Temporary Foreign Workers and Former International Students as a ...

Introduction Results Conclusion

Temporary Foreign Workers and Former International Students as a Source of Permanent Immigration Arthur Sweetman and Casey Warman School of Policy Studies and Department of Economics Queen’s University Please do not quote without permission Email: [email protected]

March 12, 2008 Arthur Sweetman and Casey Warman, Queen’s University

Temporary Foreign Workers in Canada

1/15


Introduction Data

Introduction As is well known, recent immigrants have done very poorly in the Canadian labour market.

Arthur Sweetman and Casey Warman, Queen’s University


2/15


Introduction Data

Introduction As is well known, recent immigrants have done very poorly in the Canadian labour market. Much of the decline is commonly attributed to: I

Lack of recognition of pre-Canadian human capital

I

Lack of relevance of pre-Canadian human capital

I

Canadian language/literacy gaps that reduce the productivity (transferability) of pre-Canadian human capital

I

Need for complementary Canadian knowledge/networks to make immigrant human capital more productive



2/15


Introduction Data

Two approaches



3/15


Introduction Data

Two approaches 1. What we see as more policy relevant



3/15


Introduction Data

Two approaches 1. What we see as more policy relevant I Examine what TFW or Student status add in terms of additional information to the point system I Only include variables that are controlled for under the points or admission system I Example: no differentiation or controls for gender or region of origin I We attempt to predict the number of points that a worker would have obtained under the point system → could not perfectly predict all of the points



3/15


Introduction Data




3/15


Introduction Data




3/15


Introduction Data




3/15


Introduction Data

Two approaches 1. What we see as more policy relevant I Examine what TFW or Student status add in terms of additional information to the point system I Only include variables that are controlled for under the points or admission system I Example: no differentiation or controls for gender or region of origin I We attempt to predict the number of points that a worker would have obtained under the point system → could not perfectly predict all of the points 2. More traditional economic specifications to compare to the literature and look at different questions Example: separate by gender, controls for origin region etc. Arthur Sweetman and Casey Warman, Queen’s University


3/15


Introduction Data

Data: Longitudinal Survey of Immigrants to Canada I

Sample of those who landed from outside Canada between Oct 2000 and Sept 2001

I

Three interviews: 6 months, 2 yrs, 4 yrs For those who lived in Canada prior to immigration

I

I I I

Asks if had work visa Asks if foreign student Do not know when were TFW or student



4/15


Introduction Data

Some differences or possible differences between our sample and those entering under the Canadian Experience Class (CEC)



5/15


Introduction Data

Some differences or possible differences between our sample and those entering under the Canadian Experience Class (CEC) I

Our sample applied abroad/CEC apply within Canada



5/15


Introduction Data



I

Students in our sample do not have work experience, whereas they require one year under the CEC



5/15


Introduction Data



I


I

Different type of immigrant entering under this program?



5/15


Introduction Data



I


I


I

Scale differences? → only small proportion of TFWs and students entered under the old policy



5/15


Introduction Data



I


I


I

Scale differences? → only small proportion of TFWs and students entered under the old policy

That being said, should still be informative of outcomes of immigrants entering under the Canadian Experience Class



5/15


Descriptive Statistics Policy Relevant Results Skilled Worker Principal Applicants

Table 1: Within-Class TFW and Student Distributions Student(%) TFW (%) Other (%) Total(%) Males Other Classes Skilled Workers (PA) Total(%)

1.5 5.3 3.8

1.5 5.5 3.9

97.0 89.3 92.4

100.0 100.0 100.0

Females Other Classes Skilled Workers (PA) Skilled Workers (SD) Total(%)

1.2 10.4 1.6 3.0

0.7 5.5 2.5 2.2

98.1 84.2 96.0 94.8

100.0 100.0 100.0 100.0

Notes: Sample age 19 to 62 at the time of the first cycle. Other immigrants are all immigrants who have no pre-Canadian human capital at the time of immigration. Arthur Sweetman and Casey Warman, Queen’s University


6/15



Table 2: Across-Class TFW and Student Distributions Student(%) TFW (%) Other (%) Males Other Classes Skilled Workers (PA) Total(%)

16.2 83.8 100.0

15.5 84.5 100.0

42.2 57.8 100.0

Females Other Classes Skilled Workers (PA) Skilled Workers (DS) Total(%)

17.5 61.7 20.8 100.0

12.8 43.6 43.6 100.0

44.5 15.9 39.7 100.0

Notes: Sample age 19 to 62 at the time of the first cycle. Other immigrants are all immigrants who have no pre-Canadian human capital at the time of immigration. Arthur Sweetman and Casey Warman, Queen’s University


7/15



Table 3: OLS (ln) Weekly Earnings Regressions Immigration Category, Canadian Exposure and Points, Cycle 3 (Sample includes zero $ earners, with zero set to $ 1 prior to taking ln) Family Spouses/Fiancs (PA) Parents/Grandparents (PA and S) Other Economic Provincial Nominee Business Refugees Government Private Other

