The relationship between
immigration to Australia and
the labour market outcomes
of Australian workers
Migrant Intake into Australia
Technical Supplement A
January 2016
Robert Breunig
Nathan Deutscher
Hang Thi To
Commonwealth of Australia 2016
Except for the Commonwealth Coat of Arms and content supplied by third parties, this copyright work is
licensed under a Creative Commons Attribution 3.0 Australia licence. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/3.0/au
. In essence, you are free to copy, communicate and adapt the
work, as long as you attribute the work to the Productivity Commission (but not in any way that suggests the
Commission endorses you or your use) and abide by the other licence terms.
Use of the Commonwealth Coat of Arms
For terms of use of the Coat of Arms visit the ‘It’s an Honour’ website: http://www.itsanhonour.gov.au
Third party copyright
Wherever a third party holds copyright in this material, the copyright remains with that party. Their
permission may be required to use the material, please contact them directly.
An appropriate reference for this publication is:
Breunig, R., Deutscher, N. and To, H.T. 2016, ‘The relationship between immigration to Australia and the
labour market outcomes of Australian workers’, Technical Supplement A to the Productivity Commission
Inquiry Report Migrant Intake into Australia, Canberra, April.
Publications enquiries
Media and Publications, phone: (03) 9653 2244 or email: maps@pc.gov.au
The Productivity Commission
The Productivity Commission is the Australian Government’s independent research and
advisory body on a range of economic, social and environmental issues affecting the welfare of
Australians. Its role, expressed most simply, is to help governments make better policies, in the
long term interest of the Australian community.
The Commission’s independence is underpinned by an Act of Parliament. Its processes and
outputs are open to public scrutiny and are driven by concern for the wellbeing of the
community as a whole.
Further information on the Productivity Commission can be obtained from the Commission’s
website (www.pc.gov.au).
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
1
The relationship between immigration to
Australia and the labour market outcomes
of Australian workers
Robert Breunig, Nathan Deutscher and Hang Thi To
*
Australian National University
15 January 2016
Abstract
We examine the relationship between immigration to Australia and labour market
outcomes of the Australian-born and previous immigrant cohorts. We use
immigrant supply changes in skill groups — defined by education and experience
— to identify the impact of immigration on the labour market. We find that
immigrants flow into those skill groups that have the highest earnings and lowest
unemployment. Once we control for the impact of experience and education on
labour market outcomes, we find almost no evidence that immigration has
harmed, over the decade since 2001, the aggregate labour market outcomes of
those born in Australia (natives) as well as incumbents (natives and previous
immigrants).
Keywords: immigration; Australia; native labour market outcomes; incumbent labour
market outcomes.
JEL Codes: J21,J31,J61,F22
A.1 Introduction
The impact of immigration on Australians, particularly on their wages and their
employment prospects, is a question that can provoke heated and emotional debate.
Anecdote and visceral impressions can easily dominate either side of the public
*
We gratefully acknowledge financial support from the Productivity Commission in preparing this
manuscript. This paper uses unit record data from the Household, Income and Labour Dynamics in
Australia (HILDA) Survey. The HILDA Project was initiated and is funded by the Australian
Government Department of Social Services (DSS) and is managed by the Melbourne Institute of Applied
Economic and Social Research (Melbourne Institute). The findings and views reported in this paper,
however, are those of the author and should not be attributed to either DSS, the Melbourne Institute or the
Productivity Commission. All errors are those of the authors.
2
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
conversation. In this paper, we look carefully at the data to see if we can discern an effect
of immigration on the labour market outcomes of Australian workers. We look at outcomes
for two groups: those born in Australia (natives) as well as natives and previous
immigrants (incumbents).
A standard competitive labour market model suggests that immigration should have a
negative impact on wages. An influx of immigrants shifts the supply curve to the right,
depressing wages. This simple theoretical model, however, may fail to capture a variety of
other economic phenomena that may offset the negative wage effect.
One possibility is that the immigrant influx is part of a demand shift in the overall
economy. The demand shift would have the effect of raising wages and could dominate the
supply shift, resulting in higher wages for all. Another possibility is that immigrants may
fill roles that would otherwise be unfilled (e.g. mine workers, nurses or fruit pickers) and
the presence of these workers actually lifts the productivity (and wages) of incumbent
workers in related employment. The supply of capital, the characteristics of these new
workers and the structure of technology will all matter in determining the overall effect of
immigration on wages across the economy.
Congruent with this muddy theoretical picture, the literature paints a very mixed picture of
the effect of immigration on labour market outcomes of both natives and the broad group
of incumbent workers. Early literature in the United States pointed towards very small
effects of immigration on natives in that country (Friedberg and Hunt 1995 and Smith and
Edmonston 1997). Using a novel approach that moved away from geographical
identification and more towards skill-based identification, Borjas (2003) finds that the
employment opportunities of US natives have been harmed by immigration. More recently,
Ottaviano and Peri (2012) and Manacorda, Manning and Wadsworth (2012), extending and
refining Borjas’ work, find evidence for varying effects across population subgroups in the
US and UK respectively, with at times positive effects for native-born workers as a whole
sitting alongside negative effects for less educated natives and past migrants.
The above papers differ in their assumptions about the changing nature of capital, the
definition and size of skill groups and the substitutability of different types of labour.
Varying these assumptions appears to have a significant impact on the measured effects of
immigrants on labour market outcomes.
In this paper, we employ the approach of Borjas (2003). We divide the national labour
market into skill groups based upon education and experience. We examine whether
changes in the fraction of immigrants in skill groups are associated with labour market
outcomes for those working in Australia, after controlling for other factors. There are two
main advantages of our approach. First, it is data-driven and asks a simple correlation
question in a non-parametric way. Second, it allows for geographic mobility in labour
markets, which is ruled out in approaches that use the spatial distribution of immigrants for
identification.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
3
We take two distinct approaches to defining the distinction between immigrants and
Australian workers, varying in their treatment of earlier migrants. This difference is
important, since around one-quarter of the Australian population is born overseas.
We first define immigrants as anyone born outside of Australia and focus on the labour
market outcomes of the Australian-born. We then consider the relationship between
outcomes for incumbents (those born in Australia plus those who migrated to Australia
five or more years previously) and recent (less than five years in Australia) migrants. We
examine a variety of outcomes: weekly earnings, annual earnings, hourly wage, weekly
hours worked, labour force participation and employment.
The analysis in this paper is restricted to considering effects of immigration on the labour
market outcomes of Australian workers, not their welfare more broadly considered. Such
an analysis is well beyond the scope of this paper.
We use three different data sets for our analysis. In one set of analysis we use the
Australian Bureau of Statistics (ABS) series of Surveys of Income and Housing (SIH) to
estimate the number of migrants and non-migrants in each skill group. We use the same
data to measure the labour market outcomes of the Australian born. In a second set of
analysis, we match census data to the Household, Income and Labour Dynamics in
Australia (HILDA) survey. In this case we use HILDA to estimate many of the labour
market outcomes of the Australian born but use complete census data to determine the
number of migrants and non-migrants in different skill groups. Results across both sets of
data are quite similar.
