Evaluating the Source of Low Returns to Immigrants Foreign - - PowerPoint PPT Presentation

Evaluating the Source of Low Returns to Immigrants’ Foreign Experience and Credentials in Canada: Problems with Recognition or Reflection of Productivity? Natalya N. Dygalo

University of Saskatchewan

November 19, 2008

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

The Question: Low returns to immigrants’ foreign experience and credentials: evaluating 2 explanations

◮ 1. Inefficient foreign credentials recognition

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

The Question: Low returns to immigrants’ foreign experience and credentials: evaluating 2 explanations

◮ 1. Inefficient foreign credentials recognition ◮ 2. Reflect labour markets’ fair assessment of the

relative productivity between workers with Canadian and foreign credentials

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

The Issue: Low returns to immigrants’ foreign experience and credentials:

◮ Ferrer and Riddell (2008), Table 5C: total

returns to education of immigrants from Europe, Asia and South America relative to native born are 40-55% for high school, 60-66% for U. Bachelor’s

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

The Issue: Low returns to immigrants’ foreign experience and credentials:

◮ Ferrer and Riddell (2008), Table 5C: total

returns to education of immigrants from Europe, Asia and South America relative to native born are 40-55% for high school, 60-66% for U. Bachelor’s

◮ Green and Worswick (2004): returns to foreign

experience are close to zero in many specifications for entry cohorts starting in 1990-92

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

The Issue, continued:

◮ Pendakur and Woodcock (2008): wage diff

between recent immigrants (less than 10 years) and native born male workers: white -7%; visible minority -31%

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

The Issue, continued:

◮ Pendakur and Woodcock (2008): wage diff

between recent immigrants (less than 10 years) and native born male workers: white -7%; visible minority -31%

◮ Pendakur and Woodcock (2008): wage diff

between non-recent immigrants (more than 10 years) and native born male workers: white -7%; visible minority -19%

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Contribution to existing research:

◮ Existing research: relative wages between

workers with foreign and Canadian experience, education, credentials

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Contribution to existing research:

◮ Existing research: relative wages between

workers with foreign and Canadian experience, education, credentials

◮ No research on productivity comparisons

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Contribution to existing research:

◮ Existing research: relative wages between

workers with foreign and Canadian experience, education, credentials

◮ No research on productivity comparisons ◮ This paper: compares productivity

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Objectives of this project: More direct productivity comparisons between:

◮ Canadian- and foreign-born workers

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Objectives of this project: More direct productivity comparisons between:

◮ Canadian- and foreign-born workers ◮ Workers with Canadian and foreign experience

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Objectives of this project: More direct productivity comparisons between:

◮ Canadian- and foreign-born workers ◮ Workers with Canadian and foreign experience ◮ Workers with Canadian and foreign education,

credentials

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Objectives of this project: More direct productivity comparisons between:

◮ Canadian- and foreign-born workers ◮ Workers with Canadian and foreign experience ◮ Workers with Canadian and foreign education,

credentials

◮ Foreign-born workers just after entry and

workers with Canadian experience (productivity basis for the ”entry” effect)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Contribution to policy debates I (CLSRN objective)

◮ Honourable Diane Finley, (former) Minister of

Citizenship and immigration,”[t]oo many newcomers can’t get jobs they have been trained for. That’s a terrible waste, for them and for the country”

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Contribution to policy debates II (CLSRN objective)

◮ Foreign Credentials Referral Office (FCRO). One

• f the objectives of the approximately 320

FCRO outlets is ”increasing employer awareness

• f the process for, and benefits of, hiring

internationally trained and educated professionals”

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Contribution to policy debates III (CLSRN objective)

◮ Still unaddressed issue: to what extent can

Canadian employers treat foreign credentials as Canadian equivalents? Measuring relative productivity between Canadian and foreign experience may have important implications for policy

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Measuring relative productivity between Canadian and foreign-born workers: method and data

◮ Method: production functions with relative

productivities as estimated parameters

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Measuring relative productivity between Canadian and foreign-born workers: method and data

◮ Method: production functions with relative

productivities as estimated parameters

◮ Primary data source: Workplace and Employee

Survey (WES) 1999-2005 plus selected information matched from CANSIM II

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Objectives of this presentation:

◮ Placement of the project in academic literature

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Objectives of this presentation:

◮ Placement of the project in academic literature ◮ Presentation of the method (details)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Objectives of this presentation:

◮ Placement of the project in academic literature ◮ Presentation of the method (details) ◮ Description of data work

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Labour Economics Literature I:

◮ Ferrer and Riddell (2008) - wage returns to

Canadian/ foreign experience, education, credentials

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Labour Economics Literature I:

