1
Economic Commission for Europe
Conference of European Statisticians
Group of Experts on National Accounts
Twenty-second session
Geneva, 25-27 April 2023
Item 8 of the provisional agenda
Impact of migration on national accounts
Impact of Migration on National Accounts: A UK
Perspective
Prepared by the Office for National Statistics of United Kingdom1
Summary
Official statistics come in many shapes and forms. Economic statistics are generally
governed by the System of National Accounts (2008), whilst many social statistics are
governed by separate manuals and guidance. As we move into a period where migration is
increasingly important, alongside other population and demographic trends such as aging
and urbanisation, the need to ensure national accounts and other economic measures reflect
a modern understanding of these concepts is increasingly vital to ensure statistics remain
relevant and reflect society as it is.
The proposal to include labour accounts in the System of National Accounts from
2025 (UN, 2021) provides the opportunity for the economic statistics community to reflect
not just on how to best take account of migration in economic data, but also to consider
whether economic statistics has a voice which needs to be heard whilst the social statistics
community discusses the definition of migration. The question is if it is necessary to consider
a wider alignment of definitions across both economic measures of labour and social
measures of population.
1 Prepared by Richard Heys, Sonia Carrera, David Freeman, Craig McLaren, and Chris Stickney,
ONS UK. This paper is the work of the authors and not necessarily the views of the Office for
National Statistics or the UK Government.
United Nations ECE/CES/GE.20/2023/12
Economic and Social Council Distr.: General
6 April 2023
English only
ECE/CES/GE.20/2023/12
I. Migration and the National Accounts
1. Migration across the globe has become a more complicated picture since the Covid-
19 pandemic. Countries were more likely to have international travel restrictions than school
closures, restrictions on gatherings or stay at home measures. In the case of the United
Kingdom (UK), the pandemic caused enormous volatility in the resident population, which
the existing method of population projections (which rely on a stable, predicted growth rate)
and weights used in social and economic surveys derived from these estimates were not
designed to cope with. In the year to June 2020 UK net migration was 88,000, which rose to
504,00 in year to June 20222. In addition, there were an estimated 137,000 excess deaths in
the UK between March 2020 and June 20223. This has resulted in a substantial change in the
structure of the UK’s population.
2. The UK is not alone in this experience. Migrants make a huge contribution to
economies across the globe. As of 2020 there were an estimated 281 million migrants who
live outside the country they were born, comprising 3.6% of the world’s population4. The
scale of these numbers and their capacity to stimulate rapid changes in the population of a
country highlight the importance of developing more frequent and timely information about
this group and how it changes over time, which are reflected in the ongoing transformation
of population statistics in many NSIs, including the UK’s Office for National Statistics
(ONS), to make use of more timely data from a range of sources including administrative,
commercial, surveys and other sources.
3. Given this, it is imperative, at a time when the national accounts community are
considering once-in-a-generation changes to its methods, that national accountants should be
taking active account both of this phenomenon and the efforts of population statisticians and
demographers to keep pace through changes to their methods and data sources.
4. National Accounts serve two functions: they present a view of the size and
composition of the whole economy, and they present a way to perceive the change happening
at the innovative frontier. Whilst one can simplify this to understanding stocks and flows, the
nature of change in the accounts is of fundamental importance. Growth can occur either
through the process of more output being generated by the factors of production within the
production boundary, or it can occur by activity, or factors of production moving from outside
the production boundary (and therefore out of scope of the national accounts), to moving
within the production boundary and hence into scope.
5. Movements across the production boundary (and for capital assets the equivalent asset
boundary) can make understanding economic growth inherently more complex. Given that
in the UK, as with most other economies, labour inputs are the most important factor of
production5 understanding how they change and adapt is vital to understanding whether more
is being produced or simply more is being counted. As we observe with an aging population
that unpaid work undertaken in the household is now nearly equivalent in value terms, when
estimated in the UK Household Satellite Account, to GVA generated from paid employment
in the private sector6 the importance of understanding how labour effort is deployed, how it
moves across the production boundary, and how we should conceptualise and track its change
are increasingly vital. As a commentator at a UK seminar on Beyond GDP7 recently
articulated their perception of the national accounts excluding human capital: “Trying to
understand the UK economy without taking account of human capital is like trying to
understand the Himalayas without taking account of the mountains.”
