The UK labour market and immigration.
Wadsworth, Jonathan
Rising immigration has undoubtedly been one of the most significant
demographic developments experienced by the United Kingdom over the past
fifteen years. This article reviews the evidence on the effects of
immigration on the UK labour market. On average, it seems that
immigration has not had much of an effect on either employment or pay.
However, there may be some evidence of downward pressures on pay and
jobs impact in the low skill sector, though these effects are not large.
Keywords: Immigration; employment; wages JEL Classification: J0
Introduction
Rising immigration has undoubtedly been one of the most significant
demographic developments experienced by the United Kingdom over the past
fifteen years. From the beginning of 1995 to the end of 2008, the stock
of working age immigrants rose from 2.9 to 5.4 million. Over the same
period the UK-born working age population grew by around 500,000, the
result of birth, death and emigration rates that were broadly in
balance. Since then, the immigrant population has fallen back slightly,
in part the result of the latest recession. Recessions reduce the
opportunities for work and immigration typically falls in a downturn.
The latest recession was no exception. Still, by the end of 2009, around
14.2 per cent of the working age population had been born abroad, up
from around 8 per cent at the end of the 1990-93 recession (figure 1).
The UK is not unique in having experienced increases in
immigration. Immigration has risen in almost all industrialised
countries over the past decade. The average share of migrants in the
OECD increased from 10.7 per cent of the population to 13.8 per cent
between 1998 and 2007 (OECD, 2009). The immigrant population shares in
countries like Canada, the Netherlands, Spain and Sweden are all
currently higher than in the UK (OECD, 2009). Nor is a significant rise
in the UK working age population unprecedented. Between 1977 and 1995
the UK-born population of working age rose by some 2.3 million, as the
baby boom generation reached maturity, and the immigrant stock rose by
around 500,000--almost the exact opposite pattern to that experienced
over the past fifteen years. Nevertheless rising immigration does seem
to be correlated with increased concerns over the job and wage prospects
of UK-born workers. While it is relatively trivial to achieve higher
levels of GDP by expanding the labour force, it is less obvious that GDP
per head need increase as a result of rising labour supply, whether
caused by immigration or by other factors. Hence the issue of whether
immigration puts downward pressure on the wages of incumbents or reduces
job prospects becomes central to the debate on labour market efficacy.
In what follows we assess the empirical evidence in an attempt to
establish whether these concerns are justified.
Drivers of immigration
Rising immigration means rising labour supply, unless offset by
increased emigration. Whether this puts downward pressure on wages
depends on what is happening to labour demand. Immigration will rise if
relative opportunities in the UK are better than in the source countries
or in other potential destination countries. In turn relative
opportunities reflect both the economic performance of the country
alongside any institutional barriers to entry, or exit. Indeed one of
the reasons why the UK was able to attract so many migrants over the
previous decade was undoubtedly helped by the fact the economy was
growing much faster than many other major economies over this period.
Just as with re-integrating the long-term unemployed or the inactive, or
increasing the retirement age, migration is one way of dealing with
potential obstacles to--noninflationary--growth caused by domestic
labour supply shortages.
[FIGURE 1 OMITTED]
While relative economic prosperity will attract migrants,
institutional arrangements can and do influence supply. UK government
policy in the 2000s facilitated migrant inflows, following the decision,
along with Sweden and Ireland, to allow immediate free movement of
labour to accession (A8) citizens in 2004, rather than after the
seven-year transition period applied by other existing EU member states.
Migration from outside the EU had, for some time, been restricted to
skilled workers through the work permit system. The introduction of the
points based system in 2008 essentially continued the institutional
arrangements seeking to control supply from outside the EU and influence
its skill content.
