Memorandum by the National Institute of
Economic and Social Research[1]
THE NUMBERS
AND CHARACTERISTICS
OF RECENT
IMMIGRANTS (QUESTIONS
1, 2, 13)
Data on both the numbers and characteristics
of recent immigrants are an important input to understanding the
economic impacts of immigration, which we discuss below. In this
context we discuss some of the statistics available. At present
Total International Migration (TIM) statistics produced by the
Office for National Statistics are compiled from the International
Passenger Survey (IPS). According to these, immigration net of
emigration rose sharply from less than 50,000 per annum on average
during most of the 1990s to more than 150,000 per annum on average
over the years 1998-2003. Following the enlargement of the EU
in spring 2004, net immigration rose by a further 50,000 per annum
on average in 2004 and 2005. In these years, net immigration from
the Accession 8 (A8)[2]
averaged more than 50,000 per annum.
The Workers Registration Scheme (WRS) provides
more detailed and frequently quoted statistics on migrants from
the A8. The WRS records gross, rather than net, immigration flows
and show that in the three years following Accession there have
been 656.4 thousand A8 citizens registering to work in the UK,
more than half of which registered before 2006.[3]
These numbers exceed TIM estimates of gross migration from the
A8 during 2004-05 by approximately 100,000 per annum, indicating
the temporary nature of migration from the New Member States (NMS).
The WRS counts individuals who intend to work in the UK for at
least a month. The IPS figures, on which TIM estimates are based,
count individuals who intend to stay in the UK for at least a
year. We note that comparability between these figures is complicated
by the fact that the WRS is likely to undercount the total number
of migrants, since it only counts individuals who register to
work as employees. Also, the reliability of the IPS data is often
questioned due to the way in which this data is collected through
voluntary survey participation at points of entry to and exit
from the UK, and because it is based on intentions rather than
realised outcomes.
In terms of gauging the economic impact of recent
immigration it is useful to know how many migrants are resident
in the UK at a particular time, and hence contributing to UK economic
performance, although the flows are relevant for these purposes
too. Neither the WRS nor the IPS provides this sort of information.
Of course `stock' numbers can be derived from these under certain
assumptions. These are, however, not the only sources of information
about migrants. The Labour Force Survey (LFS), which has a sample
of about 60,000 respondents asks whether respondents were born
in the United Kingdom or, if not, the year of their arrival. Although
the LFS it is not without its problems, it does provide a head
count of migrants resident in the UK.
Quite separate from the benefit of having available
data on migrant stocks as well as flows, given the concerns about
the reliability of the IPS, it is important to compare the data
which emerge from the IPS with those produced by the LFS. We note
that one means of enhancing the information available would be
to devise a means of producing combined estimates of the numbers
of recent immigrants using both sources. Such figures will result
in revised intra-censal population estimates and may shed additional
light on the accuracy of the IPS. An advantage of the LFS is that,
since it looks at people resident in the country, it automatically
covers the people who arrive not intending to stay (and who are
therefore not counted as migrants) but then change their minds;
it equally excludes the people who do intend to stay but leave
after less than a year (and should therefore not have been counted
as migrants). The IPS figures are adjusted to allow for these
but the benefit of an independent data source should not be overlooked.
Using the LFS data we can illustrate the number
of A8 migrants resident in the UK at different points in time
since the enlargement of the EU in 2004 (Table A).[4]
The figures do not provide information on most recent migrants
since the sample frame covers only people who have been at addresses
for at least six months. These data suggest there were 382,000
people, or 0.6% of the population, born in the A8 who had arrived
in the UK since Accession to the EU. Reflecting the age distribution
of A8 migrants, these individuals comprise a larger proportion,
0.9%, of the working age population. The steady increase in the
recent A8 migrant stock is partly a result of the duration of
residence criteria for participation in the LFS.