2 See ONS (2022b)
3 See ONS (2022c)
4 See UN DESA, (2021)
5 ONS (2022a) estimated that the UK's inclusive net worth, including productive assets, environmental assets, human
capital, financial assets and financial liabilities was £36.2 trillion in nominal prices in 2020. Environmental assets, as
measured in the natural capital accounts, were worth £1.7 trillion, while human capital was worth £23.8 trillion or
more than double the value of the traditional non-financial capital stock captured in the national accounts
6 See Bucknall, Christie, Heys, and Taylor (2021)
7 https://www.betterstats.net/november-2022-conference/
ECE/CES/GE.20/2023/12
6. Of course, changes in the labour allocated to productive work within the national
accounts is not merely a domestic issue: movements of inputs across national boundaries are
equally as important as movements of goods and services in terms of international trade. A
failure to account for either would fundamentally bias measures of productivity and gross
value added. Migration, in so far as it characterises this movement of labour, has to be a
primary consideration in ensuring economic statistics are meaningful and accurate. As we
move into a period where migration is increasingly important, alongside other population and
demographic trends such as aging and urbanisation, the need to ensure national accounts and
other economic measures reflect a modern understanding of these concepts is increasingly
vital to ensure statistics remain relevant and reflect society as it is.
7. Nevertheless, despite the centrality of this concept, it is not the case that there is
universal agreement on how flows of labour between countries should be measured. The
Office for National Statistics (ONS), for example, compiles social and economic statistics
for the UK which have a perspective on population and migration. The current statistical
guidance landscape requires us to produce data on three different bases:
8. Nevertheless, despite the centrality of this concept, it is not the case that there is
universal agreement on how flows of labour between countries should be measured. The
Office for National Statistics (ONS), for example, compiles social and economic statistics
for the UK which have a perspective on population and migration. The current statistical
guidance landscape requires us to produce data on three different bases:
9. Population statistics defines a person as a usual resident at their permanent address
where they spend most of their time. This usually has a residency requirement, or the
intention to reside, for at least twelve months. This includes those who migrate into or out of
the UK.
10. ILO labour statistics also capture a domestic resident population, including those
who work abroad, and excluding those who reside overseas and work in the domestic
economy (e.g. those who cross the border between Northern Ireland and the Republic of
Ireland)
11. National Accounts requires labour statistics which align to the production boundary
and therefore capture those who work in the domestic economy, that is excluding those who
work abroad, and including those workers who reside overseas (such as those living in the
Republic of Ireland and working in the UK).
12. However, this landscape is changing. UN (2021) is a guidance note for the System of
National Accounts 2025 which proposes the introduction of labour accounts into the national
accounts. Whilst there is currently ambiguity on whether this new account will feature in the
core sequence of economic accounts, or simply be a thematic satellite account, the question
of how to achieve close consistency and alignment between ILO consistent and national
accounts consistent data is clearly of key importance.
13. Many National Statistics Institutes (NSIs) will not have the resources to produce two
sets of labour data on different bases and even if they are, previous experience in the UK
suggests that users find such a model complex and challenging to interpret. Indeed, the
availability of two labour series potentially showing different growth paths opens the door to
less scrupulous politicians and commentators to ‘cherry-pick’ on a monthly basis whichever
metric best fits the political narrative they wish to communicate. Runge and Hudson-Sharp
(2020) identified that: ‘…large parts of the UK public have misperceptions about how
economic figures, such as the unemployment and inflation rate, are collected and measured,
and who they are produced and published by. This sometimes-affected participants’
subsequent views of the perceived accuracy and reliability of economic statistics.’
14. Within this work they identified that labour statistics were viewed by the UK
population with a particular dubiousness, driven by historic political decisions to change the
definitions underpinning labour market measures, such as the employment rate, which led
many to view that these data were subject to political interference rather than produced by
the independent NSI. This was mildly exacerbated by the UK previously giving greater
prominence to both national accounts and ILO consistent series. Today the UK headlines its
ECE/CES/GE.20/2023/12
ILO consistent measures and publishes its national accounts consistent series within its
productivity publications, where they are generally only accessed by ‘expert users’.
15. Clearly, clear transition tables between the new national accounts labour account and
the ILO metrics will be key to attempting to circumnavigate such issues, whilst as a
community we will also need to carefully consider the public presentation of such data,
including which measures to prioritise as headline measures for the public.
16. However, simply ensuring consistency between ILO and national accounts measures
is insufficient in a world where the definition of migration within population statistics is also
under review. The current debate around the definition of migration emerges from a UN
review of conceptual frameworks and concepts and definitions on international migration in
2021. Where previously definitions were rigid about length of stay and who to include, the
upcoming amendments (UN (2021a)) to the 1998 Recommendations on Statistics of
International Migration support the need to align statistical institutes’ measurement of
resident populations.