We find strong evidence of immigrant selection. That is, immigration flows into skill
groups where wages and employment are high. This is most likely a result of both
government policy and of the labour market decisions of immigrants. We find almost no
evidence that outcomes for those born in Australia have been harmed by immigration, with
the most statistically significant associations being with stronger labour market outcomes
for the Australian born. For incumbents, we find a negative relationship between
immigration and incumbent wages. However, this relationship is driven entirely by
highly-educated female workers with 10 years or less experience. This effect disappears
when we consider more precise skill groupings. Considered overall, the evidence suggests
that incumbent labour market outcomes have been neither helped nor harmed by
immigration.
In the next section, we discuss the definition of skill groups and the methodology that we
use. In section 3, we present the data. Empirical results are in section 4. As is the case with
all empirical work, the results are subject to certain caveats and these are discussed in
detail in section 5. We also provide some conclusions in this last section.
4
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
A.2 Methodology and related Australian literature
Our analysis examines the effect of immigration on labour market outcomes of Australian
workers using the national labour market approach (e.g. Borjas, 2003, 2006). In our
implementation of this approach, individuals are classified into five distinct educational
groups:
• high school dropouts (persons whose highest level of education was year 11 or below);
• high-school graduates (persons whose highest level of education was year 12);
• diploma graduates without year 12 education (persons who obtained a certificate or a
diploma but did not complete year 12);
• diploma graduates after completing year 12 (persons who obtained a certificate or a
diploma after having completed year 12); and
• university graduates (persons whose highest education was either a undergraduate or
post-graduate degree, or a graduate diploma certificate, after having completed
year 12).
Individuals are also classified into eight experience groups based on the number of years
that have elapsed since the person completed school.
1
We assume that the age of entry into
the labour market is:
• 17 for a typical high school dropout;
• 19 for a typical high-school graduate as well as for a typical diploma graduate without
year 12 education;
• 21 for a diploma graduate after completing year 12; and
• 23 for a typical university graduate after completing year 12.
The work experience is then given by the age of the individual minus the age at which the
individual entered the labour market. We restrict our analysis to people who have between
1 and 40 years of experience and aggregate the data into eight experience groups with
five-year experience intervals such as 1 to 5 years of experience, 6 to 10 years of
experience, and so on.
The individual data is aggregated into different education-experience cells. For each of
these cells, the share of immigrants in the population is given by:
=
+
where M
ijt
is the number of immigrants in cell (i, j, t), and N
ijt
is the number of
Australia- born individuals in cell (i, j, t).
1
In essence, we measure potential experience. This will be different for people of the same age depending
upon the age at which they finished their schooling/education. We refer to this as experience throughout.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
5
We estimate the following specification:
=
+
+
+
+ (
×
) +
(
×
)
+
×
+
(1)
where:
• y
ijt
is the mean value of a particular labour market outcome for Australia-born workers
in cell (i, j, t);
• s
i
is a vector of dummy variables for education groups (i=1 to 5);
• x
j
is a vector of dummy variables for experience groups (j=1 to 8);
•
is vector of dummy variables for time (5 time periods for the SIH data and 3 time
periods for the matched HILDA / census data);
•
is a normally distributed random error.
The model includes time dummies to account for changes in the macroeconomic
environment that affect all groups. By including dummies for education and experience
and their interaction, we account for the supply and demand factors specific to each skill
group that determine the overall level of labour market outcomes for that skill group.
2
Interacting education and experience with time dummies allows the profile of skill groups
to evolve differently over time.
Identification in the model comes from changes within skill groups over time.
3
Differences
in the changes in the proportion of immigrants within cells are related to differential
changes in labour market outcomes. The approach is non-parametric in the sense that we
are allowing the data to relate changes in immigration to changes in labour market
outcomes without imposing any structural restrictions on this relationship. (We do not
estimate a wage equation, for example.) There is no need to control for other
characteristics such as average occupation or industry within a cell since these effects and
their evolution over time are perfectly captured by the fixed effects and the interactions.
One previous Australian paper used this approach. Bond and Gaston (2011) used only the
HILDA data to assess the effects of immigration on weekly earnings and weekly hours
worked of Australian-born workers. They found that immigrant share has a positive effects
on Australian-born workers’ earnings and weekly hours worked. Their approach is flawed
however because they used HILDA for both the outcome data and the immigrant share
data.
Since HILDA is a panel with an initial sample chosen in 2001, there is no inflow of
migrants into the sample.
4
The year-on-year change in the share of immigrants in the
2
These dummies allow the observed equilibrium outcomes to differ for each skill group. These observed
equilibrium outcomes could be driven by both demand and supply factors.
3
Using a model specified in first-differences gives similar results for the key coefficient,
.
4
Prior to the top up sample in 2011.
6
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
HILDA sample is driven by two factors: differential sample attrition of migrants and
non-migrants and a small number of migrants who join the sample because they partner
with a continuing sample member (or join the HILDA sample through one of the other
following rules of the data). Overall, population immigrant flows cannot be captured in any
meaningful sense through this panel data set.
Sinning and Vorell (2011) investigate attitudes towards, and the effects of, immigration on
the labour market and crime. They estimate the effect of immigration on SLA median
income and unemployment and LGA crime rates. They use data from 1996, 2001 and 2006
Censuses and crime statistics. To address selection issues, they instrument immigration
stock in a period with a counterfactual immigration stock created under the assumption that
new immigrants settle according to the last-period distribution of immigrants. The second
stage regressions include regional controls such as median age, population size,
educational and occupational distributions and region and time fixed-effects. In neither of
these preferred models is the immigration coefficient statistically significant. However,
their instrument is weak, with a first stage F-statistic below 10 when both period and time
fixed effects are included, clouding the interpretation of these results.
The geographic approach of Sinning and Vorell (2011) (and many others) has come under
increasing attack since Borjas (2003). The approach assumes that geographic labour
markets are fixed and distinct. Yet, we know that there are important movements of both
firms and workers that tend to equalize economic conditions across cities and regions. In
Australia, this trend is strongly seen in a shift of innovative activity and employment from
Victoria and New South Wales to Queensland and Western Australia during the time of
our data window.
Our approach allows for a national-level labour market but assumes no substitutability
across skill groups. Essentially, we assume fixed and distinct labour markets defined by
skill groups (rather than by sub-national geographic). Workers and firms are assumed to be
unable to change the skill group in which they supply or demand labour in response to
prices. Given that skill groups are defined broadly and in terms of experience and
education levels that are not able to be altered by workers, this assumption seems less
problematic than strict geographical segregation. Mobility across occupations, industries
and regions does not affect identification. The restriction that workers compete in skill
groups defined by education and experience is an important one and is discussed further in
sections 4.1 and 5.
A.3 Data
Our analysis is grouped into two parts. In the first part, we use data drawn from the SIH
conducted by the ABS. We use data from five biennial surveys from 2003 to 2012. The
survey collects information from usual residents of private dwellings in urban and rural
areas of Australia, covering about 98% of all people living in Australia. Private dwellings
are houses, flats, home units, caravans, garages, tents and other structures that were used as
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
7
places of residence at the time of interview. Long-stay caravan parks are also included.
