the increasing trend of natives’ educational attainment after 1996.

Panel A in Table 2 shows the average earnings of immigrant and native employees for the full sample. All earnings are deflated at the 2006 level. Immigrant employees earned less than native employees, and this earnings gap increased over time. Immigrant employees earned about 84 % of what their native counterparts earned during their first 5 years in Hong Kong (1996–2001). The nativity gap was 16.2 % of native workers’ average earnings. The earnings gap was enlarged in 2001 and 2006 as the cohort of immigrants stayed in Hong Kong for 5–10 and 10–15 years, and earned about71 and 63 % of what natives earned 5 and 10 years after, respectively. The corresponding nativity gaps were approximately 29 % and 37%of native workers’ average earnings, indicating earnings divergence.
Table 2

The nativity warnings gap by skill levels

Duration in Hong Kong

Immigrants

Natives

Raw nativity gap

Nativity gap as % of native’s earnings

All male workers

0–5 years

12,099.99

14,436.47

2,336.48

16.18 %

5–10 years

12,747.51

17,993.80

5,246.29

29.16 %

10–15 years

11,603.03

18,475.05

6,872.02

37.20 %

High-skilled male workers

0–5 years

18,487.22

22,752.20

4,264.98

18.75 %

5–10 years

20,075.56

28,374.43

8,298.87

29.25 %

10–15 years

15,909.84

27,689.27

11,779.43

42.54 %

Low-skilled male workers

0–5 years

7,675.03

9,948.29

2,273.27

22.85 %

5–10 years

9,048.88

11,194.03

2,145.15

19.16 %

10–15 years

9,422.04

10,968.79

1,546.75

14.10 %

All earnings figures are at the 2006 price level. Self-help Tabulation Service data sets of the Hong Kong Population Censuses/By-censuses

However, when the sample was split into three groups by level of worker skill, the results of the nativity earnings gap were quite different. The initial earnings gap by nativity among high-skilled workers (panel B) was about 19 % of the native workers’ average earnings, and this gap increased over the years of immigrants’ residence in Hong Kong. In contrast, the nativity earnings gap for low-skilled male (panel C) immigrants decreased over the 10-year period. Thus, we found earnings divergence for high-skilled workers but earnings convergence for low-skilled workers for the 15-year period from 1996 to 2001. The middle-skilled group, not reported here, exhibited divergence at first and then convergence in earnings. Because workers in this group were homogenous in their education level, they were not further analyzed.6

Distribution of Immigrants in Industries and Occupations

Workers’ earnings are tied to their skill levels, which, in Hong Kong, are generally segregated by industries and occupations (Liu et al. 2004). This dynamic was true in the study data as well (see Table 3). Columns 2, 6, and 10 show the distributions of immigrants across industries in 1996, 2001, and 2006, respectively. Similar distributions can be found in columns 4, 8, and 12 for natives. The column labeled “% of managers and professionals within industry”7 shows the proportion, within various industries, of natives or immigrants whose occupations are managers and professionals. In 1996, the three top industries in which immigrant workers occupied were manufacturing, utilities and construction, and wholesale. The top three industries for native employees also included manufacturing and utilities and construction, but the third industry for native workers was transportation and communication. In subsequent years, however, immigrant workers appeared to follow the footsteps of native workers. In 2001, the concentration of immigrant workers in wholesale industries declined, whereas their concentration in the transportation-and-communications industry rose. Meanwhile, many native workers moved into business services. In 2006, business services became one of the top three industries for immigrant workers as well. It is worth noting that native workers tended to concentrate in industries that offered higher wages,8 such as business services. Since immigrant workers’ distribution in industries follow that of native workers in the lag year, it is plausible that immigrant workers were slow to be accepted into industries with higher wages.
Table 3

Distribution of male workers in industries and occupations by time and immigration status

 

1996

2001

2006

 

Immigrants

Natives

Immigrants

Natives

Immigrants

Natives

Industry

% of managers and professionals within industry

Total

% of managers and professionals within industry

Total

% of managers and professionals within industry

Total

% of managers and professionals within industry

Total

% of managers and professionals within industry

Total

% of managers and professionals within industry

Total

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

1

0.00

0.29

5.79

0.34

0.00

1.30

3.42

0.26

0.00

0.28

10.67

0.20

2

28.15

26.82

38.73

20.79

27.63

16.68

47.04

13.87

23.44

9.69

47.84

11.45

3

16.94

15.50

25.56

11.29

8.89

15.78

30.02

8.08

1.62

28.62

27.69

10.56

4

93.25

11.18

72.68

6.04

76.09

9.22

71.58

7.69

55.91

5.89

73.96

7.71

5

29.58

7.66

50.91

7.09

20.73

4.11

54.03

6.40

50.00

6.49

53.48

6.31

6

47.66

8.45

21.40

17.14

32.09

16.08

23.44

19.06

13.62

11.92

22.55

19.97

7

100.00

4.53

91.49

5.82

100.00

3.86

92.74

6.36

77.50

1.86

92.96

5.79

8

66.17

7.23

71.67

10.44

48.99

9.92

69.94

14.02

37.37

13.78

56.65

16.57

9

77.42

1.11

64.19

3.84

-

0.00

64.73

4.58

33.33

1.39

21.15

1.57

10

0.00

0.76

12.21

0.56

0.00

0.75

13.60

0.65

0.00

1.39

6.95

1.26

11

93.55

7.80

84.92

7.21

71.56

5.46

80.22

8.30

52.94

6.31

77.12

8.69

12

100.00

0.25

66.74

2.16

51.52

1.65

65.95

1.67

-

0.00

61.83

1.99

13

6.86

3.67

12.43

3.23

0.00

4.01

9.57

2.51

0.00

2.32

13.36

2.58

14

21.21

3.46

21.68

4.04

6.73

11.17

22.20

6.56

4.61

10.06

25.32

5.35

Total

 