17. The opportunity to consider whether wider alignment is possible is clearly open to the
economic statistics community at present. In particular, is there merit in considering a wider
alignment of definitions across both economic measures of labour and social measures of
population?
18. This paper will consider these questions as follows. Sections two to four will provide
headline detail on how each measure is estimated. Section five will comment on the
conceptual alignment, or lack thereof of these three measures and section six will present
some of the challenges this presents, particularly around survey design and data collection,
the necessity of being able to access data from other countries whose residents work in the
domestic economy and make proposals for how countries can co-operate to resolve these.
II. Population Statistics – their definition and scope
19. Population statistics are essential in underpinning nearly all social and economic
statistics. Accurate population counts are important in their own right for planning services
and developing policies. Equally they form the basis, or the denominator, of detailed statistics
from vaccine rollout to GDP per head and to the UN 2030 agenda for sustainable
development.
20. International guidance on population statistics is governed by ‘UN Principles and
Recommendations for Population and Housing Censuses’, 2017. They make clear the need
to measure a “usually resident population” which requires a threshold of 12 months when
considering place of usual residence according to one of the following two criteria:
(i) The place at which the person has lived continuously for most of the last 12 months
(that is, for at least six months and one day), not including temporary absences for
holidays or work assignments, or intends to live for at least six months;
(ii) The place at which the person has lived continuously for at least the last 12 months,
not including temporary absences for holidays or work assignments, or intends to
live for at least 12 months.
21. Similarly, the UN Recommendations on Statistics of International Migration, 1998,
records an international migrant as someone who changes their country of usual residence8
similarly suggesting a period of 12 months.
22. There is a reasonable argument for a ‘usual’ resident population. Having an accurate
distribution of the population supports long-term planning, particularly in small geographic
areas to avoid short term volatility. This is also important for projecting future population
estimates and is crucial for producing high quality sampling frames for surveys which draw
on population or households, such as the Labour Force Survey.
8 See paragraph 32 in particular
ECE/CES/GE.20/2023/12
23. On the other hand, the rise of global mobility has changed the way countries need to
provide services. A usual resident population doesn’t consider those who live in a country
for shorter periods but still needs access to schools, hospitals and other public services, and
might be engaged in meaningful employment. The Final Report on Conceptual frameworks
and Concepts and Definitions on International Migration, April 2021, calls on a need for a
“present population” comprised of both the resident population and a temporary population
component. The temporary population can make a significant contribution to both the
economy and society and attributing this contribution to the usually resident population (often
the denominator) misrepresents the reality.
24. A final question is whether data can meet the challenge of alternative population
bases. Historical estimates of long-term international migration in the UK using the
International Passenger Survey measured a person’s intention at the point of arrival rather
than actual observed migration. ONS (2019) presents evidence people can change their
intentions after entering or leaving the port facilities (both air and sea) where the survey takes
place. Administrative data allows for the measurement of actual observed migrations. There
are limitations with this too, however. For example, electoral roll data can provide insights
on those populations who have a right to vote in certain elections (in the UK this eligibility
can vary across local and national elections) given the caveat that registration is voluntary
and so only those who apply to vote are captured, resulting in obvious biases (see, for
example, de Coulon et al (2020)). Similarly accounting for a shorter period of migration using
methods of “interactions” with administrative data risks wider reasons other than migration
being the cause for why such interactions may stop, presenting an obvious potential bias.
25. However, the risks are greater with the continuation of an intentions-based survey
which would fail to properly address the changes to migration behaviours following the
recent shocks of the pandemic and exiting the EU. Using administrative data will allow for
quick changes to policy that may affect migration. Given travellers intentions may
subsequently change, their behaviour in the administrative data will generally capture this.
III. International Labour Organisation employment statistics –
their definition and scope
26. The number of people in employment in the UK is estimated in line with international
standards and definitions laid down by the International Labour Organisation (ILO), This
measure consists of people aged 16 years and over who did one hour or more of paid work
per week plus those who had a job that they were temporarily away from (for example,
because they were on holiday or off sick).