Updating the data summary in an earlier study
by the National Institute of Economic and Social Research (Riley
and Weale, 2006) we get a picture of the number of recent migrants
resident in the UK and their labour market characteristics. We
limit ourselves to the data for the most recent available quarter
(2007 quarter 1) to obtain an up-to-date picture, despite the
relatively small sample size which results. These data (Table
B) suggest that 4.6% of the population, and 6% of the working
age population, arrived in the UK in the last decade. In comparison
to the situation only one year ago NMS migrants account for an
increasing share of the recent migrant population. At the beginning
of 2006, migrants from the A8 accounted for a quarter of post-Accession
migrants (Riley and Weale, 2006). At the beginning of 2007, this
number has risen to a third.
We also note two distinguishing features of
migrants who arrived in the UK during the years 2004-06, the years
following Accession, in comparison to those who arrived between
1998 and 2003, the previous period over which we saw an upward
shift in net immigration to the UK. The more recent migrant group
is significantly younger, much more likely to be employed in low
skilled occupations and less likely to be employed in professional
and managerial occupations (Table B). These differences are entirely
attributable to differences in characteristics between recent
NMS migrants and migrants from other areas. Recent NMS migrants
of working age are more likely to be employed than both the indigenous
population and migrants from elsewhere.
THE MACROECONOMIC
IMPACTS OF
RECENT IMMIGRATION
(QUESTIONS 4, 6, 7, 8, 12)
It is clear that the scale of recent immigration
implied by the numbers discussed above should have been associated
with an increase in GDP. Previous analysis undertaken by the National
Institute of Economic and Social Research (Riley and Weale, 2006),
using a simple accounting framework, suggests that immigration
during 1998-2005 contributed to a rise in real GDP of around 3%
Looking at the years 2004-05 alone, immigration contributed 1%
of GDP. This does not take into account the further increases
in GDP that would accrue if capital accumulated to match the rise
in labour input. The National Institute of Economic and Social
Research has further examined the potential macroeconomic impacts
of recent immigration with the help of its macroeconomic models
(Kirby and Riley, 2006; Riley and Weale, 2006; Barrell, Guillemineau
and Liadze, 2006; Barrell, FitzGerald and Riley, 2007). These
studies provide a useful illustration of the types and magnitudes
of effects from immigration that we might observe at the national
level. Much of the debate around the economic impacts of recent
immigration relies on analyses undertaken at the individual or
sub-national level (eg Gilpin et al, 2006; Blanchflower
et al, 2007). These and similar studies provide valuable
evidence on the economic impacts of immigration, but the results
obtained do not necessarily extend to the national level, and
the results discussed here can be viewed as complementary to these.
Here we discuss some of the issues raised in
previous work by the National Institute of Economic and Social
research in the context of a simulation exercise using the Institute's
global econometric model NiGEM.[5]
This example serves to illustrate the macroeconomic impacts of
recent A8 immigration to the UK. We model this immigration as
an unanticipated shift in the population and the population of
working age commensurate to the figures in Table A. This updates
previous Institute work by taking into account further increases
in the A8 migrant population that have occurred since the summer
of 2006. The potential impacts on GDP, inflation, unemployment,
productivity and GDP per capita are illustrated in Table C. Here
we discuss key assumptions underlying these results and how alternative
assumptions would affect these.
The increase in the labour force from the immigration
raises potential output, and in the longer term output rises to
match this increase. In the short-term, productivity as measured
by output per hour in the UK is likely to have grown more slowly
than it otherwise would have, since the capital stock (including
public sector infrastructure and the housing stock) takes longer
to adjust to the increase in labour supply and in the short-term
the level of capital employed per hour worked falls. If it is
the case that migrants concentrate in sectors where the capital-to-labour
ratio is low then productivity may return to base more quickly
than this simulation suggests. Overall, GDP increases as the increase
in labour supply results in increased employment. By 2015 the
level of GDP is approximately 0.9% above baseline. Average output
per capita is reduced in the short term, due to both the short
term increase in the unemployment rate and the slow adjustment
of the capital stock, but is higher in the longer term than it
would have been in the absence of this increase in migration.