100.00

 

100.00

 

100

 

100

 

100

 

100

Industry codings: 1 agriculture and mining, 2 manufacturing, 3 utility and construction, 4 = wholesale (including export/import), 5 retail, restaurants, hotels, 6 transportation (including storage) and communication, 7 finance, 8 business services, 9 public administration, 10 sanitary and similar services, 11 social and related community services, 12 amusement and recreational services, 13 personal services, 14 = others. Source: Self-help Tabulation Service data sets of the Hong Kong Population Censuses/By-censuses.

However, even though manufacturing remained the biggest industry for both natives and Chinese immigrants in 1996 and one of the top three industries in subsequent years, the proportion of managerial and professional occupations within that industry was larger for natives than it was for immigrants. For example, 48 % of natives employed in the manufacturing industry in 2006 were managers and professionals, compared with 23 % of immigrants. The business service sector is another example. Although more immigrants than natives were employed in business services, the proportion of immigrants holding managerial and professional jobs fell from 66 % in 1996 to 37 % in 2006, while the proportion of managers and professionals for natives remained relatively high, at about 57 % to 72 %. These large differences between the industry and occupation distributions of immigrant and native workers likely confounded our analysis of skill prices on the earnings gap by nativity, so the statistical models were adjusted to include these factors.

Returns to Education and Decomposition of Earnings Differences

The economic literature on immigrants’ economic assimilation has focused on two aspects of immigrant human capital measured by their education. The first is the amount (quantity) of human capital that immigrant workers bring from their home countries, which has been referred to in the literature as the “quality” of immigrants. For example, Borjas (1985) argued that the 1960s immigration reform in the United States led to a wave of new immigrants of deteriorating quality. As discussed above, the reverse trend has been true in Hong Kong. Due to political changes and immigration policy that favors high-skilled immigrants, the quality of immigrants has been increasing over time.

The second aspect of human capital in workers’ assimilation is the returns to education or “skill price.” Using data from Hong Kong in 1981 and 1991, Lam and Liu (2002a) found that returns to education were the cause of the earnings divergence between immigrant and native male workers. The reason was that human capital obtained locally is more adaptive to technical change in the host country than human capital acquired in the immigrants’ home countries.

To understand the role of human capital on earnings differences between natives and immigrants, we used a two-equation model to decompose the log earnings differences (Lam and Liu 2002a):
$$ \ln w_{it}^h=\alpha_{it}^h+\beta_t^hEdu_{it}^h+\lambda X_{it}^{{h\prime }}+\varepsilon_{it}^h $$
(1)
and
$$ \ln w_{it}^c=\alpha_{it}^c+\beta_t^cEdu_{it}^c+\lambda X_{it}^{{c\prime }}+\varepsilon_{it}^c, $$
(2)
where the superscripts h and c refer to Hong Kong natives and Chinese immigrants, respectively, for individual i at time t. lnw is the logarithm of real monthly earnings from the individual’s main source of employment in 2006 Hong Kong dollars. Edu represents years of formal education. X is a vector for other characteristics, which include years of working experience and its squared term. Working experience is derived by subtracting 6 years from the worker’s age and then further subtracting the number of years of education for all workers who had at least 9 years of education. For workers who had fewer than 9 years of education, their work experience was derived by subtracting 15 years from their age. α and ε are the constant and error term, respectively.

To explore how the distribution of immigrants and natives in various industries and occupations may confound the results, we also include in X a dummy variable of managerial and professional occupations and 13 dummy variables for industries: agriculture and mining; manufacturing; utility and construction; wholesale (including export/import); retail, restaurants, and hotels; transportation (including storage) and communication; finance; business services; public administration; sanitary and similar services; social and related community services; amusement and recreational services; and personal services. “Other” industries were used as the reference group.

Equations 1 and 2 were estimated separately for immigrants and natives for three time points: 1996, 2001, and 2006, defined as Time 1, Time 2, and Time 3. Two sets of estimates were obtained in a stepwise manner, first without and then with the industry and occupation variables. The estimates for each time point, \( \beta_t^h \) and \( \beta_t^c \), represent the returns to education or skill prices of Hong Kong natives and Chinese immigrants at time t, respectively. These coefficients were used in the decomposition analysis.