27. The measure of employment, and related measures of unemployment and inactivity,
alongside estimates of jobs are measured in line with many other developed economies,
through a sample survey of households labelled the Labour Force Survey (LFS). The LFS
adheres to international standards and can be readily compared with equivalent surveys in
other countries. The LFS is dependent on UK population statistics through the population
weights which are applied to its sample which are derived from the decennial census of
population undertaken in every year concluding with a 1. This dataset has been collected in
the UK continuously save the Second World War since 1841. This measure consists of people
aged 16 years and over who did one hour or more of paid work per week and those who had
a job that they were temporarily away from (for example, because they were on holiday or
off sick).
28. The International Labour Organisation (ILO) defines migrant workers (ILO (2018))
as someone in employment who has changed their country of usual residence. The
information can be measured in two ways by looking at the recorded nationality or country
of birth.
29. The main source for measuring migrant workers is through labour force surveys of
households. However, there are some measurement issues to be considered when the data is
confronted with national accounts data:
ECE/CES/GE.20/2023/12
• Labour force surveys often exclude communal establishments, and so will miss
some migrant workers. Communal settings, such as barracks and hospitals can
obviously also capture domestic workers, but there is evidence that migrant workers
can be disproportionately represented in certain types of households in rented and
multiple occupation. As a result of the coronavirus (COVID-19) pandemic, the
contact method for the LFS had to change from face-to-face interviewing to
telephone-based. This change had an impact on the non-response bias of the survey,
particularly for housing tenure with lower response rate for rented accommodations.
This impacted on the estimates of non-UK born residents, who are more likely to
live in rented accommodations. ONS (2020) explains how this was mitigated
through adding a housing tenure variable into the LFS weighting methodology.
• Looking just at people in usual residence will miss those employed who have been
in the country for fewer than six months. This can particularly affect those in
seasonal employment. In the UK, there exists colloquial and media evidence that
this can be disproportionately biased towards certain industrial sectors, such as
agriculture, and sometimes particular sectors in particular geographies9.
• Household surveys can pick up people who are resident in one country but work in
another. (This could be due to commuting across a border or working remotely.)
Clearly this can cause ILO labour metrics to come out of alignment with the national
accounts production boundary which focuses on domestic production and the inputs
which feed into this.
• People can change their nationality, both whilst abroad and whilst resident in the
domestic economy.
• Finally, there can be significant differences between an individual self-reporting
their sector of employment and the industry which they would align to within the
national accounts. A simple example is a van driver who works for a chain of shops.
Whilst they may self-identify as being in the logistics industry their employer would
be categorised within the retail sector. This challenge is addressed in the next
section.
IV. National Accounts consistent labour statistics – their definition
and scope
30. Chapter 19, ‘Population and Labour Inputs’ of the present System of National
Accounts (UN (2008))10, provides the definitions of relevance to this paper. It defines the
population in general terms11 as ‘all those persons who are usually resident in the country….
that is persons are resident in the country where they have the strongest links thereby
establishing a centre of predominant economic interest. Generally, the criterion would be
based on their country of residence for one year or more.’ This clearly is a definition with a
direct antecedent in those used in population statistics, Para 19.11 caveats this by noting two
particular groups– ‘usual residents who are living abroad for less than one year are included
in the population but foreign visitors (for example, holidaymakers) who are in the country
for less than one year are excluded from the measured population’. Clearly the first group are
an erroneous inclusion from the perspective of alignment to the production boundary if their
employment is not captured within the domestic economy, whereas the second group is
correctly excluded from a production perspective but are clearly of importance in terms of
tourism and trade statistics.
31. The SNA also outlines how the national accounts uses labour estimates:
19.5 Clearly, if a ratio is to be formed between measures of output and labour input,
the concept of labour used must match the coverage of production in the SNA. The
relevant standards … confirms that the economically active population is defined in
9 See, for example: https://www.bbc.co.uk/news/uk-politics-eu-referendum-36258541
10 The Balance of Payments Manual aligns on the following definitions
11 Para 19.10
ECE/CES/GE.20/2023/12
terms of individuals willing to supply labour to undertake an activity included in
the SNA production boundary12.
19.6 Not everyone who is economically active works for a resident institutional unit.
It is therefore particularly important that the concept of residence underlying the
population estimates be consistent with that for labour force estimates and that the
residence of individuals included in employment estimates are consistent with the
criterion of resident institutional unit in the SNA.
32. How to tackle the question of those who are in active employment with less than
twelve months residency is addressed as follows:
19.18 Because the labour force is defined with reference to a short period13, the
number of persons in the labour force at any time may be smaller than the
economically active population. For example, seasonal workers may be included in
the economically active population but not in the labour force at certain times of
year.