◮ Ferrer and Riddell (2008) - wage returns to

Canadian/ foreign experience, education, credentials

◮ Ferrer, Green, and Riddell (2006) - literacy

scores

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Labour Economics Literature II:

◮ Green and Worswick (2002) - differential returns

to Canadian and foreign experience

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Labour Economics Literature II:

◮ Green and Worswick (2002) - differential returns

to Canadian and foreign experience

◮ Sweetman (2003) - adjustment for education

quality in the country of origin (international test scores)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Back to this project: measuring relative productivity - challenges

◮ Relative productivity between Canadian and

foreign-born workers

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Back to this project: measuring relative productivity - challenges

◮ Relative productivity between Canadian and

foreign-born workers

◮ Individual worker productivity measures would

be preferred but are unattainable

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Back to this project: measuring relative productivity - challenges

◮ Relative productivity between Canadian and

foreign-born workers

◮ Individual worker productivity measures would

be preferred but are unattainable

◮ Frequently used approach in economics:

coefficients in a production function specification

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Relative productivity from production functions was used to measure:

◮ relative productivity by age group

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Relative productivity from production functions was used to measure:

◮ relative productivity by age group ◮ relative productivity between males and females

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Relative productivity from production functions was used to measure:

◮ relative productivity by age group ◮ relative productivity between males and females ◮ many other studies such as the productivity of

training

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Model: a non-technical overview

◮ Ideal experiment: almost identical firms

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Model: a non-technical overview

◮ Ideal experiment: almost identical firms ◮ Firms have only one difference: the proportion

• f workers with foreign and Canadian high

school equivalent education

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Model: a non-technical overview

◮ Ideal experiment: almost identical firms ◮ Firms have only one difference: the proportion

• f workers with foreign and Canadian high

school equivalent education

◮ Under some assumptions, differences in output

bw firms measure the difference in productivity between workers with Canadian and foreign high-school diploma

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Model: specification of production ln yit = 0 + 1 ln kit + 2 ln Qit + 2( 3 − 1)P1it + 2( 4 − 1)P2it + eit; yit - output of firm i in time period t, kit - capital stock, P1it, P2it - proportions of workers with Canadian and foreign high-school ed.

β4 β3 equals relative productivity

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Two possible directions to proceed: Direction I, e.g. Hellerstein et al (1999)

◮ Write down a matching wage equation to

estimate relative wages with (possibly) the same biases in the production function

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Two possible directions to proceed: Direction I, e.g. Hellerstein et al (1999)

◮ Write down a matching wage equation to

estimate relative wages with (possibly) the same biases in the production function

◮ Advantages: answers the primary question of

whether low returns to immigrants’ education, experience are productivity-based

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Two possible directions to proceed: Direction I, e.g. Hellerstein et al (1999)

◮ Write down a matching wage equation to

estimate relative wages with (possibly) the same biases in the production function

◮ Advantages: answers the primary question of

whether low returns to immigrants’ education, experience are productivity-based

◮ Disadvantages: i) what if biases affecting the

production function and wage equation are not the same? ii) cannot credibly measure the productivity differential between foreign and Canadian workers

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Two possible directions to proceed: Direction II, e.g. Dearden and van Reenen, 2006

◮ Focus primarily on estimating the relative

productivity between foreign-born and Canadian workers

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Two possible directions to proceed: Direction II, e.g. Dearden and van Reenen, 2006

◮ Focus primarily on estimating the relative

productivity between foreign-born and Canadian workers

◮ Advantages: answers both the primary question

• f whether low returns to immigrants’

education, experience are productivity-based and (reasonably) credibly measures the productivity differential between foreign and Canadian workers

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Two possible directions to proceed: Direction II, e.g. Dearden and van Reenen, 2006

◮ Focus primarily on estimating the relative

productivity between foreign-born and Canadian workers

◮ Advantages: answers both the primary question

• f whether low returns to immigrants’

education, experience are productivity-based and (reasonably) credibly measures the productivity differential between foreign and Canadian workers

◮ Disadvantages: may run into data limitations

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Sources of bias in our equations (besides the usual simultaneity, omitted variables, measurement errors in pdn fns:

◮ Booming firms may be employing more

immigrants

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Sources of bias in our equations (besides the usual simultaneity, omitted variables, measurement errors in pdn fns:

◮ Booming firms may be employing more

immigrants

◮ Why? Immigration program discourages taking

jobs away from Canadian workers. This bias would make immigrants look more productive

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Sources of bias in our equations (besides the usual simultaneity, omitted variables, measurement errors in pdn fns:

◮ Booming firms may be employing more

immigrants

◮ Why? Immigration program discourages taking

jobs away from Canadian workers. This bias would make immigrants look more productive

◮ Firms may differ in their ability to employ

foreign workers: some occupations/industries (e.g., lawyers) require country-specific skills. The direction of the bias is unknown.