(1)

(2)

(3)

(4)

(5)

(6)

−0.618∗∗ [0.110] −1.929∗∗ [0.158] 0.477∗∗ [0.147]

−0.580∗∗ [0.110] −1.878∗∗ [0.159] 0.512∗∗ [0.147]

−0.227∗ [0.123] −0.473∗∗ [0.235] 0.770∗∗ [0.190]

−0.222∗ [0.123] −0.475∗∗ [0.236] 0.771∗∗ [0.190]

−0.287∗∗ [0.115] −1.139∗∗ [0.181] 1.153∗∗ [0.165]

−0.268∗∗ [0.115] −1.122∗∗ [0.181] 1.158∗∗ [0.165]

0.072 [0.483] −2.704∗∗ [0.158]

0.048 [0.489] −2.691∗∗ [0.157]

−0.196 [0.430] −2.297∗∗ [0.167]

−0.179 [0.426] −2.303∗∗ [0.167]

0.126 [0.456] −2.300∗∗ [0.165]

0.121 [0.460] −2.303∗∗ [0.165]

−1.204∗∗ [0.116] −0.619∗∗ [0.276] −1.333∗∗ [0.364]

−0.702∗∗ [0.137] −0.040 [0.268] −0.683∗ [0.370]

Yes

−0.691∗∗ [0.137] −0.036 [0.268] −0.677∗ [0.370] −0.191 [0.230] 0.292 [0.209] Yes

−0.569∗∗ [0.131] −0.074 [0.273] −0.594 [0.370]

No

−1.151∗∗ [0.116] −0.564∗∗ [0.276] −1.280∗∗ [0.365] 0.562∗∗ [0.217] 0.677∗∗ [0.197] No

0.07

0.08

0.13

0.13

−0.547∗∗ [0.131] −0.048 [0.273] −0.575 [0.370] 0.314 [0.213] 0.598∗∗ [0.194] 0.021∗∗ [0.002] 0.09

tfw student 46 Point Dummies or Point Estimates R2


0.022∗∗ [0.002] 0.09


8/15



Table 4: Marginal Effects from Probit Estimates on Employment, Cycle 3 (1) Family Spouses/Fiancs (PA) Parents/Grandparents (PA and S) Other Economic Provincial Nominee Business Refugees Government Private Other

(3)

−0.109 [0.019] −0.354∗∗ [0.026] 0.05 [0.033]

−0.047 [0.021] −0.077∗∗ [0.039] 0.133∗∗ [0.027]

0.047 [0.081] −0.150∗∗ [0.027]

0.032 [0.086] −0.146∗∗ [0.027]

−0.228∗∗ [0.020] −0.099∗ [0.052] −0.313∗∗ [0.063]

−0.221∗∗ [0.020] −0.092∗ [0.051] −0.306∗∗ [0.063] 0.136∗∗ [0.029] 0.015 [0.033] No

student No

∗∗

(4)

−0.115 [0.019] −0.361∗∗ [0.026] 0.045 [0.033]

tfw

46 Point Dummies or Point Estimates

(2) ∗∗


∗∗

(5)

(6)

−0.048 [0.021] −0.078∗∗ [0.039] 0.132∗∗ [0.027]

∗∗

−0.040 [0.019] −0.175∗∗ [0.031] 0.155∗∗ [0.024]

−0.038∗∗ [0.019] −0.174∗∗ [0.031] 0.156∗∗ [0.024]

−0.004 [0.092] −0.048∗ [0.027]

−0.006 [0.092] −0.047∗ [0.027]

0.063 [0.077] −0.054∗∗ [0.026]

0.052 [0.081] −0.053∗∗ [0.026]

−0.080∗∗ [0.024] 0.013 [0.047] −0.156∗∗ [0.068]

−0.083∗∗ [0.024] 0.012 [0.047] −0.158∗∗ [0.068] 0.005 [0.047] −0.049 [0.039] Yes

−0.077∗∗ [0.022] 0.022 [0.045] −0.135∗∗ [0.065]

−0.076∗∗ [0.022] 0.024 [0.045] −0.134∗∗ [0.065] 0.101∗∗ [0.032] −0.001 [0.034] 0.005∗∗ [0.000]

Yes

∗∗

0.005∗∗ [0.000]


9/15



3

log weekly earnings 4 5

6

7

Figure 1: Partial Linear Model Estimates of relationship between imputed Points and Log Earnings

0

20

40

60

67

80

100

Points

Notes: The equation: ln(Earningsi ) = Xi β + f (pointsi ) + i was estimated by the double-residual method proposed by Robinson (1988). The parametric controls include immigrant class and months since migration. Arthur Sweetman and Casey Warman, Queen’s University