We examined changes in earnings between two time points (t and t + 1) and decomposed the change in the mean earnings of immigrants relative to natives as shown in Eq. 3:
$$ \begin{array}{*{20}c} \left( {\overline{{\ln w_{i,t+1}^c}}-\overline{{\ln w_{i,t+1}^h}}} \right)-\left( {\overline{{\ln w_{it}^c}}-\overline{{\ln w_{it}^h}}} \right)=\underbrace{{\overline{Edu}_{i,t+1}^c\left[ {\left( {\mathop{{\beta_{t+1}^c}}\limits^{\hat{} }-\mathop{{\beta_{t+1}^h}}\limits^{\hat{} }} \right)-\left( {\mathop{{\beta_t^c}}\limits^{\hat{} }-\mathop{{\beta_t^h}}\limits^{\hat{} }} \right)} \right]}}_{(A) } \hfill \\ +\underbrace{{\left( {\overline{Edu}_{i,t+1}^c-\overline{Edu}_{i,t+1}^h} \right)\left( {\mathop{{\beta_{t+1}^h}}\limits^{\hat{} }-\mathop{{\beta_t^h}}\limits^{\hat{} }} \right)}}_{(B) } \hfill \\ +\underbrace{{\left[ {\left( {\overline{Edu}_{i,t+1}^c-\overline{Edu}_{i,t+1}^h} \right)-\left( {\overline{Edu}_{it}^c-\overline{Edu}_{it}^h} \right)} \right]\mathop{{\beta_t^h}}\limits^{\hat{}}}}_{(C) } \hfill \\ +\underbrace{{\left( {\overline{Edu}_{i,t+1}^c-\overline{Edu}_{it}^c} \right)\left( {\mathop{{\beta_t^c}}\limits^{\hat{} }-\mathop{{\beta_t^h}}\limits^{\hat{} }} \right)}}_{(D) }+other\;terms. \hfill \\\end{array} $$
(3)

Equation 3 contains four different types of education effects: Effect A is the “relative price effect,” which measures the effect of the changes in relative skill prices for immigrants compared with natives. A narrowing difference in the returns to education for immigrants and natives over time would indicate an earnings convergence by nativity. B is the general price effect. If immigrants have more education than natives and if the returns to education increase over time, the earnings gap will converge. C is the relative quantity effect. If the education-level differences between Chinese immigrants and natives do not vary over time, this term would be close to zero.9 Note that the term “effect” is used in a loose manner to refer to the amount of earnings differences that are accounted for. It does not imply any causal relationship. By our definitions of high-skilled and low-skilled workers, we expected that a very small part of the earnings gap would be explained by C for the low-skilled workers because they should have finished their highest level of education by age 15 years. D is the general quantity effect. If immigrants’ mean level of education increases over time, but they are consistently paid less than natives for the same amount of education (i.e., lower returns to education), then the earnings gap by nativity will increase. The “other terms” are a sum of the effects of all other factors that could contribute to the earnings changes in two periods for immigrants and natives, including the constant terms and experience variables (see Eqs. 1 and 2).

The returns to education obtained from Eq. 1 for natives and Eq. 2 for immigrants are reported in Table 4 and illustrated in Fig. 1. For all workers, regardless of skill level (see Fig. 1a), the returns to education for natives were consistently higher than that for immigrants. Without controlling for industries and occupations, the returns clearly diverged. However, the divergence was not obvious after controlling for industries and occupations. A closer examination of high- and low-skilled workers revealed a different pattern. High-skilled immigrants had similar returns to education as their native counterparts at Time 1. They surpassed natives at Time 2 and Time 3 (Fig. 1b). Immigrant workers’ disadvantage at Time 1 was revealed when industries and occupations were controlled (Fig. 1c), suggesting that these high-skilled immigrants entered some relatively high-paying industries or occupations when they first arrived in Hong Kong, which drove up average earnings and the returns to education. Within industries and occupations, however, they earned less than natives with similar education levels. For low-skilled workers, the returns to education were lower for immigrants than for natives at Time 1, but the difference disappeared at Time 2 and Time 3. The occupational and industrial differences did not change the results very much. In other words, when we take into account the type of occupations and industry immigrants occupied, we observed a convergence by nativity of the returns to education for both high- and low-skilled workers. Lumping all skill levels in a general analysis obscured significant heterogeneity.
Table 4

The returns to education by nativity for all male workers, high-skilled workers, and low-skilled workers, with and without industry and occupation variables

 

Time 1

Time 2

Time 3

 

(I)

(II)

(I)

(II)

(I)

(II)

All workers

Native

0.115 (0.000)

0.069 (0.000)

0.133 (0.000)

0.082 (0.000)

0.127 (0.000)

0.074 (0.000)

N

174,166

187,774

177,502

Immigrant

0.091 (0.002)

0.033 (0.003)

0.087 (0.002)

0.040 (0.003)

0.063 (0.002)

0.036 (0.002)

N

4,760

4,399

4,951

Immigrant − native

−0.024

−0.036

−0.046

−0.042

−0.064

−0.038

High-skilled workers

Native

0.147 (0.001)

0.107 (0.001)

0.185 (0.001)

0.140 (0.001)

0.154 (0.001)

0.107 (0.001)

N

49,800

62,314

67,069

   

Immigrant

0.141 (0.006)

0.054 (0.007)

0.209 (0.007)

0.149 (0.009)

0.166 (0.007)

0.103 (0.007)

N

1,813

1,456

1,681

Immigrant − native

−0.006

−0.053

0.024

0.009

0.012

−0.004

Low-skilled workers

Native

0.030 (0.001)

0.025 (0.001)

0.038 (0.001)

0.029 (0.001)

0.039 (0.001)

0.030 (0.001)

N

64,781

62,397

54,823

Immigrant

0.014 (0.005)

0.013 (0.005)

0.036 (0.005)

0.028 (0.004)

0.031 (0.005)

0.026 (0.005)

N

1,955

2,119

2,130

Immigrant − native

−0.016

−0.012

−0.002

−0.001

−0.008

−0.004

Estimates on returns to education are based on Eqs. 1 and 2 for natives and immigrants, respectively. Standard errors are in parentheses. N is the number of observations. The sample includes male employees age 15 years and above with positive wages only. All regressions include years of education, experience, and experience-squared. Estimates in all rows labeled I do not control for industry and occupation; estimates in all rows labeled II do control for industry and occupation. Source: Self-help Tabulation Service data sets of the Hong Kong Population Censuses/By-censuses

https://static-content.springer.com/image/art%3A10.1007%2Fs12134-012-0262-4/MediaObjects/12134_2012_262_Fig1_HTML.gif
Fig. 1