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Specification of the error term: eit = i + vit + mit; vit = vit−1 + "it; "it; mit ∼ MA(0); and labor input and capital stock are possibly correlated with the time-invariant firm-specific effect i, measurement error mit, and (possibly autoregressive) productivity shocks vit.

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Main identifying assumption in equations in first differences after transformation to eliminate the AR(1) shock: E(xi1"it) = E(xi1mit) = 0 where xi1 - logged capital,

• utput, labour, and proportion of workers with

foreign and Canadian high-school diploma

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

The model in one piece ln yit = 0 + 1 ln kit + 2 ln Qit + 2( 3 − 1)P1it + 2( 4 − 1)P2it + eit; eit = i + vit + mit; vit = vit−1 + "it; "it; mit ∼ MA(0); E(xi1"it) = E(xi1mit) = 0

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Transformed equation after elimination of AR(1) process in residuals, continued ln yit = 0 + 1 ln yit−1 + 2 ln kit + 3 ln kit−1+ 4 ln Qit + 5 ln Qit−1 + 6P1t + 7P1t−1 + 8P2t + 9P2t−1 + ∗i +"it + mit − mit−1 where the common factor restrictions: 3 = − 1 2; 5 = − 1 4; 7 = − 1 6; 9 = − 1 8 can be imposed and tested after estimating vector = ( 1; :::; 9) using minimum distance

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

First differencing the main equation, System GMM

◮ We will use first differences of the above

equation

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

First differencing the main equation, System GMM

◮ We will use first differences of the above

equation

◮ Following Blundell and Bond (2000), Arellano

and Bond (1995), we will also use equations in levels

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Moment conditions: system GMM

◮ Moment conditions in first differences:

instruments lags 2 and above of levels

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Moment conditions: system GMM

◮ Moment conditions in first differences:

instruments lags 2 and above of levels

◮ Moment conditions in levels: lagged differences

starting with lagged first difference

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Moment conditions: system GMM

◮ Moment conditions in first differences:

instruments lags 2 and above of levels

◮ Moment conditions in levels: lagged differences

starting with lagged first difference

◮ Windmeijer (2005) standard errors, 2-step GMM

estimates

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Moment conditions: system GMM, technical specification

◮ Moment conditions in first differences:

E[xi,t−s△("it + mit − mit−1)] = 0 where xi,t include ln yit; ln kit; ln Qit; P1t; P2t s is 2 or above if there is no measurement error (3 or above otherwise)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Moment conditions: system GMM, technical specification

◮ Moment conditions in first differences:

E[xi,t−s△("it + mit − mit−1)] = 0 where xi,t include ln yit; ln kit; ln Qit; P1t; P2t s is 2 or above if there is no measurement error (3 or above otherwise)

◮ Moment conditions in levels:

E[△xi,t−s( ∗i +"it + mit − mit−1)] = 0

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Firm-Level Wage Equation (similar to the production function) ln Wit = 0 + ( 1 − 1)P1t + ( 2 − 1)P2t + jt where Wit is the average wage in firm i in time period t, and δ2

δ1 measures the relative wage between

workers with Canadian and foreign high-school education

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Other methods to estimate the main equation:

◮ Olley, G. and A. Pakes (1996)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Other methods to estimate the main equation:

◮ Olley, G. and A. Pakes (1996) ◮ Levinsohn, J. and A. Petrin (2003)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: Workplace and Employee Survey (WES)

◮ Canadian matched employer-employee

1999-2003 (2005)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: Workplace and Employee Survey (WES)

◮ Canadian matched employer-employee

1999-2003 (2005)

◮ Workers re-sampled every two years

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: Workplace and Employee Survey (WES)

◮ Canadian matched employer-employee

1999-2003 (2005)

◮ Workers re-sampled every two years ◮ Sample of firms updated every 5 years

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

WES: approximate data counts from Pendakur and Woodcock (2008)

◮ 58,298 employees, 7,641 workplaces

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

WES: approximate data counts from Pendakur and Woodcock (2008)

◮ 58,298 employees, 7,641 workplaces ◮ mean number of employees observed: 7.6,

median 6

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

WES: approximate data counts from Pendakur and Woodcock (2008)

◮ 58,298 employees, 7,641 workplaces ◮ mean number of employees observed: 7.6,

median 6

◮ 3,064 firms in 1999, 2001, 2003; 2,063 in two

years

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: inputs into production, descriptives