33. As such, National Accounts looks to capture all labour involved in production,
irrespective of length of tenure in the domestic economy, whilst also trying to align on the
fundamental definition of the population for GDP per head type metrics.
34. The exact practice which countries use to calculate the national accounts consistent
data can vary, as mapped by Ward et al (2018):
‘In most countries, labour force surveys (LFS) form a primary source of information
for employment related statistics, such as persons employed, employees and hours
worked. However, because the coverage of LFS does not fully align with the coverage
of activities used to estimate GDP, additional adjustments relying on complementary
sources, such as administrative or business statistics, are often applied to bridge
conceptual differences, and in many countries, the use of these sources is often
preferred to LFS data. Evidence from the 2018 OECD/Eurostat national accounts
labour input survey shows that the adjustments made to align measures of labour input
with the corresponding measures of production according to the domestic concept,
vary considerably across countries, with many countries making no adjustments, in
particular, for the measurement of hours worked.’
35. In essence countries can use three alternative methods to source the required data:
• Household surveys, such as the Labour Force Surveys
• Other surveys of employment, generally business surveys. In the UK these include:
The Annual Survey of Hours and Earnings (ASHE), carried out in April each
year, is the most comprehensive source of information on the structure and
distribution of earnings in the UK. ASHE provides information about the
levels, distribution and make-up of earnings and paid hours worked for
employees in all industries and occupations14.
The Short-Term Employment Surveys (STES). STES is a group of surveys that
collect employment and turnover information from private sector businesses.
In December of each year, the jobs estimates are "benchmarked" to the latest
estimates from the Business Register and Employment Survey (BRES).
the Business Register and Employment Survey (BRES) captures employee and
employment estimates at detailed geographical and industrial levels and is
regarded as the official source of employee and employment estimates by
detailed geography and industry15.
• Administrative sources
12 Bold text as contained in UN (2008). It should be noted that paras 19.17, 19.19 and 19.20 all replicate some version
of this definition from slightly different perspectives
13 Noted as ‘usually a week’ in para 19.17
14 For more detail see
https://www.ons.gov.uk/surveys/informationforbusinesses/businesssurveys/annualsurveyofhoursandearningsashe
15 For more detail see
https://www.ons.gov.uk/surveys/informationforbusinesses/businesssurveys/businessregisterandemploymentsurvey
ECE/CES/GE.20/2023/12
Real-Time Pay-As-You-Earn (PAYE RTI) Tax administrative data provides a
count of workers on employer payrolls that feed information into the tax
system. This covers the vast majority of employees and can provide counts and
earnings information by geographical area and can be matched with data that
provide nationality information.
Other tax and benefits administrative data, for example from Self-Assessment
tax returns, can provide a count of self-employed workers that feed information
into the tax system, similarly to PAYE RTI.
36. As Ward et al note specific adjustments are generally required to be made to a) achieve
conceptual alignment with the production boundary and b) address known biases in the data,
in a fashion tailored to the limitations of the particular data source used.
A. Conceptual Adjustments
37. Broadly four types of conceptual adjustment are required to align with national
accounts requirements if one begins with household survey data, such as the LFS:
• Territoriality – the issues of cross-border work have to be adjusted for in a
household survey, whereas a business survey, which only covers domestic
businesses does not face the same biases.
• Seasonal work adjustments where the population weights which are used to derive
whole economy estimates are generated from a population estimate which fails to
take account of short-term economic migration, which we have already observed is
of more significance in particular industries. As with the other issues described in
this section, the effect here is to bias the geographical or industrial distribution of
potentially both estimates of output and productivity.
• Alignment with industry groupings. As Ward et al describe the challenge –
‘industry coding is often conducted on the basis of information given by the
respondent about the type of product, service or function provided by his/her place
of work, which may not align with the industry coding of that firm in the business
register, and hence national accounts (although in some countries this alignment is
improved by matching respondents information, such as the name and address of
the firm with equivalent information on the business register)’. In the UK, LFS data
is reweighted using the STES to address this distributional issue, which would
otherwise serve to bias measures of productivity derived from the national accounts.
• Coverage – Ward et al (2018) note that ‘the LFS does not cover some groups of the
population such as persons below or above certain age thresholds (which varies by
country), those living and working in communal establishments (such as prisons or
long-term care facilities), collective households (such as religious institutions) and
the armed forces, all of whose output is included, at least in theory, in estimates of
GDP.’
