Research and analysis

Refugee Integration Outcomes (RIO) Insights: Embarks, Economic Activity, and Housing between 2015 and 2021

Published 26 September 2025

Applies to England and Wales

Analysis from linked Census 2021 and administrative data for asylum and resettled refugees in England and Wales between 2015 and 2021.

Executive summary

The Refugee Integration Outcomes (RIO) longitudinal cohort study is a joint project between the Home Office (HO), the Office for National Statistics (ONS) and the Ministry of Housing, Communities and Local Government (MHCLG). Finding robust quantitative data on refugee integration outcomes, particularly over the longer term, is challenging. Existing studies on refugee integration outcomes are scarce and often exhibit weaknesses in the robustness of their data. The aim of the RIO cohort study is to create a comprehensive dataset of refugee outcomes to better inform our understanding of integration outcomes and provide a more robust evidence base for interventions by government and other providers.

The current RIO dataset includes refugees resettled under the Vulnerable Persons Resettlement Scheme (VPRS) or Vulnerable Children’s Resettlement Scheme (VCRS) (resettled refugees) and asylum seekers who were granted refugee status between 2015 and 2020 (asylum route refugees). These groups represent a subset of all refugees in England and Wales and therefore conclusions from this study should not be generalised across other refugee groups.

This report presents analysis results from refugee cohorts linked to Census 2021 data. There is a significant difference in the quality of data matching for our 2 refugee groups. ONS previously published the Refugee integration outcomes data-linkage pilot: Census 2021 linkage methodology article reporting that Census 2021 linkage rates were 91% for the resettled refugee (RR) group and 70% for the asylum route refugee (ARR) group. The linkage rate for the ARR group is 72% in this latest study. The difference in linkage rates between the 2 cohorts emphasises that these are 2 distinct cohorts with different characteristics and behaviours. Analysis of each group should be interpreted and used with caution, but particularly for the ARR considering the significantly lower level of data matching for this group.

These are not official statistics of the refugee population, and the numbers quoted here may differ from official statistics. The analysis and results may change as we continue to update linkage methods. The linkage rates for the ARR group were lower and therefore analysis of this group should be used and interpreted with more caution. These groups represent a subset of refugees in England and Wales and conclusions from this study should not be used to make inferences about the wider refugee population. Official statistics on asylum and resettled refugees are available in regularly published HO Asylum and resettlement datasets.

Refugee departure rates

Linkage to Home Office Borders and Immigration (HOBI) data has revealed new insights into how many refugees stay in the UK long term, by allowing us to flag those who had not left the country at the time of the HOBI data extract (31 December 2022). Our analysis suggests that most refugees remain living in the UK. We found that 6% of ARR and 2% of the RR cohort were flagged as having left the country and appeared absent based on their last recorded departure with no return by 31 December 2022. Of those whose data suggested they had left the UK, approximately 60% did so after 3 years or more in the UK.

Labour market

Economic activity

Employment is a fundamental aspect of refugee integration, as highlighted in a wide range of academic research and the Home Office Indicators of Integration framework. Employment provides people with an income to sustain themselves and become less reliant on the state, helps them to build social networks, and improves their language acquisition and confidence.

Analysis of the Census 2021 data confirms what we have seen in other data, that refugees were overall less likely to be economically active than the usual resident population, but that the percentage of those who were economically active and in employment increased with more time spent living in England and Wales. We also found that people who were granted refugee status following an asylum application (the ARR group) were more likely to be economically active and in employment than resettlement refugees (RR), and that generally, men were more likely to be economically active and in employment than women.

Over time, we saw greater labour market participation and employment outcomes for refugees. We saw the highest rate of employment and labour market participation for the earliest cohort (2015) of ARR men, who, in 2021, were approaching similar levels to the usual resident population. The lowest rates of economic activity and employment were, overall, seen for RR women. Our data suggests that caring for home and/or family was a commonly cited reason for RR women not being economically active.

Working patterns

Of those who were in employment, we found the ARR group were more likely to be in full-time employment (31 to 49 hours per week) than RR, who were more likely to be working part-time. Refugee women, overall, were more likely to be in part-time employment (working zero to 30 hours per week). This pattern did not change with increasing time in the UK. In the usual resident population, women who are working are also more likely to be in part-time employment than men.

For refugee men, we saw very different patterns of full-time and part-time working for the 2 cohorts. For ARR men, we saw an equal split of full-time and part-time, whereas RR men were predominantly working part-time.

Occupation types

Refugees in this study tended to work in occupations that require few qualifications or little prior experience. The largest proportions of refugees reported working in elementary occupations, which are defined as those that comprise simple and routine tasks that mainly require the use of hand-held tools and often some physical effort. Other occupations frequently reported include skilled trades, sales and customer service.

Language proficiency

Overall language proficiency 

Proficiency in speaking English is an important facilitator of a range of integration outcomes, including employment and social connections. Analysis of Census 2021 data showed that ARR reported higher language proficiency than RR, with higher percentages reporting English as their main language. Additionally, we saw that language proficiency was higher among refugees who had been in the country for longer, suggesting an improvement over time.