If labour demand exceeds labour supply in the receiving country,
the impact of immigration will be different from that in a country
already at full employment. Concerns about substitution and displacement
of the UK-born workforce become more prevalent when output is demand
constrained, as in a recession. In the context, the distribution of
skills and demand within a country also matters. Any effects of
immigration are likely to be most profound among those groups that are
closest substitutes for immigrant labour. If the skill composition of
immigrants differs from that of the receiving country there will be
differential pressures across skill groups. If not, then the economy
effectively grows by replicating itself. Similarly, the output mix, the
level of technological intensity and prices are other additional
adjustment mechanisms that need not mean that wages are the only source
of adjustment to an increase in supply.
Trends and stylised facts
Rising immigration to the UK has been associated with a change in
the country of origin mix. The immigrant stock disaggregated by country
of origin is becoming much more heterogeneous in Britain (table 1).
Twenty years ago, one third of all immigrants came from just two
countries, Ireland and India. Now these two countries account for just
12 per cent of all immigrants. The top three sender countries for the
new arrivals to the UK in 2009 were Poland, India and the United States,
while the share of immigrants from EU Europe has remained broadly
constant over this period and the share of immigrants from Africa and
Asia has risen. The mix of European Union migrants has, however, changed
considerably over time, particularly after the UK, along with Ireland
and Sweden, allowed immediate entry to the inhabitants of the A8
accession countries in 2004. Poland is now the second largest source of
immigrants to the UK.
Immigrants are, on average, more educated than their UK-born
counterparts, and the educational attainment gap has been rising over
time, since more recent immigrants are more educated, on average, than
other immigrants (see table 2). While more than half of the UK-born
workforce left school at 16 or earlier, fewer than one sixth of new
immigrants finished their education by the age of 16. Just under one in
five UK-born members of the workforce finished education at 21 or after
compared with more than one in three immigrants and more than 50 per
cent of all new immigrants.
While the stock of immigrants has risen in all regions over time,
it has risen most in London. Although there is some anecdotal evidence
to suggest that new immigrants are more regionally dispersed than in the
past, the coefficient of variation for the immigrant share across
eighteen UK regions suggests that immigration has become more
concentrated in certain areas, over time, falling from 0.86 in 1985 to
0.75 in 2009 (according to the LFS). Thirty-two per cent of immigrants
still live in London and, in 2009, 39 per cent of London's
population were immigrants (see figure 2). The geographic dispersion of
immigrant share across more disaggregated local areas is much larger.
Some 60 per cent of the working age populations of Brent and Westminster
were born overseas compared to under 3 per cent of the populations of
Knowsley or Redcar & Cleveland.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
New immigrants also tend to live in areas with higher existing
immigrant shares. This reflects a combination of the nature of the local
labour market demand, a migrant's perception of the best location
to maximise opportunities (Borjas, 2001) and the role of networks
facilitating the arrival of new immigrants (see for example McKenzie and
Rapport, 2010). Figure 3 compares the share of new immigrants (those who
have been living in the UK for less than one year) in each of 201 UK
local authorities identified in the Annual Population Survey (APS) with
the share of other immigrants in the same local authority. The line
through the data summarises the positive correlation that is apparent.
Figure 4 compares the numbers of immigrants in each age group with
the numbers of UK-born in the same age group. The average age of an
immigrant is, in 2009, 37.5 (and 28 on arrival) compared to the UK-born
average age of 40. Since most immigrants to the UK arrive as young
adults, there are few children among the stock of immigrants. With
regard to the adult population there is clearly most supply pressure
among adults in their late twenties, though whether immigrants are
substitutes within or across age groups is an issue we return to below.
[FIGURE 4 OMITTED]
Estimating the labour market impact of immigration
One way to try to assess the effect of immigration is to compare
average wages or unemployment in the years before and after a change in
labour supply. However, because other mechanisms like the economic cycle
or shifting sectoral demands will also have affected wages and
employment over the same period, this will not answer the question of
whether immigration has an effect on the labour market prospects of
resident workers.