These are distinct from non-private dwellings, such as hotels, boarding schools, boarding
houses and institutions, whose residents are excluded. The SIH contains a wide range of
information on demographic and economic characteristics of individuals and households.
In the second part of our analysis, we use data drawn from the Household, Income and
Labour Dynamics in Australia (HILDA) combined with data from the Australian Census of
Population and Housing (Census).
The HILDA survey is a household-based panel study that collects information on
respondents’ economic and demographic characteristics. The wave 1 HILDA survey was
conducted in 2001 and has been conducted annually since. The vast majority of data was
collected through face-to-face interviews and a small fraction of the data was collected
through telephone interviews. 13 969 people were interviewed in wave one from
7682 households. The survey has grown slightly over time as all individual sample
members and their children are followed. The sample was replenished in wave 11 with a
top-up sample of 4009 people added in the survey.
The Australian Population and Housing Censuses provide information on the number of
people in each part of Australia, what they do and how they live. The data record the
details of all people (including visitors) who spend the night in each dwelling on Census
Night. Immigrants are included in the census provided that they intend to stay in Australia
for at least one year. The census data thus excludes those who intend to stay in Australia
for less than one year.
5
Census data contains information on topics such as age, gender,
education, birthplace and employment status of all people in Australia on Census Night.
6
In the first part of our analysis, we estimate the model of equation (1) using SIH data for
five financial years 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012. We only
use data from 2003 onwards. Survey years prior to 2003-04 group education in broader
categories that are different than those used in 2003-04 and onwards. This makes it
impossible for us to extend our chosen skill group definitions further back in time than
2003.
We estimate the model for six different dependent variables relating to the labour market
outcomes of Australian-born workers: annual earnings from wage and salary, weekly
earnings from wage and salary, log hourly wage rate, weekly hours worked, the labour
force participation rate and the unemployment rate. The key explanatory variable of
interest, the share of immigrants in each education/experience cell, is also extracted from
5
We thank Jenny Dobak of the ABS for clarifying this.
6
We use the entire census data to construct the fraction of immigrants in each skill group. For 2006 and
2011, this data is available online through ABS table builder. For 2001, the data was constructed for us by
the ABS and provided through the Productivity Commission. We thank Meredith Baker and Troy
Podbury of the Productivity Commission and Steve Gelsi and Dominique O’Dea of the ABS for their
assistance in procuring the data. We also thank Sharron Turner at ANU for her assistance in helping us to
access ABS data.
8
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
the SIH as the survey samples, properly weighted, are representative cross-sections in each
year.
In the second part, we estimate the model of equation (1) using HILDA data combined
with complete Census data for 2001, 2006 and 2011. The explanatory variable of interest,
the share of immigrants in each skill group, is extracted from Census data. For the
dependent variables (labour market outcomes) we use the Census data for the
unemployment rate and the labour force participation rate of Australian-born workers. Data
for weekly hours worked, weekly and annual earnings (i.e. labour income) and hourly
wage rates are extracted from HILDA data as Census data do not provide individual
earnings in continuous values. We use cross-sectional weights from HILDA to make the
cell means representative. The weighted and unweighted means are almost identical. The
necessity of using immigrant share from Census data comes from the fact that the share of
immigrants in HILDA is not an appropriate indicator for the changing immigrant share in
Australia over time, as discussed above.
Descriptive statistics, from the SIH, of the main variables used in the analysis are provided
in figures A.1 to A.6. Figure A.1 presents the migrant share for each education-experience
cell, grouped by education category. For young people, migrant shares are relatively higher
in groups with university education compared to groups without university education. This
reflects the shift towards a higher skill requirement in Australian immigration policy in
recent years as well as strong labour market demand in Australia for highly educated
people.
Figure A.2 presents the mean values of annual earnings of Australian-born workers by
education and experience, grouped by education category. With the same experience,
annual earnings are higher for people with higher educational attainment. Annual earnings
increase faster for the young. The effect of experience is smaller after 20 years of
experience. For all groups we see the usual inverted U-shape earnings/experience profile.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
9
Figure A.1 Migrant share by Education and Experience: SIH
Figure A.2 Annual earnings of Australian born workers by education
and experience: SIH
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
Year 11 Year 12 Diploma without
year 12
Diploma after
year 12
University
Proportion o migrants to total population
1-5 year experience 6-10 year experience 11-15 year experience 16-20 year experience
21-25 year experience 26-30 year experience 31-35 year experience 36-40 year experience
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
Year 11 Year 12 Diploma without
year 12
Diploma after
year 12
University
1-5 year experience 6-10 year experience 11-15 year experience 16-20 year experience
21-25 year experience 26-30 year experience 31-35 year experience 36-40 year experience
$ per year
10
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
Figure A.3 shows the mean annual earnings of Australian born workers by education and
experience, respectively. We see very strong returns to university education and again an
inverted U-shape experience/earnings profile.
Figure A.3 Annual earnings of Australian born workers by education
and experience groups
Education
Experience
0
10000
20000
30000
40000
50000
60000
70000
80000
High school
dropout
High school
graduate
Diploma
graduates
without year
12
Diploma
graduates
after year 12
University
graduates
Total
$ per year
0
10000
20000
30000
40000
50000
60000
70000
1-5
years
6-10
years
11-15
years
16-20
years
21-25
years
26-30
years
31-35
years
36-40
years
Total
$ per year
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
11
Figure A.4 presents the unemployment rate of Australian born workers by education and
experience groups. The figures show that the unemployment rate decreases with the level
of education and with experience; the exception is slightly higher unemployment for those
in the highest experience group.
Figure A.4 Unemployment rate of Australian born workers by education
and experience groups
Education
Experience
0
1
2
3
4
5
6
7
8
9
High school
dropout
High school
graduate
Diploma
graduat es
without year
12
Diploma
graduat es
after year
12
University
graduat es
Total
Per cent
0
1
2
3
4
5
6
7
8
9
1-5
years
6-10
years
11-15
years
16-20
years
21-25
years
26-30
years
31-35
years
36-40
years
Total
Per cent
12
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
Figure A.5 presents migrant share by education and experience from the Census data and
figure A.6 shows annual earnings by education and experience from HILDA. The overall
impression provided by the two data sets is quite similar.
Figure A.5 Migrant share by education and experience, Census
Figure A.6 Annual earnings by education and experience, HILDA
0.0
0.1
0.2
0.3
0.4
0.5
0.6
1-5 years 6-10 years 11-15
years
16-20
years
21-25
years
26-30
years
31-35
years
36-40
years
Share
Dropout Y12 Cert w/o Y12 Cert w Y12
Degree
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
1-5 years 6-10 years 11-15
years
16-20
years
21-25
years
26-30
years
31-35
years
36-40
years
Income ($)
Dropout Y12 Cert w/o Y12 Cert w Y12 Degree
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
13
Figures A.7 and A.8 show the distribution of changes over time in the key variable
in
the two data sets — SIH and Census. The model is identified from these changes and the
key empirical question is: are changes in the share of immigrants in total workers
statistically related to labour market outcomes of Australian-born workers over the sample
period? We can see that in both data sets, the changes in the share of migrants is centered
around zero and is fairly small – while we do observe both positive and negative changes,
this will limit our ability to detect any effect of immigration on labour market outcomes.