10/15



0

.2

employment .4 .6

.8

1

Figure 2: Partial Linear Model Estimates of relationship between imputed Points and probability of being employed

0

20

40

60

67

80

100

Points

Notes: The equation: Employedi = Xi β + f (pointsi ) + i was estimated by the double-residual method proposed by Robinson (1988). The parametric controls include immigrant class and months since migration. Arthur Sweetman and Casey Warman, Queen’s University


11/15



Table 5: Log weekly earnings regressions for Skilled Worker Principal Applicants

Males tfw student R2 Females tfw student Full set of controls R2

(1) Cycle1

(2) Cycle2

(3) Cycle3

(4) Cycle1

(5) Cycle2

(6) Cycle3

0.908∗∗ [0.076] 0.111 [0.083] 0.11

0.745∗∗ [0.067] 0.088 [0.084] 0.09

0.629∗∗ [0.066] 0.028 [0.074] 0.08

0.723∗∗ [0.078] 0.111 [0.086] 0.24

0.595∗∗ [0.069] 0.078 [0.081] 0.24

0.476∗∗ [0.072] −0.004 [0.075] 0.19

0.443∗∗ [0.160] 0.092 [0.136] No 0.03

0.369 [0.242] 0.186∗∗ [0.093] No 0.03

0.129 [0.194] 0.206 [0.136] No 0.03

0.358∗∗ [0.148] −0.063 [0.165] Yes 0.21

0.273 [0.210] 0.084 [0.116] Yes 0.19

0.155 [0.158] 0.103 [0.176] Yes 0.18

Notes: Sample age 19 to 62 at the time of the first cycle. All regressions control for months since migration. Full set of controls include: age, highest degree prior to landing, region of origin, region of residence, English and French language ability, marital status and number of kids. Robust standard errors in brackets. * significant at 10%; ** significant at 5% Arthur Sweetman and Casey Warman, Queen’s University


12/15



Table 6: Marginal Effects from Probit Estimates on Employment for Skilled Worker Principal Applicants

Males tfw student

Females tfw student Full set of controls

(1) Cycle1

(2) Cycle2

(3) Cycle3

(4) Cycle1

(5) Cycle2

(6) Cycle3

0.242∗∗ [0.037] 0.03 [0.053]

0.183∗∗ [0.028] 0.043 [0.044]

0.099∗∗ [0.027] −0.027 [0.042]

0.160∗∗ [0.054] 0.016 [0.062]

0.171∗∗ [0.029] 0.083∗∗ [0.039]

0.085∗∗ [0.029] 0.009 [0.038]

0.144∗ [0.086] 0.116∗ [0.070] No

0.121 [0.079] 0.095 [0.059] No

0.129∗∗ [0.054] 0.007 [0.057] No

0.117 [0.108] 0.07 [0.091] Yes

0.113 [0.082] 0.046 [0.071] Yes

0.081 [0.067] −0.018 [0.072] Yes

Notes: Sample age 19 to 62 at the time of the first cycle. All regressions control for months since migration. Full set of controls include: age, highest degree prior to landing, region of origin, region of residence, English and French language ability, marital status and number of kids. Robust standard errors in brackets. * significant at 10%; ** significant at 5%



13/15



Table 7: Log earnings and Marginal Effects from Probit Estimates on Employment for Skilled Worker Principal Applicants, males from non-Western Countries

Earnings tfw student R2 Employment tfw student Full set of controls

(1) Cycle1

(2) Cycle2

(3) Cycle3

(4) Cycle1

(5) Cycle2

(6) Cycle3

0.853∗∗ [0.090] 0.073 [0.090] 0.09

0.739∗∗ [0.073] 0.07 [0.091] 0.07

0.585∗∗ [0.078] 0.019 [0.079] 0.06

0.834∗∗ [0.094] 0.100 [0.098] 0.19

0.688∗∗ [0.077] 0.080 [0.088] 0.21

0.519∗∗ [0.088] −0.005 [0.081] 0.16

0.225∗∗ [0.046] 0.012 [0.057] No

0.188∗∗ [0.034] 0.038 [0.048] No

0.077∗∗ [0.038] −0.034 [0.046] No

0.154∗∗ [0.067] −0.001 [0.068] Yes

0.187∗∗ [0.032] 0.082∗ [0.044] Yes

0.073∗∗ [0.036] 0.008 [0.042] Yes

Notes: Sample age 19 to 62 at the time of the first cycle. All regressions control for months since migration. Full set of controls include: age, highest degree prior to landing, region of origin, region of residence, English and French language ability, marital status and number of kids. Robust standard errors in brackets. * significant at 10%; ** significant at 5%



14/15


Conclusion

Conclusion

B TFWs have superior outcomes in all specifications → Both employment and earnings B Foreign students have more mixed results B Both entry characteristics appear to have predictive power beyond that in the points system B Overall people in these groups appear to succeed in the Canadian labour market more smoothly/quickly



15/15