The returns to education by skill levels and immigrant status, with and without taking into account workers’ industry and occupation. All natives (N) are the estimates from Eq. 1 without controls for industry and occupation. All immigrants (N) are the estimates from Eq. 2 without controls for industry and occupation. All natives (Y) are the estimates from Eq. 1 with controls for industry and occupation. All immigrants (Y) are the estimates from Eq. 2 with controls for industry and occupation. Source: Self-help Tabulation Service data sets of the Hong Kong Population Censuses/By-censuses

Table 5 reports the results of the decomposition analysis. As specified in Eq. 3, the change in the mean log earnings gap between immigrants and natives was decomposed into four components, labeled as A, B, C, and D, which were calculated using the years of education (quantity) and the returns to education (price effect) of different workers by nativity and skill levels at various time points. Rows I and II show the results first without and then with controlling the industry and occupation variables.
Table 5

Decomposition of changes in relative mean earnings, with and without controls for workers’ industry and occupation

  

Relative price effect (A)

General price effect (B)

Relative quantity effect (C)

General quantity effect (D)

Total education effect

Other terms

Change of relative mean earnings

Panel A: all workers

Time 2–time 1

I

−0.2347

−0.0074

−0.1081

0.0146

−0.3355

0.2633

−0.0723

 

II

−0.0640

−0.0053

−0.0649

0.0220

−0.1122

0.0400

Time 3–time 2

I

−0.1949

0.0039

−0.0313

−0.0074

−0.2297

0.1721

−0.0576

 

II

0.0433

0.0052

−0.0193

−0.0067

0.0225

−0.0801

Panel B: high-skilled workers

Time 2–time 1

I

0.4455

−0.0102

−0.1079

0.0043

0.3311

−0.4616

−0.1305

 

II

0.9196

−0.0089

0.0786

0.0377

0.9698

−1.0003

Time 3–time 2

I

−0.1790

0.0133

−0.0297

0.0021

−0.1933

0.0394

−0.1538

 

II

−0.1940

0.0142

−0.0224

0.0008

−0.2014

0.0476

Panel C: low-skilled workers

Time 2–time 1

I

0.1083

0.0041

0.0071

−0.0028

0.1167

−0.0190

0.0977

II

0.0851

0.0020

0.0059

−0.0021

0.0910

0.0068

Time 3–time 2

I

−0.0456

0.0003

−0.0083

0.0003

−0.0533

0.1282

0.0749

II

−0.0228

0.0003

−0.0063

0.0001

−0.0287

0.1036

Relative mean earnings were calculated by subtracting the earnings of natives from the earnings of immigrants. The sample includes male employees age 15 years and above with positive wages only. All regressions include years of education, experience, and experience-squared. Estimates in all rows labeled I do not control for industry and occupation; estimates in all rows labeled II do control for industry and occupation. Source: Self-help Tabulation Service data sets of the Hong Kong Population Censuses/By-censuses

Among all four components, the relative price effect was the most useful in explaining the change in the earnings gaps by nativity for all workers (panel A), when industries and occupations were not controlled (row I). The relative price effect explains the widening earnings gap to the advantage of natives between Time 1 (1996) and Time 2 (2001) and between Time 2 and Time 3 (2006). The relative quantity effect also explains the earnings divergence, albeit to a much lesser extent. However, the general quantity effect from Time 1 to Time 3 was positive and cannot explain the diverging trend in relative mean earnings for immigrants and natives. When industries and occupations are taken into account (Row II), the explanatory power of the relative price effect was weak between Time 1 and Time 2. Also, between Time 2 and Time 3, both the relative price effect and the total education effect turned positive, and their magnitudes became smaller after controlling for industry and occupation. This suggests a small narrowing of the earnings gap between immigrants and natives. So, the diverging earnings gap between Time 2 and Time 3 was due to factors not related to educational attainment.

Next, we examined the decomposition by skill level. We found that the total education effects and relative price effects did not fully predict the earnings divergence among high-skilled workers (panel B), neither did they predict convergence among low-skilled workers (panel C). Our decomposition results showed that a clear trend of earnings divergence continued despite the relatively higher skill prices for immigrants (Fig. 1). This was most pronounced for high-skilled immigrants in their first 10 years in Hong Kong. The relative price effect predicted a .4455 positive log-earnings gap for immigrants and natives between Time 1 and Time 2, which was in the opposite direction of the mean log-earnings difference (−.1305). The relative price effect was even larger (.9196) after controlling for industries and occupations, which was about seven times larger than the mean log earnings gap.

In fact, the earnings divergence observed among high-skilled workers from Time 1 to Time 2 was captured well by the effect of the “other terms.” This was true even after the industry and occupation variables were controlled in the regressions, and the magnitude of the other terms was larger than the relative quantity effect, which also explained the earnings divergence. These results suggest that unobserved skills or institutional factors in earnings determination may be important factors behind the earnings divergence among high-skilled workers.