In the Census, we find that the average proportional change in migrant share (pooling
across the two time periods) is 0.0022. The minimum is -0.07 and the maximum is .10. In
the SIH, the average is slightly negative (-0.0049), the minimum is -0.13 and the maximum
change is 0.18. The migrant share changes calculated from the SIH have a slightly higher
variance than those calculated from the Census. In general, across both data sets, the larger
changes are for the most highly educated groups who saw positive increases in the share of
immigrants over time. The two groups with certificates (year 12 and no year 12) saw the
largest decreases in immigrant share.
Figure A.7 Distribution of migrant share changes between periods: SIH
data
0
5
10
15
20
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
14
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
Figure A.8 Distribution of migrant share changes between periods:
Census data
A.4 Empirical results
We estimate models of the labour market outcomes of Australian-born workers (including
annual earnings, weekly earnings, weekly hours worked, hourly wage rate, labour force
participation, and unemployment rate) against the share of migrants with different
specifications: (i) models that include only the time dummy variables; (ii) models
controlling for all dummy variables including dummies for education groups, for
experience groups, and dummies for time but without any interaction terms; (iii) models
controlling for education, experience, time and the interactions between dummy variables
that allow for changing skill premia over time.
We present weighted regressions using the weights defined as the number of
Australian-born in each education-experience cell for whom the relevant outcome variable
is defined. That is, we weight labour force participation regressions by the native
population, unemployment regressions by the native labour force, and hours and earnings
regressions by the number of natives employed. We also present unweighted estimates for
comparison. In all of our models, we present standard errors that control for clustering on
education-experience cells to allow for serial correlation in the estimates.
The results from SIH data are presented in tables A.1 and A.3 and results from HILDA
wage and earnings data matched to census data for immigrant shares by
experience/education cells are reported in tables A.2 and A.4. In our discussion of these
results we begin with the broad, overarching story coming out of the coefficients, before
turning to individual coefficients that may be of particular interest.
0
5
10
15
20
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
15
Empirical results: Survey of Income and Housing
Table A.1 presents the results for the full sample from the SIH. In the first row, we
estimate a model that includes only time dummies and no controls for education or
experience. Row two presents results where we add the controls for education and
experience levels, but no interactions between the two. Row three presents the results when
we add the full set of skill controls including interactions between education and
experience and interactions with time which allow skill premia to vary across time.
Unweighted estimates are provided in row four for comparison.
Table A.1 Estimated values of
from equation (1): SIH, full sample
Log annual
earnings
Log weekly
earnings
Log of wage
rate
Weekly
hours
Participation
rate
Unemployment
rate
Weighted, time dummies only
θ 1.879*** 1.650*** 1.510*** 7.480** 0.240* -0.205***
(0.360) (0.301) (0.231) (2.991) (0.120) (0.055)
Weighted, education, experience and time dummies but no interactions
θ -0.090 -0.086 -0.144** 0.089 0.108 -0.017
(0.143) (0.135) (0.068) (3.124) (0.111) (0.053)
Weighted; education, experience and time dummies and their interactions — preferred estimates
θ 0.175 0.021 -0.077 6.983 0.525** -0.021
(0.154) (0.169) (0.205) (4.190) (0.250) (0.043)
Unweighted; education, experience and time dummies and their interactions
θ .388** 0.179 0.035 8.549*
.464** -0.035
(0.177) (0.186) (0.196) (4.662) (0.207) (0.04)
Note: *,**,*** indicate statistical significance at the 10%, 5%, and 1% significance level respectively.
The weighted estimates with a full set of shift and interaction dummies (row three) are our
preferred model in all of the tables. We primarily discuss these weighted results.
For models that only include time dummies, we find a positive relationship (and
statistically significant) between immigration and wages (measured as yearly earnings,
weekly earnings or hourly wage) in the sense that more immigration is correlated with
higher wages. Immigration is also correlated with higher labour force participation and
lower unemployment.
For the models that include all dummy variables and their interactions, we find little
statistical relationship between immigration and wages or other labour market outcomes
(participation or unemployment). There does appear to be some small statistical association
16
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
between immigration and a higher participation rate. This association is quite small. If the
share of immigrants goes up by 1 percentage point (from say 20% to 21%), this is
associated with a 0.5 percentage point increase in the participation rate. Recall from
figure A.9 that the typical changes are very small — on the order of one percentage point.
Empirical results: HILDA combined with Census data
The results for the HILDA/Census data are quite similar (table A.2). We find a strong
association between Australian-born labour market outcomes and immigrant shares when
we do not control for different returns to experience and education. Once we include a full
set of dummies, these associations disappear. We find no statistically significant
associations.
Table A.2 Estimated values of
from equation (1): HILDA and Census,
full sample
Log annual
earnings
Log weekly
earnings
Log of wage
rate
Weekly
hours
Participation
rate†
Unemployment
rate†
Weighted, time dummies only
θ
2.016*** 1.821*** 1.686*** 4.682 0.241** -0.244***
(0.404) (0.337) (0.245) (4.193) (0.119) (0.066)
Weighted, education, experience and time dummies but no interactions
θ
0.210 0.455*** 0.243* 6.315 -0.007 -0.015
(0.185) (0.154) (0.130) (6.010) (0.089) (0.058)
Weighted; education, experience and time dummies and th
eir interactions — preferred estimates
θ
0.267 0.752 0.612 11.349 0.074 0.076
(0.666) (0.607) (0.413) (14.997) (0.081) (0.047)
Unweighted; education, experience and time dummies and their interactions
θ
-0.061 0.534 0.622 13.922 0.034 0.061
(0.714) (0.634) (0.476) (14.987) (0.071) (0.038)
Note: *,**,*** indicate statistical significance at the 10%, 5%, and 1% significance level respectively.
†Calculated from Census; otherwise calculated from HILDA.
Overall, the results show strong evidence for migrant selection. We reach this conclusion
because we observe that when we add no controls (except time dummies), there is a very
strong positive association between labour market outcomes and immigration. This could
lead one to erroneously conclude that immigrants are ‘causing’ positive labour market
outcomes.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
17
When we control for differential returns to experience and education and changes to those
returns over time (by including the full set of education and experience dummy variables
and their interactions), we find that the positive association disappears. The positive
correlation observed in row 1 of tables A.1 and A.3 is thus ‘spurious’ in that what we are
picking up is that immigrants are associated with high skill levels and labour market
outcomes are also associated with high skill levels. Once we control for this association,
the ‘causal’ effect of immigration on labour market outcomes (in row 3 of tables A.1
and A.3) becomes mostly statistically insignificant. Further, there is not a clear and
consistent story looking at the signs and sizes of the coefficients – some are consistent with
stronger labour market outcomes (higher wages, higher participation rates and hours and
lower unemployment), and some with weaker labour market outcomes.
Thus, migrants are flowing into those skill groups that have the highest earnings and the
best employment opportunities. This is the result of government policy but also of the
decisions by potential migrants, which determine which type of migrant comes to
Australia.
Once we account for the differential returns to experience and education, we find no
evidence across the sample that immigration is associated with worse labour market
outcomes for Australian-born workers. In the SIH data, there is a small statistical
association between immigration and a higher participation rate among Australian-born
workers. This association is small in size and only significant at the 10 per cent level.