Ideally we want to compare the observed change in the outcome of
interest with the 'counterfactual' change that would have
taken place had immigration not occurred but everything else (e.g.
technological change) had and to relate this difference to the size of
the change in the labour force caused by immigration. The effect
immigration has on the wages (or employment) of native-born workers is
then
([W.sup.observed.sub.after] - [W.sup.observed.sub.before]) -
([W.sup.counterfactual.sub.after] - [W.sup.counterfactual.sub.before]) /
[DELTA]immigration
However such counterfactuals are not observed and so there is a
need to circumvent this problem. Typically the labour market is divided
into sub-groups in order to compare wages in groups that experienced a
lot of immigration with wages for groups that did not (so that this
latter group effectively becomes the counterfactual). The search for the
appropriate choice of counterfactual and hence identification of an
immigration effect--has led to the development of a variety of
techniques. The 'spatial correlations' approach identifies
immigration effects by comparing changes in wages across local labour
markets with different immigrant shares. Dustman, Fabbri and Preston
(2005) use regional-level data disaggregated by skill over time for the
UK and conclude that migration had no discernible effect on the level of
native wages.
However some object to this approach because of worries over the
endogeneity of both immigrant and native-born location choice.
Immigrants may be attracted to the best performing local labour markets
(so wages determine immigrant location rather than immigrants determine
wages). Similarly concerns over the efficacy of this strategy have been
raised because native-born workers may move in response to an inflow of
immigrants. This may in turn affect local labour supplies in both source
and destination areas of these secondary internal migrants and hence any
impact of immigration is not confined to the local area of arrival.
Figure 5 plots the percentage change in the outflow rate of UK-born
individuals from each of 349 local areas in the UK between 2004 and 2008
against the percentage change in the immigrant share in each local
authority over the same period. The figure suggests that the correlation
between UK-born mobility and immigrant inflows is very weak and, if
anything, negative, so that this particular issue may not be of too much
concern in the UK.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
However the fact is that, as figure 6 shows, immigrant inflows do
appear to be correlated with local area performance, as measured by the
hourly wage. Certainly the concentration of immigrants in London
suggests that there is a correlation between location and local economic
opportunities and this does tend to reduce the usefulness of these
spatial correlation exercises to identify an immigration effect.
One way to deal with the endogeneity of location choice is to look
at the consequence of exogenous residential location of immigrants
caused, for example, by randomised allocations of refugees across areas.
However such interventions are rare (see Glitz, 2006). Instrumentation
of immigration using pattern of previous settlements (correlated with
immigration location decisions and hopefully uncorrelated with current
local area performance) is a possibility, though concerns remain over
this strategy since previous residential choice may have been correlated
with economic performance in the past and local area economic
performance appears to be quite persistent over time.
Consequently recent attention has shifted towards a more aggregate
economy approach. This involves dividing the working population into
age/skill/year cells rather than areas since it is hard to choose age
and gender--and, more contentiously, skill--and then look to see if
differences in the shares of immigrants across cells are associated with
differences in wages. This avoids any local labour market endogeneity
concerns but has, in the past, assumed that immigrants are perfect
substitutes for natives within a cell (so that everyone does much the
same job within a given age/skill cell). Borjas (2003) adopts this
approach for the USA and finds negative effects of immigration.
However, just as skilled and unskilled labour may not be able to do
the same tasks immediately, there are reasons to think that because of
institutions--for example barriers to immigrants with overseas
qualifications practising in certain occupations, barriers caused by
language assimilation or the time needed to adapt existing skills in a
new country--immigrants may not always be able to do the same job to the
same degree immediately as a native-born individual. The degree of
substitutability therefore becomes a matter for empirical verification.
Standard economic theory suggests that we can estimate this by looking
at the wage of UK-born workers relative to that of immigrants
conditional on the supply of UK-born workers (N) relative to immigrants
(M) in the same age (a) and education (e) group.