◮ Output: revenue (sales)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: inputs into production, descriptives

◮ Output: revenue (sales) ◮ Employment

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: inputs into production, descriptives

◮ Output: revenue (sales) ◮ Employment ◮ Capital: will be matched by industry/year from

table 310002, CANSIM II

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: inputs into production, descriptives

◮ Output: revenue (sales) ◮ Employment ◮ Capital: will be matched by industry/year from

table 310002, CANSIM II

◮ Price indexes will be matched by industry/year

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: creating the main variables, see Dostie (2007)

◮ Canadian - born (close to often reported figure

• f about 80%)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: creating the main variables, see Dostie (2007)

◮ Canadian - born (close to often reported figure

• f about 80%)

◮ Year of immigration, birth year are available

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: creating the main variables, see Dostie (2007)

◮ Canadian - born (close to often reported figure

• f about 80%)

◮ Year of immigration, birth year are available ◮ Years since immigration: mean 3.988 (10.181)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: creating the main variables, see Dostie (2007)

◮ Canadian - born (close to often reported figure

• f about 80%)

◮ Year of immigration, birth year are available ◮ Years since immigration: mean 3.988 (10.181) ◮ Country of origin: starting in 2003

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Data: creating the main variables, see Dostie (2007)

◮ Canadian - born (close to often reported figure

• f about 80%)

◮ Year of immigration, birth year are available ◮ Years since immigration: mean 3.988 (10.181) ◮ Country of origin: starting in 2003 ◮ Use ethnicity and immigration status in 1999,

2001

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Education: from Dostie (2007) for 1999

◮ 14 categories can be identified (completed

degrees)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Education: from Dostie (2007) for 1999

◮ 14 categories can be identified (completed

degrees)

◮ Less than high-school, high-school, some

college, u.: 0.463

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Education: from Dostie (2007) for 1999

◮ 14 categories can be identified (completed

degrees)

◮ Less than high-school, high-school, some

college, u.: 0.463

◮ Bachelor degree or above: 0.194

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Education: from Dostie (2007) for 1999

◮ 14 categories can be identified (completed

degrees)

◮ Less than high-school, high-school, some

college, u.: 0.463

◮ Bachelor degree or above: 0.194 ◮ Assume a foreign degree if age at immigration is

above a given cut-off

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Experience: from Dostie (2007)

◮ mean 16.167 (10.714)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Experience: from Dostie (2007)

◮ mean 16.167 (10.714) ◮ My challenge: compute Canadian and foreign

experience

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Experience: from Dostie (2007)

◮ mean 16.167 (10.714) ◮ My challenge: compute Canadian and foreign

experience

◮ Canadian experience: years since immigration

(Ferrer and Riddell, 2008)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Experience: from Dostie (2007)

◮ mean 16.167 (10.714) ◮ My challenge: compute Canadian and foreign

experience

◮ Canadian experience: years since immigration

(Ferrer and Riddell, 2008)

◮ Foreign experience: impute years of education

by country/degree using the Census and use potential experience

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Conclusion/Restatement of objectives in light of data limitations:

◮ At a minimum: measure relative productivity

between Canadian- and foreign-born workers (relative productivity of bundles of education/experience)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Conclusion/Restatement of objectives in light of data limitations:

◮ At a minimum: measure relative productivity

between Canadian- and foreign-born workers (relative productivity of bundles of education/experience)

◮ If possible, productivity by Canadian/foreign

experience

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Conclusion/Restatement of objectives in light of data limitations:

◮ At a minimum: measure relative productivity

between Canadian- and foreign-born workers (relative productivity of bundles of education/experience)

◮ If possible, productivity by Canadian/foreign

experience

◮ Productivity by Canadian/ foreign degree

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Conclusion/Restatement of objectives in light of data limitations:

◮ At a minimum: measure relative productivity

between Canadian- and foreign-born workers (relative productivity of bundles of education/experience)

◮ If possible, productivity by Canadian/foreign

experience

◮ Productivity by Canadian/ foreign degree ◮ Entry effects

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Challenges:

◮ Ensuring a closer match of the project objectives

with policy needs

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Challenges:

◮ Ensuring a closer match of the project objectives

with policy needs

◮ Maximizing contribution to academic literature

• n immigration

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials

Challenges:

◮ Ensuring a closer match of the project objectives

with policy needs

◮ Maximizing contribution to academic literature

• n immigration

◮ Completing this project given institutional

constraints (Stat Canada, RDC limitations)

Natalya N. Dygalo Productivity: Foreign vs Canadian Credentials