The longer term effects reflect the age composition of migrants,
which is skewed towards working age. Although not illustrated
here this also results in small improvements in the public finances
(Kirby and Riley, 2006). In this example we have assumed that
A8 migrants and natives are equally productive. Taking into account
the evidence that recent A8 migrants tend to be concentrated in
low-skilled occupations the GDP effects would be smaller than
shown here, as we illustrate in Barrell, FitzGerald and Riley
(2007).
With the labour supply curve shifting down the
labour demand curve, there is a short-run increase in unemployment
until the economy adjusts and labour demand is able to meet the
extra capacity. The increase in unemployment can occur, either
through migrant labour searching for employment or the displacement
of existing workers. Indeed there is some evidence from the LFS
that suggests that the employment rate of people of working age
from the A8 countries was initially much lower than is now the
case (64% in 2005Q1 in comparison to 81% in 2007Q1). Our simulations
show the ILO unemployment rate returning to its baseline rate
by around 2012 and, at its peak, the unemployment rate is only
0.4 percentage points higher than baseline. We have assumed that
migrants and natives are perfect substitutes in production. We
have not assumed that the inflows of migrants are composed of
workers who reduce mismatch in the labour market by meeting currently
unfulfilled labour demand, relieving bottlenecks. If there were
such effects occurring in the economy then the impact on unemployment
in the short run would be less than the numbers suggested here,
while the immediate impact on output would be greater due to this
larger increase in the labour input. Conversely, if immigration
is concentrated amongst the low-skilled and low-paid workers,
where wages are likely to be less downwardly flexible, we may
expect to see smaller disinflationary effects and larger upward
pressure on unemployment, as illustrated in (Riley and Weale,
2006) using the model in Riley and Young (2007). In these simulations
we make no allowance for additional remittance payments from migrant
workers to their home country. These are likely to be large for
A8 migrants (Fihel et al, 2006), partly because of the
temporary nature of migration from these countries. Taking these
into account we might expect to see larger reductions in inflationary
pressure than indicated in Table A, as the rise in demand would
be less strong.
This example highlights some of the differences
between the macroeconomic impacts of unanticipated migration in
the longer and the shorter term. In the shorter term the economy
is affected by short run adjustment factors and the extent to
which these matter is largely a consequence of the assumption
that the change in migration is unanticipated, which seems a reasonable
assumption to make in the case of recent A8 migration. This is
also one of the key differences in determining the economic effects
of migration versus general population change. In many ways we
can think of the change in migration in the same way that we think
about any other population change. However, there are likely to
be important differences. In addition to whether or not these
changes can be regarded as expected or unexpected, we have highlighted
the importance of the age distribution of the change in the population,
the role of remittances in macroeconomic adjustment, existing
productivity differentials, and the degree of substitutability
at the economy level between different types of labour, including
migrants and natives.
The example here is useful in illustrating the
potential macroeconomic impacts of a sudden change in the population,
similar to that which has been observed following the enlargement
of the EU in 2004. A separate question is the effectiveness of
macroeconomic policy in a small open economy, such as the UK,
where labour is increasingly mobile across national borders. In
this context migration may both affect the parameters of wage
bargaining and make economies more flexible, changing the short
term trade-off between output and price effects in response to
policy innovations.
FUTURE TRENDS
FOR IMMIGRATION
(QUESTIONS 1, 13, 14)
An understanding of the behavioural factors
determining migration is necessary if the reliability of long-term
forecasts of population growth and output growth are to be assessed
and judgements made about the policy instruments available to
affect the evolution of migrant flows. Such an understanding is
also essential for the production of meaningful forecasts of the
migration inflow. The National Institute of Economic and Social
Research (Mitchell and Pain, 2003) has developed an econometric
model of the economic and demographic determinants of annual migrant
inflows into the UK. We found that the change in migration over
the decade to 1998-2000 is primarily due to population growth
in the source locations and the continuing pull effects from the
rise in the migrant stock and per capita incomes in the UK relative
to the source location. This finding is in contrast to the conclusions
of Hatton (2005), which suggest that the upward shift in UK immigration
towards the end of the 1990s is in large part attributable to
changes in immigration policy. In this regard we note that rising
immigration is not a feature peculiar to the UK. Several OECD
countries have experienced rising immigration in the last decade.