As for low-skilled workers (panel C), earnings convergence was observed between Time 1 and Time 2. This convergence can be explained by the convergence of the effect of relative skill price for that period. The effect of “other terms” was also prominent for its sign and magnitude in explaining earnings convergence for low-skilled workers between Time 1 and Time 2 and between Time 2 and Time 3 (i.e., after they had been in Hong Kong for 5 to 15 years). It is likely that, with the increase of tourists from the mainland after Hong Kong’s handover, low-skilled immigrant workers who speak Mandarin may have an increasing advantage, particularly in the service sector.10

In sum, although the relative price effect had strong explanatory power for the earnings gap between immigrants and natives, the effect was different for high- and low-skilled workers in different periods. The “other terms” also played an important role in earnings inequality by nativity, suggesting that other unobserved institutional factors need to be considered in future research on the nativity earnings gap. In general, our decomposition results are different from those reported in Lam and Liu’s study (2002a). One possible explanation for the difference is that Lam and Liu’s (2002a) work captured a specific historical period when illegal immigrants were prevalent. Although immigrants who entered Hong Kong illegally in the 1980s were allowed to apply for permanent residency under the reach-base policy, discrimination against them likely happened when they first entered the Hong Kong labor market while having their applications for legal status processed. The economic assimilation of these older cohorts may not represent the economic assimilation of legal immigrants, who are more common in Hong Kong today.

Returns to Education and Experiences Obtained Before and After Immigration

Our main goals in this paper were to decompose the earnings gap among immigrants and natives in different time periods and to compare the importance of the relative skill price effect and total education effect on the earnings divergence or convergence we observed. However, in this section, we further examine the skill prices of immigrants by the place of education they received because the imperfect transferability of education in one’s home country can possibly explain the wage differentials between natives and immigrants (Basilio and Bauer 2010; Friedberg 2000). Although most Chinese immigrants in this study, especially low-skilled workers, had completed their education before they moved to Hong Kong, some continued to invest education in Hong Kong. In this the previous section, we investigated the returns to two types of education for immigrants: education received in mainland China and education received in Hong Kong. We followed Friedberg (2000) to estimate the following equation:
$$ \begin{array}{*{20}c} \hfill \\ \ln\;{w_{it }}=\alpha +{\beta_1}Edu_{it}^h+{\beta_2}Edu_{it}^c+\gamma\;\mathrm{immigran}{{\mathrm{t}}_{it }}+{\varphi_1}Edu_{it}^h*\mathrm{immigran}{{\mathrm{t}}_{it }}+{\mu_1}Exp_{it}^h+{\mu_2}Exp_{it}^c+{\varphi_2}Exp_{it}^h*\mathrm{immigran}{{\mathrm{t}}_{it }}+\lambda {X_{it}}^{\prime }+{\varepsilon_{it }}, \hfill \\\end{array} $$
(4)
where X is a vector for a dummy variable indicating managerial and professional occupations, dummy variables for 13 industries, and a dummy variable named “immigrant” that takes the value of one if individual i is a Chinese immigrant from Mainland China. Exp is derived from age and years of education (Edu). α and ε are the constant and error term, respectively. Same as Eqs. 1, 2, and 3, the superscripts h and c refer to Hong Kong natives and Chinese immigrants, respectively, for individual i at time t. Because language proficiency might affect the productivity of immigrants as service industries in Hong Kong become more developed, we also included in X a dummy variable for spoken Cantonese proficiency and a dummy variable for spoken English proficiency. This regression model not only allowed the returns to education received in China among Chinese immigrants to be different from their education received in Hong Kong; it also allowed the return to immigrants’ labor market experience (Exp) in China to be different from that in Hong Kong. The portability of education and work experience obtained in China to Hong Kong’s labor market was measured by β2 and μ2. To the extent that immigrants gradually sort themselves into occupations that reward their home-country education and work experience, this model would show an increase in both coefficients.
Results in Table 6 are consistent with previous literature that compared immigrants and natives in other countries (Basilio and Bauer 2010; Friedberg 2000). Returns to education and experiences obtained in the country of origin (in this study, China) were lower than those obtained in the host country (in this study, Hong Kong). For instance, in panel A in Table 6, controlling for industry and occupation differences, returns to education obtained in China were about 5 %, whereas the returns to education obtained in Hong Kong were 7 % to 8 % for all three time points. However, returns to local education in Hong Kong are quite low for Chinese immigrants. The interaction term of education and immigrant status implies that the gap in returns to education between Hong Kong natives and Chinese immigrants became even larger in the later periods.11 In general, the variation of skill prices for high- and low-skilled immigrants in our results could imply that the quality of education received in home country are different for low-skilled and high-skill workers. As we can see, returns to English proficiency is much higher than the returns to Cantonese proficiency, although both coefficients are positive and statistically significant. For high-skilled workers, interestingly, the returns to English proficiency decreases over time.
Table 6

Regression of log earnings on immigrants’ pre- and postmigration education and experience, by skill level

 

Time 1

Time 2

Time 3

 

I

II

I

II

I

II

Panel A: all workers

Immigrant

0.282* (0.030)

0.195* (0.028)

0.592* (0.053)

0.462* (0.049)

1.060* (0.078)

0.853* (0.070)

Education in HK

0.117* (0.000)

0.069* (0.000)

0.135* (0.000)

0.083* (0.000)

0.127* (0.000)

0.075* (0.000)

Education in China

0.091* (0.002)

0.053* (0.002)

0.088* (0.002)

0.051* (0.002)

0.063* (0.002)

0.036* (0.002)

Education_HK*immigrant

−0.050* (0.009)

−0.032* (0.008)

−0.073* (0.007)

−0.058* (0.006)

−0.102* (0.007)

−0.077* (0.006)

Experience in HK

0.030* (0.000)

0.025* (0.000)

0.029* (0.000)

0.024* (0.000)

0.021* (0.000)

0.018* (0.000)