Empirical results: Separate estimation by male and female
Tables A.1 and A.2 pooled all individuals. We also re-estimate the models, splitting the
sample by male/female. (See tables A.3 and A.4 for SIH and HILDA/Census,
respectively.) In what follows, unless otherwise specified, we present results from our
preferred specification where we control for a full set of dummies and interactions. The
patterns that we observe in tables A.1 and A.2 — positive selection by immigrants when
we do not control for returns to education and experience and weighted and unweighted
estimates which are roughly similar — are repeated for all of our models. These full results
are available from the authors upon request.
For males, in both data sets, we find no statistically significant association between
immigration and labour market outcomes. In SIH, we find positive associations at the
10 per cent significance level between immigration and hours worked and labour force
participation in the female sub-sample. Using the Census data, we find a positive
association between immigration and the unemployment rate for females. More
immigration seems related to more unemployment. The effect is significant at the 5 per
cent level, but very small and only for females. If the share of immigrants goes up by
5 percentage points, the unemployment rate for females increases by about 0.6 percentage
points. Note that we only find this effect in the Census data. The coefficient for females in
the SIH data is actually negative, although not statistically significant.
18
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
The model of equation (1) imposes a constant response parameter, , across all experience
and education groups. Given the large number of fixed effects in the model, it is not
possible to estimate a model with a parameter that varies by skill group.
It may be that the labour market outcomes of different types of workers have different
responses to immigration in which case the assumption of a constant response parameter
would be incorrect. To test this hypothesis, at least somewhat, we estimate the model for a
sub-population of people with experience less than or equal to 15 years. We again estimate
models where we pool across all individuals as well as separately by male and female.
The results are broadly consistent with what we find in the main sample. For the SIH
(table A.3) the only statistically significant relationship that we find is for females.
Specifically, we find that increased immigration is associated with decreased
unemployment. If the share of immigrants goes up by 5 percentage points, this is
associated with a drop in the unemployment rate for females of about 0.9 percentage
points.
Table A.3 Estimated values of
from equation (1): SIH, selected
sub-samples
Log annual
earnings
Log weekly
earnings
Log of wage
rate
Weekly
hours
Participation
rate
Unemployment
rate
Males only
θ 0.064 0.064 0.068 -0.848 0.131 -0.037
(0.164) (0.181) (0.196) (3.226) (0.101) (0.051)
Females only
θ 0.155 0.153 -0.029 8.112* 0.209* -0.039
(0.184) (0.170) (0.203) (4.803) (0.104) (0.050)
All individuals with 15 years of experience or less
θ 0.247 -0.082 -0.254 3.465 0.175 -0.098
(0.332) (0.445) (0.406) (9.117) (0.207) (0.094)
Males with 15
years of experience or less
θ 0.298 0.240 0.359 -5.202 -0.049 0.033
(0.222) (0.278) (0.398) (3.885) (0.106) (0.087)
Females with 15 years of experience or less
θ 0.071 -0.122 -0.038 7.417 0.100 -0.189*
(0.348) (0.354) (0.586) (7.253) (0.160) (0.099)
Models include full set of time dummies, education and experience fixed effects and full set of interactions
Note: *,**,*** indicate statistical significance at the 10%, 5%, and 1% significance level respectively.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
19
In the combined HILDA / Census data (table A.4), we find no relationship between any of
the earnings variables and immigration for this less experienced group. We do find a weak
positive association between immigration and participation in the full sample of less
experienced people. We again find a positive relationship between immigration and
unemployment for females. Note the contrast with SIH where we find a negative
relationship between immigration and unemployment for females.
Table A.4 Estimated values of
from equation (1): HILDA and Census,
selected subsamples
Log annual
earnings
Log weekly
earnings
Log of wage
rate
Weekly
hours
Participation
rate†
Unemployment
rate†
Males only
θ
0.792 1.213 1.166 16.878 0.009 0.037
(0.814) (0.832) (0.704) (16.506) (0.053) (0.039)
Females only
θ
-1.105 -0.486 -0.673 8.443 -0.033 0.112**
(0.784) (0.747) (0.531) (18.539) (0.092) (0.050)
All individuals with 15 years of experience or less
θ
0.038 0.593 0.230 -4.133 0.180* 0.167
(0.432) (0.504) (0.694) (24.168) (0.096) (0.110)
Males with 15 years of experience or less
θ
0.335 0.975 1.020 5.704 0.059 0.083
(0.841) (0.809) (0.735) (26.580) (0.076) (0.079)
Females with 15 years of experience or less
θ
-0.691 -0.370 -0.773 -7.373 -0.002 0.256*
(1.295) (1.259) (0.840) (32.681) (0.101) (0.134)
Models include full set of time dummies, education and experience fixed effects and full set of interactions
Note: *,**,*** indicate statistical significance at the 10%, 5%, and 1% significance level respectively.
†Calculated from Census; otherwise calculated from HILDA.
Empirical results: Incumbents
Throughout this paper so far, we have compared immigrants (as those born outside
Australia) to those born in Australia. But Australia has a very large stock of immigrants
who, while born outside of Australia, have lived in Australia for a long time. To check if
our results are driven by how we classify individuals, we re-estimate the model comparing
‘incumbents’ to ‘recent immigrants’. We define incumbents as those born in Australia plus
20
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
those who have migrated to Australia more than five years previously. ‘Recent
immigrants’ are now defined as those who migrated to Australia within the last five years.
We estimate the labour market outcomes of incumbents as a function of the share of recent
immigrants in overall population. Weights are now defined based upon the number of
incumbents rather than the number of Australian-born. We only estimate models using the
Census / HILDA data. In the SIH, we do not have precise enough information about year
of arrival in Australia to distinguish between incumbents and recent arrivals. Results for
the full sample are provided in table A.5. We show results without controls and with
controls and weighted and unweighted for comparison with table A.2.
Table A.5 Estimated values of
from equation (1): HILDA and Census,
full sample
incumbents compared to recent immigrants
Log annual
earnings
Log weekly
earnings
Log of wage
rate
Weekly
hours
Participation
rate†
Unemployment
rate†
Weigh
ted, time dummies only
θ
0.142 0.529 0.564 -0.411 0.915*** -0.116
(1.260) (1.116) (0.813) (14.295) (0.235) (0.079)
Weighted, education, experience and time dummies but no interactions
θ
0.211 0.141 -0.028 9.603 0.298** -0.434***
(0.316) (0.296) (0.287) (12.951) (0.132) (0.125)
Weighted; education, experience and time dummies and their interactions
— preferred estimates
θ
0.437 0.519 -0.516 35.527 0.287** 0.101
(1.108) (1.024) (0.654) (31.419) (0.135) (0.095)
Unweighted; education, experience
and time dummies and their interactions
θ
-0.224 -0.049 -0.647 26.260 0.280* 0.111
(1.220) (1.181) (0.917) (32.177) (0.146) (0.084)
Note: *,**,*** indicate statistical significance at the 10%, 5%, and 1% significance level respectively.