38. When administrative data or business surveys are used, alternative adjustments are
required. This is specifically to convert the number of hours worked from usual or contracted
hours to actual hours worked. This can be sufficient to bias productivity estimates directly
or, if FTE numbers of staff in particular occupation classifications are used to derived
estimates of output for particular products or assets (see for example own-account software
and database assets), through biasing both the numerator and the denominator in productivity
calculations.
B. Bias Adjustments
39. Of more general concern are the issues relating to LFS biases, specifically those which
affect measures of actual hours worked. As Ward et al (2018) identify these can be
significant, either due to cultural issues of deliberate mis-reporting (certain professions /
societies over-report hours as an issue of personal pride), error (where self-declared actual
hours cannot be replicated from self-declared ‘usual hours’ plus over-times minus absences
of all types), or methods issues (rolling forward hours worked for survey respondents who
are absent for a month who may be on leave can lead to over-estimation of key variables.
These biases can all be expected to be relatively consistent through time.
40. Other biases which may be inconsistent through time result from periods of relatively
high (or low) immigration or emigration of labour, specifically during the period up to twelve
ECE/CES/GE.20/2023/12
months of residency point where such individuals start to be included in population measures
and hence LFS weights. In addition, survey weights, which often use projections, can be
subject to time lags: changes in migration patterns can take a number of years to be reflected
in survey weights. The UK method for adjusting weights in the Labour Force Survey, for
example, requires updated annual population estimates, followed by biennial population
projections, and finally new survey weights.
41. Administrative or business surveys of employment are less likely to suffer from such
issues as it can be relatively safely assumed the survey respondent will be doing so from a
staff list or wage report and are unlikely to differentiate in their report between those who
have met the twelve-month residency threshold.
V. Conceptual alignment
42. As can be seen above, the two key issues relating to migration, taken at its widest to
mean any movement of labour across borders, and those captured within the measurement of
the economy are: a) around those in paid employment with fewer than twelve months tenure,
and b) those who work in the domestic economy and live overseas or vice versa, as
demonstrated in the following table:
Table1: Conceptual alignment of the three metrics
Those living overseas
and working in the
domestic economy16
Those working in the domestic
economy, but fewer than
twelve months residency
Those working in the
domestic economy, and
more than twelve months
residency
Population
Statistics
Excluded Excluded Included
ILO Labour
Measures
Excluded Included conceptually, but
excluded in the weights as these
come from population statistics
Included
National
Accounts
Included Included conceptually, but
excluded in the weights as these
come from population statistics
Included
43. Being able to have the clarity on where these differences lie allow us to then consider
where these may lead to our metrics behaving differently as patterns of migration change.
This aspect of change is vital to consider, as with many economic statistics this can wash
through to different challenges in terms of levels and growth rates:
• In normal times, if rates of migration or cross-border working remain relatively
consistent across time periods then the growth rates should not be significantly
biased, although the stock level may be more subject to bias.
• However, in circumstances when migration trends change this can result in three
problems:
• Faster growth in net migration, where these individuals are allowed to work, will
result in national accounts capturing faster GVA growth in the first year whilst not
observing labour inputs growth. This results in accelerated growth in GDP per head
and labour productivity measures.
• Wider discrepancies in dis-aggregations, both of industry and geography if
migration relatively greater affects some industries and regions than others, which
can distort the appearance of where the drivers of growth can be observed.
• If migration also has a greater or lesser impact than average on particular occupation
classifications which are used for cost of production type estimates, such as for
intangible capital investment (e.g. software and databases) this could distort
16 Those living domestically and working overseas are obviously included in Population Statistics and ILO measures,
but excluded from National Accounts
ECE/CES/GE.20/2023/12
perceptions of investment rates and again the drivers of growth in the national
accounts.
• The impact will also be dependent on the relative size of the sector of the economy.
VI. Simple example of potential impact
44. It is possible to illustrate the impact of these challenges using the agricultural industry,
which is one of those who are relatively intense users of migrant labour for short-term, often
seasonal roles, particularly during the harvest season in late summer / early autumn. If one
was to incorporate these seasonal workers in population estimates the following impacts
could be observed:
• The population weights for this industry would be increased in the Labour Force
Survey.
• Seasonal statistical adjustments applied to the Labour Force Survey would be
applied. This would have the effect of spreading this workforce throughout the year.
Overall number of workers and hours worked would however be affected
• National Accounts data on gross value added would remain unchanged, but where
this is already seasonally adjusted the relationship over the year between
employment should improve in accuracy. Growth rates should be relatively
unaffected.