Experience in China

0.014* (0.001)

0.010* (0.001)

−0.001 (0.001)

−0.002 (0.001)

0.002 (0.001)

0.003 (0.001)

Experience_HK*immigrant

0.004 (0.006)

0.007 (0.005)

0.009 (0.006)

0.012 (0.006)

−0.037* (0.006)

−0.034* (0.005)

English proficiency

0.363* (0.022)

0.293* (0.020)

0.403* (0.022)

0.394* (0.020)

0.267* (0.023)

0.273* (0.020)

Cantonese proficiency

0.051* (0.008)

0.037* (0.008)

0.078* (0.010)

0.058* (0.009)

0.061* (0.010)

0.047* (0.009)

Adj. R2

0.356

0.452

0.392

0.493

0.322

0.458

Panel B: high-skilled workers

Immigrant

0.206 (0.086)

0.584* (0.078)

−0.200 (0.131)

0.378** (0.121)

0.368 (0.157)

1.118* (0.138)

Education in HK

0.151* (0.001)

0.107* (0.001)

0.192* (0.001)

0.143* (0.001)

0.157* (0.001)

0.110* (0.001)

Education in China

0.142* (0.005)

0.078* (0.005)

0.211* (0.007)

0.128* (0.006)

0.166* (0.007)

0.097* (0.006)

Education_HK*immigrant

−0.101* (0.015)

−0.076* (0.014)

−0.011 (0.013)

−0.045* (0.012)

−0.042* (0.012)

−0.082* (0.011)

Experience in HK

0.050* (0.000)

0.042* (0.000)

0.045* (0.000)

0.039* (0.000)

0.033* (0.000)

0.029* (0.000)

Experience in China

0.032* (0.002)

0.022* (0.002)

0.014* (0.003)

0.009* (0.002)

0.004 (0.003)

0.007 (0.003)

Experience_HK*immigrant

−0.069* (0.010)

−0.054* (0.009)

−0.022 (0.011)

−0.016 (0.011)

−0.073* (0.010)

−0.087* (0.009)

English proficiency

0.467* (0.032)

0.373* (0.029)

0.423* (0.031)

0.359* (0.028)

0.298* (0.031)

0.289* (0.027)

Cantonese proficiency

0.123* (0.018)

0.076* (0.017)

0.101* (0.020)

0.014 (0.018)

0.139* (0.019)

0.084* (0.017)

Adj. R2

0.370

0.486

0.410

0.498

0.260

0.429

Panel C: low-skilled workers

Immigrant

0.106 (0.052)

0.082 (0.051)

0.073 (0.077)

−0.007 (0.073)

0.388** (0.118)

0.145 (0.110)

Education in HK

0.030* (0.001)

0.026* (0.001)

0.037* (0.001)

0.029* (0.001)

0.041* (0.001)

0.031* (0.001)

Education in China

0.013 (0.005)

0.011 (0.005)

0.035* (0.006)

0.032* (0.005)

0.031* (0.006)

0.030* (0.005)

Education_HK*immigrant

−0.021 (0.013)

−0.020 (0.013)

−0.007 (0.013)

−0.003 (0.012)

−0.074* (0.016)

−0.035 (0.015)

Experience in HK

0.017* (0.000)

0.015* (0.000)

0.014* (0.000)

0.012* (0.000)

0.010* (0.000)

0.009* (0.000)

Experience in China

0.003 (0.001)

0.002 (0.001)

−0.004** (0.001)

−0.003 (0.001)

−0.003 (0.002)

−0.002 (0.002)

Experience_HK*immigrant

0.030* (0.008)

0.032* (0.008)

0.014 (0.008)

0.015 (0.008)

−0.018 (0.009)

−0.003 (0.008)

English proficiency

0.134 (0.052)

0.078 (0.051)

0.385* (0.064)

0.298* (0.061)

0.077 (0.081)

0.116 (0.076)

Cantonese proficiency

0.064* (0.011)

0.062* (0.011)

0.124* (0.013)

0.119* (0.012)

0.037 (0.014)

0.047* (0.013)

Adj. R2

0.089

0.135

0.064

0.147

0.031

0.160

The sample includes male employees age 15 years and above with positive wages only. In addition to the variables shown in the table, both Model 1 and Model 2 include dummies for spoken Cantonese and English proficiency. Additionally, Model 2 includes 13 industry dummies and a dummy for white-collar jobs. Standard errors are in parentheses. Source: Self-help Tabulation Service data sets of the Hong Kong Population Censuses/By-censuses

*p < .001, **p < .01

As shown in Table 6, whether controlling for industry and occupation, high-skilled immigrants had positive returns to education received in China. Taking into account the interaction term between education in Hong Kong and immigration status, the returns to education received in Hong Kong were actually lower than the returns to education received in China. For example, at Time 1, without considering occupation and industry differences (column I), the return to education in Hong Kong was .05 (i.e., subtracting .101 from .151), about .092 lower than their return to education received in China (.142).