ln ([W.sup.N.sub.ea] / [W.sup.M.sub.ea]) = [beta] -
1/[sigma].sub.l] ln([N.sub.ea] / [M.sub.ea]) (1)
The essential idea is that any increase in the labour supply of any
group will, other things equal, put downward pressure on the wages of
that group. Rising immigration constitutes an increase in labour supply
and so the issue is whether migrants compete with other migrants for
jobs or with UK-born workers, or both. To answer this question we need
an idea of the degree of substitutability between groups and equation
(1) is a straightforward way of estimating this relationship. (1) The
coefficient 1/[[sigma].sub.l] is the inverse of the elasticity of
substitution. A coefficient of zero implies that the elasticity of
substitution is infinite and so the immigrants and native-born workers
of the same age and with the same level of education are perfect
substitutes. This means that rising labour supply caused by immigration
will put more downward pressure on wages and/or employment of UK-born
workers. A non-zero coefficient suggests that the two groups are
imperfect substitutes. The larger is 1/[[sigma].sub.l] (in absolute
terms) the more inelastic the degree of substitution, the less easy it
is to substitute migrant labour for domestic labour and the greater is
the concentration of the wage effects caused by rising immigrant supply
on existing immigrants (who are assumed to be closer substitutes for new
immigrants) rather than on UK-born workers. This means that, if true,
then any pressures on wages and employment of the UK-born will be
reduced if any newcomers are imperfect substitutes.
Work by Manacorda, Manning and Wadsworth (2006) suggests that
immigrants and UK-born workers may well be imperfect substitutes in the
UK, but that the degree of substitution may be greater among the less
skilled. Ottoviani and Peri (2006) arrive at a similar conclusion for
the United States. The idea that less skilled labour is easier to
substitute than skilled labour seems intuitive and is consistent with
recent work by Dustmann, Frattini and Preston (2008) and Nickell and
Salaheen (2008) who also find that competition for jobs from migrants
may be greater among the less skilled. Some evidence supportive of this
is given in table 3, which outlines the percentage shares of immigrants
and UK-born workers in different occupations conditional on the level of
education.
The majority of immigrants are concentrated in certain unskilled
jobs, despite being more highly qualified, on average, than UK-born
workers. This is particularly noticeable for immigrants educated to
degree level. Whether this is because of employer preferences or
perceptions of immigrant quality, immigrant preferences for temporary
work, or language barriers has yet to be established, but together these
factors may help explain why there could be more downward pressure in
less skilled jobs. The national minimum wage does of course provide an
important level below which wages cannot fall.
[FIGURE 7 OMITTED]
There is, however, some anecdotal evidence to suggest that the
degree of substitutability may have increased over time, since there is
now very little difference between the employment rates of immigrant and
UK-born men. Historically the employment gap has widened in recessions
and narrowed in economic recoveries. This has not been observed during
the latest recession. Unemployment rates for immigrants and UK-born have
risen together by similar amounts (see figure 7). This may be, in part,
because immigrants, as shown in table 2, are more skilled, on average,
than in the past, making them less vulnerable to any downturn. Lemos and
Portes (2008) find no association between the increase in A8 migration
in local areas between 2004 and 2006 and local unemployment rates,
either in aggregate or among youth or the less skilled. More research on
the effects of immigration in a downturn is, however, probably still
needed.
Conclusion
On balance, the evidence for the UK labour market suggests that
fears over the consequences of rising immigration have not been borne
out. It is hard to find evidence of much displacement of encumbent
workers or lower wages, on average. That is not to say there have been
no effects. As ever, the less skilled may have experienced greater
downward pressure on wages and greater competition for jobs than others,
but these effects still appear to have been relatively modest. Since
these effects are averaged across good and bad times, we know less about
the effects in downturns and there is a need to look at these effects in
more detail. We also need to understand more about how capital and
sectoral shifts in demand respond to immigration over the medium and
longer term. Future migration trends will, as ever, depend on relative
economic performance and opportunity but we still need to know more
about the effects of rising immigration beyond the labour market in
areas like prices, consumption, housing, health, crime and welfare.
doi: 10.1177/0027950110380324
REFERENCES
Borjas, G. (2001), 'Does immigration grease the wheels of the
labor market?', Brookings Papers on Economic Activity, No. 1, pp.
69-119.
--(2003), 'The labour demand curve is downward sloping:
reexamining the impact of immigration on the labour market',
Quarterly Journal of Economics, 118, pp. 1335-74.