†Calculated from Census; otherwise calculated from HILDA
The only statistically significant effect we find is a positive association between the
participation rate and immigration. If the share of recent immigrants goes up by 5
percentage points, this is associated with an increase in labour force participation of
incumbents of about 1.4 percentage points. When we compare tables A.2 and A.5, it
appears that the effect of selection is much stronger when we compare Australian-born to
all immigrants than when we compare incumbents to recent immigrants. This is a
somewhat counterintuitive result – more recent migrants might be expected to be more
likely to enter strong labour markets. That said, the signs of the coefficients remain broadly
consistent with positive selection.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
21
We also split the samples by male and female. For males, none of the coefficients are
statistically significant. For females, we find a positive association between recent
immigration and incumbents’ weekly hours.
7
Empirical results: overarching summary
Overall, across all of these estimates, our results indicate that immigration is higher into
those skill groups (defined by education and experience) that have higher wages and better
labour market prospects. This is consistent with immigrants coming to Australia with
knowledge of where returns are high and is also consistent with selective migration
policies.
Once we control for this selection into skill groups by immigrants, there is very little
evidence of any negative labour market effects, in aggregate, on those born in Australia or
the broader group of incumbents resulting from immigration.
Are immigrants and Australian-born workers in same skill groups
comparable?
A key element of our model is the assumption that migrants and Australian-born workers
compete within the same education/experience cells (skill groups). It could be that
experience and education obtained outside of Australia has a lower value in the local
labour market and that in fact migrants are competing with the Australian-born at lower
levels of experience and education. This would mean that we have misclassified some
individuals as competing in one skill group when they should actually be in another, lower
skill group.
First, it is important to note that misclassification by itself poses no threat to our
identification strategy. We identify the effects in the model from changes in the share of
migrants. Mis-classification poses no problem unless the degree of misclassification is also
changing over time.
Nonetheless, it is important to see if immigrants and Australian-born individuals within
skill group cells look similar. In table A.6, we present the three most common occupations
for migrants and natives by education and 10-year experience groupings. The two groups
look very similar, particularly where levels of education are highest. If we think of
anecdotes where overseas-trained doctors are driving taxis in Australia, this might be the
group for whom we would worry the most about misclassification. Yet, the top three
occupations are the same, and in the same order for both immigrants and Australian-born.
Australian-born individuals with higher education are between 6 and 15 percentage points
more likely to be professionals than comparable immigrants, so there is some evidence for
7
These results are available from the authors.
22
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
higher occupational status for the highly educated if they are Australian-born. However,
within our sample there is not evidence of large-scale occupational downgrading by
migrants.
In tables A.7 and A.8, we present the Duncan index of dissimilarity comparing native and
migrant occupational distributions (at the one digit level) holding either education
(table A.7) or experience (table A.8) constant. This index captures the proportion of either
group that would need to change occupations to make the two distributions equal. The
more similar the occupational distributions, the smaller the index. We have highlighted the
smallest values in each row and column.
The results are comforting in the sense that the occupational profiles of immigrants and
natives are most similar within the same education-experience cell, in general. Within
education groups, less experienced migrants look most similar to less experienced natives.
However, highly experienced migrants look more similar to moderately experienced
natives, so there may be some discount placed on overseas experience. Within experience
groups, migrants almost always look most similar to natives with the same education.
Robustness check: broader skill classifications
As a final check on our classification of skill groups, we re-estimate all of the models with
fewer education-experience cells. Some authors have argued that wider skill groups are
better as the assumption of no competition across skill groups is more likely to hold when
skill groups are more broadly defined. We re-estimate all the models using 12 groups —
3 educational groups (high school dropout; university graduates; all others) and
4 experience groups defined by 10 year groupings.
8
The results are quite similar to those already presented.
9
We begin by discussing the effect
of immigration on outcomes for the Australian-born. For the SIH data, the only significant
associations are a positive relationship between hours and immigration and a negative
relationship between unemployment and immigration when we pool male and female
together. The coefficients are 11.8 and -0.08 and are just significant at the 10% level.
When we split the sample by sex we find no statistically significant coefficients. For the
combined HILDA/Census data, we only find a statistically significant association between
immigrants and the participation rate. The coefficient in the pooled sample is 0.40. We find
a statistically significant estimate of .252 for males. We find no effect for females.
8
Figure A.3, the middle 3 educational categories which we have combined together have very similar
average earnings.
9
For this reason we only discuss the results and do not present full tables. These are available from the
authors upon request.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
23
Table A.6 Three most common occupations by skill group and migrant
/ Australian-born status
Calculated from 2011 Census data
Education
Experience
Top 3 professions (and fraction of workers in occupation)
Migrants
Dropout
1-10 years
Labourers
0.285
Trades
0.191
Machinery
0.139
Dropout
11-20 years
Labourers
0.276
Machinery
0.185
Trades
0.160
Dropout
21-30 years
Labourers
0.235
Machinery
0.171
Clerical
0.154
Dropout
31-40 years
Labourers
0.233
Clerical
0.178
Machinery
0.160
Y12
1-10 years
Sales
0.216
Community
0.183
Labourers
0.175
Y12
11-20 years
Clerical
0.174
Labourers
0.169
Trades
0.119
Y12
21-30 years
Clerical
0.202
Labourers
0.155
Managers
0.149
Y12
31-40 years
Clerical
0.203
Labourers
0.172
Managers
0.153
Cert w/o Y12
1-10 years
Trades
0.410
Community
0.140
Labourers
0.121
Cert w/o Y12
11-20 years
Trades
0.374
Community
0.125
Clerical
0.102
Cert w/o Y12
21-30 years
Trades
0.323
Community
0.136
Managers
0.124
Cert w/o Y12
31-40 years
Trades
0.310
Community
0.133
Managers
0.125
Cert w Y12
1-10 years
Trades
0.256
Community
0.178
Labourers
0.126
Cert w Y12
11-20 years
Trades
0.254
Professionals
0.152
Clerical
0.150
Cert w Y12
21-30 years
Trades
0.226
Professionals
0.169
Clerical
0.152
Cert w Y12