• Labour productivity measures which conflict gross value-added and measures of
hours worked or workers would fall recognising the increase in the total number of
employees and hours worked, more accurately capturing the true position. The
present model where the same output is shared amongst the resident workforce
naturally exaggerates their productivity levels, although again, if the number of
seasonal workers remains constant over the years, annual growth rates will be
relatively unaffected.
VII. Conclusions
45. Whatever the issues affecting economic measurement, it is important to be aware that
this is not a conversation which occurs in isolation of wider challenges. Waiting twelve
months to confirm usual residency, in addition to collecting and processing data, creates an
inevitable time lag, which reduces the timeliness of population and migration statistics. This
is an issue of increased salience for many statistical systems and political debate, and clearly
has the ability to impact economic statistics.
46. The issues around the conceptual alignment between the three treatments described
above indicate clearly that economic measures can be differentially affected both when
migration patterns change, but could also shift if one or more of the three definitions of
population and migration statistics methods are revised. This leads us to twin questions: 1)
whether these conceptual differences aide or hinder the production of statistics and their
improvement, and 2) if it does aide, are new data sources or methods available to permit us to
consider viable alternatives?
47. In a fast moving, modern, digital economy, where data is becoming more readily
available in ever-increasing quantities, the prospect of moving to a system where issues of
population and migration do not need to wait for the twelve-month threshold, and can be
measured within a shorter period obviously present themselves as increasingly feasible
options. Where these may better align to user need, whether this concerns students, seasonal
workers, cross-border workers or other groups, or may reduce mis-alignment between
measures this should clearly be explored in greater depth to ensure we reach an end-state
which is optimal for all users of these data. Noting any revised definition of migration would
need to meet the needs of a variety of population and migration statistics and above all be
coherent between stocks and flows, it would also be the case that economic measurement
would need to consider the potential for any such change to impact key measures of GVA,
investment, GDP per head and productivity measures, as well as human capital and education
satellite accounts.
ECE/CES/GE.20/2023/12
48. As national statistics institutes continue to improve on the timeliness and quality of
their population statistics, many looking to replace a decennial census, now is the opportunity
to ensure the entire statistical system benefits from this. Users may no longer have the appetite
to tolerate a large time lag in estimating a usual resident population that doesn’t reflect the
entire population who still contribute to the economy and society.
49. To move this debate forward, it is clearly vital to understand the current debate within
population statistics. The agenda at the 54th UNSC (see UN(2023a)) will consider the role of
temporary mobility and its importance to economy and society. Having greater recognition
and clear conceptual framework of this group enables wider possibilities of how international
statistical institutes integrate these into future economic and social statistics. Much like the
System of National Accounts, a new System of Population and Social Accounts/Statistics
could give greater clarity to definitions and their use. In addition, the recent creation of a
Friends of Chair Social and Demographic Statistics (see, UN(2023b)) can help accelerate
progress on horizontal integration across social, economic and environmental statistics - and
closer alignment.
50. However, even once we understand this debate, the next step has to be for the
economic measurement community to consider how far do we want to align economic
measures, how we wish to use population statistics in the future and whether wider alignment
would allow users to transition across these datasets in a way which is more intelligible to the
general public and relates better to their lived experience.
51. The current debate around including Labour Accounts into the updated SNA 2025 is
the perfect opportunity to consider these points and to reflect on whether as a community we
have sufficiently tackled this vital question. However, the draft annotated chapter outline
circulated in December 2022 on Labour Accounts, did not mention the word ‘migration’,
despite for many users and members of the public migration is a fundamental economic
question, in terms of the impact on labour markets, but also their wider net contribution to
society, if only through taxes and use of public services. Ensuring we tackle this question
means we need to consider three fundamental questions:
• Does our reliance on population statistics on a different basis affect the validity of
our data?
• Does our current data reflect real experience if we exclude migrants with fewer than
twelve month’s residence?
• How should we consider inflows and outflows of human capital / education output,
if we wish to consider the issue from a stocks perspective?
ECE/CES/GE.20/2023/12
Bibliography
Bucknall, R., Christie, S., Heys, R., and Taylor, C. (2021): ‘GDP and Welfare: Empirical
Estimates of a Spectrum of Opportunity’ ESCoE Discussion Paper 2021-08. Available at
https://www.escoe.ac.uk/publications/gdp-and-welfare-empirical-estimates-of-a-spectrum-
of-opportunity/
Cathro, C., Runge, J., Whitwell-Mak, J., Stockland, K., Broughton, N., and Rostron, J.