It is worth noting that the difference of returns to education in Hong Kong for high-skilled and low-skilled immigrants compared with the natives (i.e., ϕ1 in Eq. 4) revealed the same pattern of the relative price effect that was shown in the decomposition of changes in relative mean earnings (Table 5). Here are some examples after controlling for industry and occupation. In panel B in Table 6, for high-skilled workers, the negative coefficient φ means relatively lower returns to local education for the Chinese immigrants compared with Hong Kong natives. The difference in the returns to local education became significantly smaller from Time 1 (−.076, with a standard error of .014) to Time 2 (−.045 with a standard error of .012) but enlarged again from Time 2 to Time 3 (−.082, with a standard error of .011). Recall that in our decomposition analysis for high-skilled workers (panel B of Table 5), the relative price effect was positive on the change of relative mean earnings between Time 1 and Time 2 (.9196), but the relative price effect was negative between Time 2 and Time 3 (−.1940). This means that the skill price of immigrants has increased more than that of natives from Time 1 to Time 2 (i.e., a narrowing nativity gap in the returns to education between Time 1 and Time 2), but the skill price of immigrants has increased less than that of natives from Time 2 to Time 3 (i.e., an enlarging nativity gap in the returns to education between Time 2 and Time 3). Therefore, whichever method we used, whether estimating the returns to education received in Hong Kong in the pooled sample for both natives and immigrants (Eq. 4), or examining the returns to education, local and nonlocal alike, in the two-equation model (Eqs. 1 and 2), the changes in the relative skill prices could not fully explain the earnings assimilation patterns we observed. In other words, the place of education did not help to explain the earnings divergence for high-skilled male employees.

Panel C (Table 6) illustrates a similar story for low-skilled workers. The negative difference in the returns to education also shrank from Time 1 to Time 2 (although both coefficients of −.02 and −.003 are not statistically significant) and became larger at Time 3 (−.035 with a standard error of .015). This pattern was the same in models without the controls for industry and occupation as well. The changes of returns to local education between two time points is similar to the changes of returns to education in the decomposition analysis (panel C in Table 5) that did not separate the education received in Hong Kong from the education received in China. In short, our estimates from Eq. 4 yield similar results as our estimates from the two-equation model. The place where immigrants received their education did not appear to matter for the overtime change in the nativity earnings gaps. Thus, our results did not support the hypothesis that human capital obtained locally is more adaptive to technical change in the host country than human capital acquired in the immigrants’ home countries.

Conclusion

Using the three most recent, publicly available years of the Hong Kong Census (1996, 2001, and 2006), this study examined the economic assimilation of male Hong Kong immigrants with different skill levels as defined by their human capital. We decomposed earnings differentials to reveal whether the returns to immigrant men’s human capital contributed to an earnings divergence or convergence with native male workers.

The results showed an earnings divergence between immigrant and native workers, indicating an increase in economic disadvantage of recent Chinese immigrant workers in Hong Kong. However, this general trend masked differences by skill level. Separate analyses of high- and low-skilled workers led to different assimilation patterns. The nativity gap enlarged over time for high-skilled immigrants even though their skill prices, or returns to education, were higher than the skill prices of natives. By contrast, low-skilled immigrant workers achieved earnings parity with low-skilled native workers. This finding is consistent with past literature on economic assimilation of low-skilled immigrant workers (Hall and Farkas 2008).

An unusual feature of Hong Kong is that Chinese immigrant male workers who entered the country in the past two decades were on average more educated than their native counterparts. In this respect, recent Chinese immigrants are similar to Asian Indian and African immigrants in the United States, who generally have higher educational attainment than the average American (Portes and Rumbaut 2006; Massey et al. 2007). Over the 15-year period in the present study, the immigrant–native gap in years of education narrowed. However, the immigrant–native difference in the returns to education diverged. Separate analyses of workers with different skill levels were revealing. Among male workers who had a lower secondary education or less, the returns to education converged between immigrants and natives. Among male workers who had a post-secondary education, high-skilled Chinese immigrants had lower returns in 1996 but enjoyed higher returns to their education than did high-skilled native workers in subsequent years. Only among male workers who had an upper secondary education did we observe a divergence in the returns to education between immigrant and native workers.

Did the change in the returns to education over time explain the change in the nativity gap in earnings? Although Lam and Liu (2002a, 2002b) found that the divergence in the returns to education was the major cause of the earnings divergence for immigrants and natives in the 1980s, our analyses do not produce the same results. Specifically, converging returns to education explained converging earnings among low-skilled workers from 2001 to 2006, but converging returns to education were actually related to an earnings divergence among high-skilled workers. We find that the recent Chinese immigrant cohort who possess more education have gained relative earnings benefits due to rising skill prices. This pattern was especially clear after the immigrants stayed in Hong Kong for 5 to 15 years and for low-skilled immigrants. However, there seemed to be other unobserved factors affecting the high-skilled Chinese immigrants’ earnings that enlarged the earnings gap. Other explanations, such as occupational segregation or the glass ceiling, may offer more insight into the earnings-assimilation patterns observed in this study.

This study highlights the usefulness of separating skill level in the study of economic assimilation. According to Lam and Liu (2002b), Hong Kong’s earnings divergence by nativity was driven by technical change that raised the returns to local education relative to the returns to mainland education, with the assumption that nonlocal education is less adaptable to technical change. Here we found that, among mainland immigrants, the skill price of non-local education (received in China) is higher than the skill price of local education (received in Hong Kong). This finding challenges the assumption that nonlocal education is less adaptable to technical change. In addition, if technical change was the reason for divergence in earnings and skill price, the effect on technical change should be minimal for low-skilled workers. We found both converging returns to education and converging earnings for low-skilled immigrants and natives in this paper.

We further estimated the returns to local (Hong Kong) and nonlocal (mainland China) education for immigrants. The results revealed that the returns to nonlocal education were not low, suggesting that the imperfect transferability of home country’s education found in previous literature might not apply to Hong Kong’s recent Chinese immigrants.12 We conclude that, although the earnings divergence in Hong Kong continues to exist for all workers, the driving forces of the divergence may be diverse and warrant further investigation. One area for future research is to examine the factors that contribute to the skill-price differences for immigrants and natives. Another area is to find out whether the influx of immigrants affects the skill price of natives and the extent to which immigrant workers’ skill distribution matters (Jaeger 2007).