It is probably fair to say that the sharp increase
in immigration following the enlargement of the EU in Spring 2004
was attributable to a sudden and one-off change in policy—ie
the relaxation of border restrictions for people from the NMS—the
impact of which will have been amplified by the temporary restrictions
on labour mobility put in place in most other EU countries (the
exceptions being Ireland, which has probably seen the largest
change in the population due to immigration from the NMS if this
is measured relative to the local population, and Sweden). Because
this policy change can be regarded as a one-off event, and because
most EU countries have now relaxed their initial restrictions
on migration from the NMS, allowing migrants to move freely across
their borders, it seems likely that the change in the level of
immigration in 2004 will have been a temporary one. However, there
are a number of factors that may help to sustain a relatively
high level of NMS immigration to the UK, in historical comparison.
Income differentials between the old EU countries and the NMS,
particularly Poland, Latvia, Lithuania, and Slovakia, are likely
to persist in the near future. Specific to the UK (and Ireland),
the establishment of new NMS communities in the wake of Accession
will act as a "pull" factor for further NMS migrants
in the years to come. Our empirical work suggests that this "friends
and family" effect is one of the most significant determinants
of migration into the UK (Mitchell and Pain, 2003). Finally, the
strength of the UK labour market vis-a"-vis many other old
EU countries, if this persists, is likely to be a factor contributing
to the popularity of the UK as a destination country for NMS migrants
and migrants from outside the EU.
While it is possible to point to factors influencing
future migration trends to the UK, it is clear from our modelling
work that forecasts of immigration are associated with substantial
uncertainty. With this in mind it is not a question of one forecast
proving to be right and another forecast proving to be wrong.
Point forecasts are best seen as the central points of ranges
of uncertainty. It is therefore important to provide a description
of the uncertainty associated with migration forecasts. Just as
the public have learnt to understand that the Bank of England's
"fan" chart for inflation does not mean the Bank does
not control inflation, similarly we should expect the public to
understand "uncertainty bands" published around migration
forecasts. Awareness of the uncertainty around the central estimate
will enable better decisions to be made. The degree of uncertainty
around central migration estimates should be estimated from a
model of migration inflows.
Table A
NUMBER OF A8 MIGRANTS IN THE UK ARRIVING
SINCE MAY 2004
|
| All Thousands
| % of population |
Working age Thousands |
% of population |
|
| 2004q4 | 54.5
| 0.09 | 46.0
| 0.13 |
| 2005q1 | 57.5
| 0.10 | 50.5
| 0.14 |
| 2005q2 | 89.5
| 0.15 | 80.8
| 0.22 |
| 2005q3 | 122.1
| 0.21 | 112.5
| 0.31 |
| 2005q4 | 164.5
| 0.28 | 149.5
| 0.41 |
| 2006q1 | 199.5
| 0.34 | 175.1
| 0.48 |
| 2006q2 | 211.9
| 0.36 | 189.6
| 0.52 |
| 2006q3 | 263.0
| 0.45 | 241.7
| 0.66 |
| 2006q4 | 336.3
| 0.57 | 298.6
| 0.81 |
| 2007q1 | 382.3
| 0.63 | 338.6
| 0.90 |
|
| Source: Quarterly Labour Force Surveys 2004Q1-2007Q1; numbers weighted with population weights.