31-40 years
Trades
0.213
Professionals
0.185
Clerical
0.150
Degree
1-10 years
Professionals
0.511
Clerical
0.139
Managers
0.094
Degree
11-20 years
Professionals
0.537
Managers
0.166
Clerical
0.117
Degree
21-30 years
Professionals
0.528
Managers
0.189
Clerical
0.110
Degree
31-40 years
Professionals
0.554
Managers
0.177
Clerical
0.105
Australian born
Dropout
1-10 years
Trades
0.249
Labourers
0.229
Sales
0.155
Dropout
11-20 years
Labourers
0.220
Machinery
0.192
Clerical
0.141
Dropout
21-30 years
Clerical
0.211
Labourers
0.182
Machinery
0.163
Dropout
31-40 years
Clerical
0.239
Labourers
0.177
Machinery
0.151
Y12
1-10 years
Sales
0.255
Community
0.174
Clerical
0.162
Y12
11-20 years
Clerical
0.249
Managers
0.160
Sales
0.130
Y12
21-30 years
Clerical
0.294
Managers
0.191
Sales
0.115
Y12
31-40 years
Clerical
0.293
Managers
0.213
Professionals
0.107
Cert w/o Y12
1-10 years
Trades
0.482
Community
0.105
Clerical
0.094
Cert w/o Y12
11-20 years
Trades
0.386
Managers
0.116
Clerical
0.108
Cert w/o Y12
21-30 years
Trades
0.310
Managers
0.146
Clerical
0.132
Cert w/o Y12
31-40 years
Trades
0.282
Managers
0.143
Clerical
0.139
Cert w Y12
1-10 years
Trades
0.288
Clerical
0.175
Community
0.168
Cert w Y12
11-20 years
Trades
0.247
Clerical
0.186
Managers
0.147
Cert w Y12
21-30 years
Professionals
0.209
Clerical
0.179
Managers
0.175
Cert w Y12
31-40 years
Professionals
0.283
Managers
0.180
Clerical
0.161
Degree
1-10 years
Professionals
0.655
Managers
0.112
Clerical
0.101
Degree
11-20 years
Professionals
0.601
Managers
0.199
Clerical
0.096
Degree
21-30 years
Professionals
0.621
Managers
0.212
Clerical
0.083
Degree
31-40 years
Professionals
0.643
Managers
0.198
Clerical
0.077
24
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
Table A.7 Duncan index of dis-similarity for Australian-born and
immigrant workers calculated from 2011 Census data
holding education constant
Experience of corresponding immigrant group
Education
-experience of native group 1-10 years 11-20 years 21-30 years 31-40 years
High school dropouts
1-10 years
0.097
0.182
0.197
0.209
11-20 years
0.173
0.097
0.040
0.063
21-30 years
0.240
0.195
0.107
0.081
31-40 years
0.261
0.225
0.137
0.108
Year 12
1-10 years
0.099
0.244
0.266
0.282
11-20 years
0.271
0.148
0.104
0.121
21-30 years
0.332
0.209
0.169
0.188
31-40 years
0.354
0.222
0.183
0.197
Certificate
(w/o Year 12)
1-10 years
0.082
0.122
0.175
0.186
11-20 years
0.108
0.057
0.080
0.091
21-30 years
0.172
0.094
0.041
0.035
31-40 years
0.195
0.119
0.056
0.040
Certificate (w Year 12)
1-10 years
0.114
0.132
0.168
0.199
11-20 years
0.198
0.080
0.078
0.101
21-30 years
0.294
0.150
0.108
0.105
31-40 years
0.355
0.211
0.163
0.146
Degree
1-10 years
0.161
0.122
0.138
0.116
11-20 years
0.195
0.096
0.083
0.069
21-30 years
0.228
0.130
0.116
0.102
31-40 years
0.236
0.138
0.124
0.110
Numbers in table indicate the proportion of individuals who would have to change occupation to make the
occupational distribution identical for two groups. Highlighted cells are the lowest – indicating the most
similar distributions – in their row or column.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
25
Table A.8 Duncan index of dis-similarity for Australian-born and
immigrant workers calculated from 2011 Census data
holding education constant
Education of corresponding immigrant group
Education-experience of native
group
High
school
dropout Year 12
Certificate
(w/o
Year 12)
Certificate
(w Year 12) Degree
1-10 years
High school dropout
0.097
0.252
0.246
0.200
0.585
Year 12
0.324
0.099
0.305
0.187
0.488
Certificate (w/o Year 12)
0.328
0.399
0.082
0.227
0.550
Certificate (w Year 12)
0.353
0.280
0.220
0.114
0.427
Degree
0.711
0.640
0.668
0.622
0.161
11-20 years
High school dropout
0.097
0.155
0.346
0.332
0.568
Year 12
0.345
0.148
0.331
0.270
0.441
Certificate (w/o Year 12)
0.315
0.275
0.057
0.175
0.537
Certificate (w Year 12)
0.387
0.223
0.225
0.080
0.422
Degree
0.685
0.566
0.632
0.536
0.096
21-30 years
High school dropout
0.107
0.112
0.349
0.346
0.556
Year 12
0.324
0.169
0.325
0.275
0.421
Certificate (w/o Year 12)
0.319
0.241
0.041
0.119
0.492
Certificate (w Year 12)
0.374
0.258
0.242
0.108
0.333
Degree
0.696
0.594
0.623
0.524
0.116
31-40 years
High school dropout
0.108
0.096
0.354
0.346
0.564
Year 12
0.304
0.197
0.324
0.275
0.447
Certificate (w/o Year 12)
0.324
0.252
0.040
0.103
0.493
Certificate (w Year 12)
0.388
0.306
0.283
0.146
0.271
Degree
0.703
0.614
0.622
0.512
0.110
Numbers in table indicate the proportion of individuals who would have to change occupation to make the
occupational distribution identical for two groups. Highlighted cells are the lowest — indicating the most
similar distributions — in their row or column.
Interestingly, we find stronger effects when we consider broad skill groupings for
incumbents, but the results are inconclusive on the question as to whether immigration
leads to stronger or weaker labour market outcomes for incumbents (See table A.9.) We
find a negative association between incumbent wages and the fraction of recent
immigrants. We find statistically significant positive associations between immigration and
weekly hours worked and participation. The fraction of recent immigrants is significant at
the 5% level for participation, but only at the 10% level for wages and weekly hours. The
wage and hours effects are fairly strong. If the share of recent immigrants goes up by 1
26
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
percentage point, this is associated with a drop in wages of 2.6 per cent, an increase in
weekly hours of 32 minutes and an increase in the participation rate of one-half of one
percentage point.
When we split the sample by sex (table A.9) we again find mixed results. For males we
find a positive association between recent migration and the participation rate but also a
positive association with the unemployment rate. The wage and hours effects from the
pooled sample are concentrated amongst female workers — for men the effects are smaller
and not statistically different from zero.
Table A.9 Estimated values of
from equation (1): HILDA and Census
(incumbents compared to recent immigrants); Broad
experience groups and education categories
3 education categories and 4 experience categories
Log annual
earnings
Log weekly
earnings
Log of wage
rate
Weekly
hours
Participation
rate†
Unemployment
rate†
All incumbents
θ
0.618 0.307 -2.587* 53.607* 0.580** 0.257
(1.104) (1.082) (1.243) (29.675) (0.235) (0.153)
Males only
θ
0.430 0.371 -0.266 33.002 0.366* 0.306**
(1.944) (2.170) (1.596) (50.258) (0.186) (0.120)
Females only
θ
1.226 0.444 -5.471*** 91.568*** 0.440 0.130
(1.999) (1.815) (1.452) (24.716) (0.338) (0.340)
Models include full set of time dummies, education and experience fixed effects and full set of interactions
Note: *,**,*** indicate statistical significance at the 10%, 5%, and 1% significance level respectively.
†Calculated from Census; otherwise calculated from HILDA.
It is important to note that the negative wage effect in the pooled sample is also very
fragile and driven by one skill group: degree holders with 1-10 years of experience. If we
add a dummy variable for that group (or drop them from the analysis), the coefficient on
immigrant share in the wage regression becomes positive, 0.5376, but insignificant.
Between 2001 and 2011, this group of individuals had lower wage growth than expected
but this could plausibly be for other reasons, such as differential effects of the Global
Financial Crisis or the mining boom across education-experience groupings.