Dustmann, C., Fabbri, F. and Preston, I. (2005), 'The impact
of immigration on the UK labour market', Economic Journal, 115, pp.
F324-41.
Dustmann, C., Frattini, T. and Preston, I. (2008), 'The effect
of immigration along the distribution of wages', CREAM Discussion
Paper, No. 03/08.
Glitz, A. (2006), 'The labour market impact of immigration:
quasi-experimental evidence', CREAM Discussion Paper No. 12/06.
Lemos, S. and Portes, J. (2008), 'New labour? The impact of
migration from central and eastern European countries on the UK labour
market', IZA Discussion Paper No. 3756.
Manacorda, M., Manning, A. and Wadsworth, J. (2006), 'The
impact of immigration on the structure of wages in Britain', London
School of Economics, Centre for Economic Performance Discussion paper
No. 754.
McKenzie, D. and Rapoport, H. (2010), 'Self-selection patterns
in U.S.-Mexico migration: the role of migration networks', Review
of Economics and Statistics (forthcoming).
Nickell, S. and Salaheen, J., (2008), 'The impact of
immigration on occupational wages: British evidence', working
paper, Nuffield College, Oxford.
--(2010), 'Immigration in the UK', in Gregg, P. and
Wadsworth, J. (eds), The State of Working Britain, Oxford, Oxford
University Press.
Ottaviano, G. and Peri, G. (2006), 'Rethinking the effects of
immigration on wages', NBER Working Paper No. 12497.
OECD (2009), International Migration Outlook, Paris, OECD.
NOTE
(1) For more discussion on the appropriate functional form see
Manacorda, Manning and Wadsworth (2006).
Jonathan Wadsworth, Royal Holloway College, University of London,
CEP, CREAM and IZA. E-mail: J.Wadsworth@rhul.ac.uk.
Table 1. Immigrant shares and area of origin, 1979-2009
1979 1989 1999 2009
Immigrant
share 7.6 8.3 9.7 12.7
of which % from:
Europe 41.6 35.9 32.2 31.1
Asia 28.0 32.9 32.4 35.1
Africa 11.8 15.0 19.7 21.2
Americas 16.3 13.2 11.8 9.4
Oceania 2.3 3.0 3.9 3.3
Largest senders (% of all immigrants)
1 Ireland Ireland India India
(19.30 (15.8) (9.6) (9.8)
2 India India Ireland Poland
(13.3) (12.3) (8.9) (8.6)
3 Jamaica Pakistan Pakistan Pakistan
(6.2) (7.3) (6.8) (7.6)
4 Pakistan Germany Germany Germany
(5.8) (4.4) (5.3) (5.1)
5 Italy Jamaica Bangladesh South Africa
(4.3) (4.0) (4.4) (3.5)
Source: Labour Force Survey. Percentage share of all immigrants in
brackets.
Table 2. Education and immigrant status (working age
population), 2009, per cent
Percentage of group with each level
of education
Age finished UK-born All New
education immigrants immigrants
<=16 53.1 24.8 15.4
17-20 28.2 36.5 34.0
21+ 18.7 38.7 50.6
Source: Labour Force Survey (excluding students).
Table 3. Occupation, immigrant status and education
Manager/ Assoc. Admin Skilled
profess- Prof. manual
ional
Secondary
UK-born 20.9 13.2 13.1 13.6
A8 3.4 2.7 4.4 17.7
Other imm. 20.0 13.5 10.3 10.5
University
UK-born 60.6 21.1 7.2 2.9
A8 13.2 6.9 6.4 10.7
Other imm. 49.8 19.6 7.6 3.3
% in each occupation
Personal Sales Element-
services ary
Secondary
UK-born 10.1 7.6 21.5
A8 5.8 3.1 63.0
Other imm. 10.4 7.7 27.6
University
UK-born 3.0 2.5 2.8
A8 8.8 4.0 49.9
Other imm. 5.8 4.2 9.6
Source: Labour Force Survey.