|
Table B
THE STRUCTURE OF RECENT IMMIGRATION
|
| Population
structure
| Of which came to the UK
|
| First quarter
2007
| 1998-2003
All
| 2004-06
All |
2004-06
NMS |
|
| Numbers (thousands) | 58,988
| 1,454 | 1,234
| 382 |
| Age 18-24 | 5,297
| 161 | 279
| 115 |
| (% of total) | (9.0)
| (11.1) | (22.6)
| (30.1) |
| Age 25-34 | 7,480
| 560 | 469
| 161 |
| (% of total) | (12.7)
| (38.5) | (38.0)
| (42.1) |
| Age 35-49 | 13,166
| 376 | 192
| 43 |
| (% of total) | (22.3)
| (25.9) | (15.5)
| (11.4) |
| Working age | 36,679
| 1,188 | 1,004
| 339 |
| (% of total) | (62.2)
| (81.7) | (81.4)
| (88.6) |
| Working age outside labour force and in full-time education
| 1,863 | 100
| 117 | -
|
| (% of working age population) | (5.1)
| (8.4) | (11.6)
| |
| Working age employed | 27,124
| 783 | 648
| 276 |
| (% of working age population) | (74.0)
| (66.0) | (64.6)
| (81.4) |
| In professional, managerial, and associate professional occupations
| 11,620 | 355
| 209 | 31
|
| (% of employed working age population) |
(42.8) | (45.4)
| (32.3) | (11.2)
|
| In intermediate occupations | 10,467
| 253 | 182
| 72 |
| (% of employed working age population) |
(38.6) | (32.4)
| (28.0) | (26.2)
|
| Process, plant and machine operatives and elementary occupations
| 5,037 | 175
| 257 | 173
|
| (% of employed working age population) |
(18.6) | (22.3)
| (39.7) | (62.6)
|
|
Source: Quarterly Labour Force Survey 2007Q1; Numbers weighted with population weights.
Notes: (-) LFS sample size too small to report.
|
Table C
AN ILLUSTRATION OF THE POTENTIAL ECONOMIC IMPACTS OF RECENT
A8 MIGRATION
(% difference from base)*
|
| GDP
| Inflation rate |
Unemployment
rate
| Productivity
(GDP per person
hour)
| GDP per capita |
|
| 2005 | 0.15 |
0.13 | 0.18
| 0.03 | -0.03
|
| 2006 | 0.29 |
0.00 | 0.36
| -0.05 | -0.14
|
| 2007 | 0.42 |
-0.15 | 0.43
| -0.16 | -0.21
|
| 2008 | 0.50 |
-0.25 | 0.28
| -0.23 | -0.12
|
| 2009 | 0.59 |
-0.24 | 0.18
| -0.25 | -0.03
|
| 2010 | 0.67 |
-0.19 | 0.10
| -0.25 | 0.05
|
| 2011 | 0.73 |
-0.15 | 0.04
| -0.23 | 0.11
|
| 2012 | 0.78 |
-0.10 | 0.00
| -0.22 | 0.17
|
| 2013 | 0.82 |
-0.07 | -0.03
| -0.20 | 0.21
|
| 2014 | 0.86 |
-0.04 | -0.05
| -0.18 | 0.24
|
| 2015 | 0.88 |
-0.02 | -0.06
| -0.16 | 0.27
|
|
*All numbers shown as per cent difference from baseline except
for the inflation rate and the unemployment rate which are shown
as percentage point difference from base.
30 September 2007
1
The authors are Research Fellows at the National Institute of
Economic and Social Research (NIESR). The views expressed here
reflect those of the authors and do not necessarily reflect those
of the NIESR. Back
2
The eight Central and Eastern European countries that joined the
EU on 1 May 2004 gaining new rights to live and work in the UK.
These are the Czech Republic, Estonia, Hungary, Latvia, Lithuania,
Poland, Slovakia and Slovenia. Back
3
Accession Monitoring Report: May 2004-June 2007, Border &
Immigration Agency, Home Office (2007). Back
4
The population weights used in deriving these have not been adjusted
to the latest mid-year population estimates. Estimates of the
total population also differ from the mid-year estimates as the
sample for the LFS does not include communal residences. Back
5
For a description of the National Institute Global Econometric
Model (NiGEM) see Al-Eyd et al. (2006). Back
|