Our results on low-skilled workers are consistent with findings in the U.S. on the economic assimilation of low-skilled immigrant workers (Hall and Farkas 2008). These results support in part the view that immigrant groups who share the same ethnicity with the majority group in the host country, and who receive government support are likely to achieve upward socioeconomic mobility over time (Portes and Rumbaut 2006). However, this same view cannot explain our results on high-skilled workers. Zeng and Xie (2004) suggested that high-skilled Asian immigrant workers faced a glass ceiling that hindered their economic assimilation. Such glass ceiling for Asian immigrants in the United States likely involves racial/ethnic bias, but this would not apply to high-skilled mainland Chinese workers in Hong Kong who come primarily from the same Han ethnicity as do Hong Kong natives. Despite sharing the same ethnicity, exclusion and prejudice toward Mainlanders immigrants has long existed, albeit covertly, which can be traced back to the Colonial period. Stephen Chiu and his colleagues concluded a study of the barriers facing immigrants in the Colonial period that, “….even when educational attainment was taken into account, Chinese immigrants were consistently penalized in their initial class placement, subsequent mobility, and current income attainment. They were unable to close their initial deficit in labor market attainment over time.” (2005, p. 222). Our study suggests that the immigrant penalty continued to exist after the Handover. In Hong Kong, nativity, not ethnicity, is an important source of economic inequality.

Footnotes
1

The 20 % random sample is drawn by the first author from the full datasets of the Population Censuses/by-censuses. The design of the 1991 and 2001 Population Census comprised a complete but simple enumeration of all persons on their basic characteristics such as age and sex, and a detailed 1/7 sample enquiry on a broad range of socio-economic characteristics of the population. The design of the 1996 Population by-census and the 2006 Population by-census were a 1/7 and 1/10 sample enquiry on a broad range of socio-economic characteristics of the population, respectively.

 
2

The Chinese ethnicity made up 95 % of its population in 2006. Other ethnic groups comprised 1.6 % Filipinos, 1.3 % Indonesians, and 2 % of a variety of ethnicities, including Caucasian, Indian, Nepalese, Japanese, Thai, Pakistani, and other Asian (Hong Kong Census and Statistics Department 2006).

 
3

The minimum work age was changed to 18 years in 2008.

 
4

We also conducted the analysis using the age range of 25 to 50 to restrict the sample to individuals who had completed their education and were economically active at the time they were first observed. The results were virtually identical to those reported here.

 
5

Hong Kong’s education system contains 6 years of primary school, 3 years of lower secondary school, and 2 years of upper secondary school. After secondary school, students can either study craft courses in technical institutions, or attend two years of matriculation courses, which prepare them for college (i.e. first degree programs). College education spanned 3 years during the period we studied.

 
6

Results are available from the authors upon request.

 
7

We defined managers and professionals’ occupations as managers and administrators; professionals and associate professionals. Other occupations include clerks, service workers and shop sales workers, skilled agricultural and fishery workers, craft and related workers, plant and machine operators and assemblers, and elementary occupations.

 
8

Average earnings for each industry in 1996, 2001, and 2006 are available upon request.

 
9

As discussed earlier and shown in Table 1, there were some changes in the average years of education over time for both immigrants and natives. Even if we restrict our sample to the cohort of workers ages 25 years and above (results not shown), who were likely to have completed their education, there are still some, albeit small differences between t and t + 1. It is reasonable that the average levels would increase over time because individuals age 15 years and above will continue education. However, sample selection could also change the mean levels of education of the artificial cohorts we track. For example, if immigrants with more education left Hong Kong after a short period of stay and only those with relatively less education remained, the average education level for immigrants should become lower over time, which was the case in our data. It is also plausible that the pursuit of further education (possibly due to occupational needs) for immigrants and natives are different. Nonetheless, the changes were less than a year in our study.

 
10

Restaurant business hired the most immigrant male employees in all Census years we studied: about 11.8 % in 1996, 14.4 % in 2001, and 19.0 % in 2006.

 
11

Returns to experience in China were less than half of the returns to experience in Hong Kong in the first 5 years after arrival. Furthermore, the returns to experience in China became statistically insignificant at the later time points. It is interesting that the returns to experience in Hong Kong actually became negative for Chinese immigrants. This seems to indicate a devaluation of experience by their employers, an improved quality of education in the host society, a lack of internal labor market opportunities, an obsolescence of accumulated human capital and effects of aging (Casanova 2010), or some combination of these factors. Just as we observed for all immigrants, the returns to foreign work experience for high-skilled immigrants became very small and even insignificant at Time 3. There was also a negative return to local work experience for high-skilled Chinese immigrants. It is even more interesting that the return to local work experience was positive for low-skilled Chinese immigrants in their first 5 years in Hong Kong. Even at Time 2 and Time 3, the estimated returns to work experience, regardless of whether controls were included for industrial and occupational differences, were very similar and positive for both Chinese-immigrant and native low-skilled workers. This might explain the earnings convergence that we observed for this group.

 
12

This might due to the high-skilled immigration policy, which ensured that only those who possessed skills that would contribute substantially to Hong Kong’s labor force were recruited and permitted to enter Hong Kong’s labor market.

 

Copyright information

© Springer Science+Business Media Dordrecht 2012