10
10
For example, either the mining boom or Global Financial Crisis could have plausibly eroded the wages of
young university graduates, relative to other workers with little experience or university graduates well
into their careers through rapid wages growth for trades or limited employment growth in traditionally
well-paid graduate jobs.
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
27
A priori, it is difficult to say whether the more narrow skill groups or the broader skill
groups provide better estimates. Comparing table A.2 to table A.9, we can see that the
standard errors are two to three times larger when we use the broader groups and the
incumbent sample. The broader groups will provide more imprecise estimates and
potentially more volatile estimates since we are estimating on a much smaller effective
sample size. The narrower groups will give biased estimates if skill groups are too
narrowly defined and if there is leakage and competition across skill groups. As others in
the literature have pointed out, the results do depend upon the definition of skill groups.
A.5 Discussion and conclusion
In this paper we use a simple and data driven approach to address whether or not the labour
market outcomes of the Australian-born and incumbents are related to patterns of
migration. We do this by constructing skill groups which are defined by education and
years of (potential) experience. We look at whether changes in the share of immigrants in
these cells over time is related to changing labour market outcomes for the Australian-born
and incumbents. We control for a variety of fixed effects as well as macroeconomic
conditions and we allow the return to skills to vary over time.
Overall, and looking across the full suite of our results, we find little evidence that the
labour market outcomes of Australian-born workers are negatively related to immigration.
The few statistically significant associations we do find are inconclusive, and cover both
stronger and weaker labour market outcomes. They may arise simply from statistical
chance or reflect the influence of omitted variables on relative labour market outcomes of
education-experience cells over time. Moreover, these associations are economically small
and only just statistically significant, so the evidence is scant. Our results are consistent
across two very different data sets.
We do find some negative effects of recent migrants (those who arrived in Australia in the
last five years) on employment and wage of incumbents (Australian-born and immigrants
who have resided in Australia for more than five years) when we consider very broadly
defined skill groups. However, we also find positive associations between recent migration
and weekly hours and labour force participation of incumbents.
The approach that we use has an advantage over approaches that use the uneven
geographical spread of immigrants to identify the impact of immigration on labour market
outcomes. In those approaches, geographical labour markets are assumed to be distinct and
movement between labour markets which might be driven by differences in employment
opportunities and wages are ruled out. In Australia, this looks like a very bad assumption
given the large flows of workers from one state to another which we observed during the
mining boom which took place during our data period, 2001–2011.
28
MIGRANT INTAKE INTO
AUSTRALIA – TECHNICAL SUPPLEMENT A
The disadvantage of our approach is that we assume that each skill group (defined by
education and experience) is an individual labour market and that there is no
substitutability of workers across different labour markets. Specifically, the approach is
assuming that the arrival of immigrants in one skill group is not causing Australian-born or
incumbent workers to move to competing in another skill group. Given that skill groups
are defined on relatively immutable categories, education and potential experience, this
seems less problematic than the geographical assumption.
Our results are dependent both upon the immigration policies in place during the period
2001–2012 and the overall economic conditions. As we are estimating over a period of
very robust economic growth, it is perhaps not surprising that we find very little negative
impact of immigration on natives and incumbents. It could be that in periods of slow
growth or contraction there are negative effects, but we would not be able to identify these
in our data. Given that our approach is non-parametric and data-driven, our results are
dependent upon policy settings. The results do not give any insight into how different
policies might affect the relationship between immigration and labour market outcomes of
the Australian-born and incumbents.
One reason why we may fail to find statistically significant results is that the amount of
variation in immigrant shares in our data is pretty small. Recalling figures A.9 and A.10,
most of the skill groups show little or no change in the proportion of immigrants over time.
A longer time window and more variability in immigration would assist in identification
— as available in the original Borjas (2003) paper — but we do not currently have either
of these things.
Our data does not account for short-term migrants. They are absent in the census data by
construction. In the SIH, they would only be counted if they were living in private
dwellings. If short-term migrants are living in hostels or other non-private dwellings, they
will not be in our data. While this group may be important for certain low-skill jobs in the
economy, the results across all skill groups should not be substantially impacted by their
absence.
Throughout, we have discussed changes in the percentage of migrants in skill groups as
being related to in-flows of migration. But, they can also be related to outflows. Immigrant
shares in skill groups can drop if Australian-born workers are out-migrating even in the
absence of any change in immigration. Our intuition, again, is that this is not an important
determinant of the results. Out-migration has been important in highly skilled groups in
Australia, but less so during the economic boom of the 2000s. For most groups,
in-migration dominates out-migration and it is this effect that we are mostly capturing.
Despite these caveats, the paper provides important new information about the relationship
between immigration and the labour market outcomes of the Australian-born and
incumbents at an aggregate level. If there were strong negative effects, the approach used
here should reveal a more consistent picture — in the signs, sizes and statistical
significance of the coefficients. The fact that we find associations with both stronger and
weaker labour market outcomes, with the few results that are statistically significant
IMMIGRATION TO AUSTRALIA AND THE LABOUR
MARKET OUTCOMES
29
relatively sensitive to assumptions such as the classification of skill groupings, suggests
that, at least at the level of the overall economy and the vast majority of workers,
immigration does not appear to have been a major factor in the labour market outcomes of
the Australian-born and previous immigrant cohorts over the period studied.
References
Bond, M., and Gaston, N. 2011, ‘The impact of Immigration on Australian-born workers:
An assessment using the National Labour Market Approach’, Economics Papers, vol.
30, no. 3, pp. 400–13
Borjas, G. J. 2003, ‘The Labor Demand Curve Is Downward Sloping: Reexamining the
Impact of Immigration on the Labor Market’, The Quarterly Journal of Economics, vol.
118, no. 4, pp. 1335–74.
Borjas, G. J. 2006, ‘Native Internal Migration and the Labor Market Impact of
Immigration’, Journal of Human Resources, vol. 41, no. 2, pp. 221–58.
Duncan, O. B. and Duncan, B. 1955, ‘Residential Distribution and Occupational
Segregation’, American Journal of Sociology, vol. 60, no. 5, pp. 493–503.
Friedberg, R. M., and Hunt, J. 1995, ‘The Impact of Immigrants on Host Country Wages,
Employment and Growth’, The Journal of Economic Perspectives, pp. 23–44.
Manacorda, M., Manning, A., and Wadsworth, J. 2012, ‘The Impact of Immigration on the
Structure of Wages: Theory and Evidence from Britain’, Journal of the European
Economic Association, vol. 10, no. 1, pp. 120–51.
Ottaviano, G. I., and Peri, G. 2012, ‘Rethinking the Effect of Immigration on Wages’,
Journal of the European Economic Association, vol. 10, no. 1, pp. 152–97.
Sinning, M. and Vorell, M. 2011, ‘People’s Attitudes and the Effects of Immigration to
Australia’, Ruhr Economic Papers 0271, Rheinisch-Westfälisches Institut für
Wirtschaftsforschung, Ruhr-Universität Bochum, Universität Dortmund, Universität
Duisburg-Essen.
Smith, J. P., and Edmonston, B. (eds.) 1997, The new Americans: Economic, demographic,
and fiscal effects of immigration, National Academies Press.