(2022) ‘Improving Public Understanding of Economic Statistics: Presenting Labour Market
Statistics to the Public’ ESCoE Discussion Paper 2022-26. Available at https://escoe-
website.s3.amazonaws.com/wp-content/uploads/2022/11/10105516/ESCoE-DP-2022-
26.pdf
de Coulon, A., Egyei, R.K., & Wadsworth, J. (2020) ‘Immigration Stocks and Flows, APS
and Electoral Register Data’ ESCoE Discussion Paper 2020-13. Available at
https://www.escoe.ac.uk/publications/immigration-stocks-and-flows-aps-and-electoral-
register-data/
ILO (2018) ‘ILO Guidelines concerning statistics on International Labour Migration.’
Available at https://www.ilo.org/wcmsp5/groups/public/---dgreports/---
stat/documents/meetingdocument/wcms_648922.pdf
ONS(2019) ‘Understanding different migration data sources – August 2019 Progress Report
’. Available at
https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/internatio
nalmigration/articles/understandingdifferentmigrationdatasources/augustprogressreport
ONS (2020) ‘Coronavirus and its impact on the Labour Force Survey’. Available at
https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/employmentandemplo
yeetypes/articles/coronavirusanditsimpactonthelabourforcesurvey/2020-10-13
ONS (2022a) ‘Inclusive capital stock, UK: 2019 and 2020’. Available at
https://www.ons.gov.uk/economy/economicoutputandproductivity/output/articles/inclusivec
apitalstockuk/2019and2020
ONS (2022b) ‘Long-term international migration, provisional: year ending June 2022’.
Available at
https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/internatio
nalmigration/bulletins/longterminternationalmigrationprovisional/yearendingjune2022
ONS (2022c) ‘Excess deaths in England and Wales: March 2020 to June 2022’. Available at
https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/a
rticles/excessdeathsinenglandandwalesmarch2020tojune2022/2022-09-20
Runge, J., and Hudson-Sharp, N. (2020) ‘Public Understanding of Economics and
Economic Statistics’ (ESCoE Occasional Paper 03). Available at
https://www.escoe.ac.uk/publications/public-understanding-of-economics-and-economic-
statistics/
United Nations Department of Economic and Social Affairs (UN DESA), (2021)
‘International Migrant Stock 2020’. New York. Available at
www.un.org/development/desa/pd/content/international-migrant-stock
UN (1998) ‘1998 Recommendations on Statistics of International Migration Revision 1’.
Available at https://unstats.un.org/unsd/publication/seriesm/seriesm_58rev1e.pdf
UN (2008): ‘System of National Accounts’. Available at
https://unstats.un.org/unsd/nationalaccount/docs/sna2008.pdf
UN (2017) ‘UN Principles and Recommendations for Population and Housing Censuses’.
Available at https://unstats.un.org/unsd/publication/seriesM/Series_M67rev3en.pdf
UN (2021): ‘WS.4 Labour, Human Capital and Education’. Available at
https://unstats.un.org/unsd/nationalaccount/RAconsultation.asp?cID=12
ECE/CES/GE.20/2023/12
UN(2021a): ‘The Final Report on Conceptual frameworks and Concepts and Definitions on
International Migration, April 2021’ Available at https://unstats.un.org/unsd/demographic-
social/migration-expert-group/task-forces/TF2-ConceptualFramework-Final.pdf
UN(2023a): ‘Report of the UN Expert Group on Migration Statistics on Indicators for
international migration and temporary mobility’. Available at
https://unstats.un.org/UNSDWebsite/statcom/session_54/documents/BG-3b-EGMS-E.pdf
UN(2023b): 'Terms of Reference of the Friends of the Chair Group on Social and
Demographic Statistics’. Available at
https://unstats.un.org/UNSDWebsite/statcom/session_54/documents/BG-3b-
ToR_FoC_Social-E.pdf
Ward, A., Zinni, M.B., Marianna, P., (2018) ‘International productivity gaps: Are labour
input measures comparable?’ Available at:
https://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=SDD/DOC(2018)
12&docLanguage=En
- Group of Experts on National Accounts
- Twenty-second session
- Impact of Migration on National Accounts: A UK Perspective
- Prepared by the Office for National Statistics of United Kingdom0F
- I. Migration and the National Accounts
- II. Population Statistics – their definition and scope
- III. International Labour Organisation employment statistics – their definition and scope
- IV. National Accounts consistent labour statistics – their definition and scope
- V. Conceptual alignment
- VI. Simple example of potential impact
- VII. Conclusions
- Bibliography