Language proficiency differences by sex and age category

Overall, there was not a large difference in language proficiency between men and women in each cohort. We observed that language proficiency was strongest in the 16 to 24 age category, with over 50% of both women and men in this age group self-reporting speaking English ‘well’ or ‘very well’. There was a steady trend of decreasing English proficiency with increasing age. A disparity in English proficiency between men and women became apparent in the older age groups, where men were more proficient in speaking English.

Reasons for the above observations may include refugees in younger age categories having had more English experience before moving to England and Wales, given the increased prevalence of English-speaking media across the wider world in recent years (Sundqvist & Sylvén, 2016), or being more likely to be in employment, which provides more opportunities for development of their language skills, both in formal and informal situations (Dustmann & Fabbri, 2003).

Language proficiency and economic activity

For both cohorts, a higher percentage of refugees who reported better English skills were economically active and in employment compared to those reporting lower levels of English. However, even when accounting for the reported differences in English language skills, a larger percentage of ARR was in employment than RR.

The RR cohort includes a greater proportion of women,^{[footnote 1]} dependent children (see Table AE in our published data tables) and people with potentially greater vulnerabilities, whose exposure to English may be more limited, compared with the ARR cohort. This will likely affect not only language proficiency but the chances of entering stable employment.

Housing and living conditions

Accommodation type

Like economic integration, housing is not solely an outcome, but an important facilitator of further integration. From analysis of Census 2021 data, we found that nearly all refugees in our dataset were living in households (99%) rather than communal establishments (1%), such as hotels, guest houses, hostels or other temporary accommodation. We saw differences in the types of accommodation between ARR and RR cohorts. Most ARR households were living in a flat, maisonette or apartment; whereas for RR households, we observed a more even split of accommodation types which matched more closely the proportions seen in the usual resident population.

Overcrowding

Most ARR and RR households were occupied-to-standard (neither overcrowded nor under-occupied). The percentage of refugee households living in overcrowded accommodation was similar for both ARR and RR households, and was higher than the overall proportion of overcrowded households reported by Census 2021 for the usual resident population.

For both ARR and RR households, we observed higher rates of overcrowding for earlier arriving cohorts. The reasons for an increased likelihood of overcrowding with length of time in the country are not known, but it may be due to factors such as further family reunification^{[footnote 2]} or having more children.^{[footnote 3]}

Tenure

Overall, most refugee households live in rented accommodation, with most ARR households living in private rented accommodation and most RR households living in social rented accommodation. We saw a higher percentage of refugee households living in social rented accommodation the longer they had been in England and Wales.

Compared to the usual resident population of England and Wales, home ownership was very low amongst the refugees in the cohorts we analysed here. Looking at differences between cohorts, we observed a slightly higher percentage of owner-occupied households for ARR compared with RR. There was only a modest increase in home ownership with length of time in England and Wales for ARR, but no trend was observed for RR by year of resettlement. However, analysis for RR was limited by the small sample size and substantial data suppression. Therefore, the findings should be interpreted with greater caution. The previous ONS report 2011 Census Analysis: Social and Economic Characteristics by Length of Residence of Migrant Populations in England and Wales has shown that, in general, home ownership tends to increase over time amongst migrants, with levels of home ownership rising closer to or above the average for other households.

Future plans

The RIO cohort study currently covers England and Wales, but we hope to expand this study to cover Scotland and Northern Ireland, subject to data quality and availability, and agreement from data suppliers. We also intend to include refugees arriving via other humanitarian and protection routes in the future, such as the UK Resettlement Scheme, Afghan Citizens Resettlement Scheme, Afghan Relocation and Assistance Programme. We also hope to conduct further linkage to income and benefit data, additional housing data, health data and education data.

1. Background and introduction

The integration of refugees is a complex issue. The nature of integration differs between countries and is a statistically complex phenomenon to measure. The United Nations High Commissioner for Refugees (UNHCR) defines integration from a refugee perspective as “a preparedness to adapt to the lifestyle of the host society without having to lose one’s own cultural identity”. Equally, the host society needs to be “welcoming and responsive to refugees and for public institutions to meet the needs of a diverse population”.

The Government’s Integrated Communities Strategy defines the result of successful integration as “communities where people, whatever their background, live, work, learn and socialise together, based on shared rights, responsibilities and opportunities”. The Home Office Indicators of Integration framework identifies the practical processes and changes that contribute to integrating individuals and communities, including refugees.

1.1 The Refugee Integration Outcomes (RIO) project

The RIO longitudinal cohort study is a joint data linking project between the Home Office (HO), the Office for National Statistics (ONS) and the Ministry for Housing, Communities and Local Government (MHCLG). The project aims to link HO refugee data with other administrative data across government departments to create an anonymised analytical dataset.

RIO was conceived as a pilot project in 2019. The pilot study tested the feasibility of linking refugee data to administrative data sources, developed specific linkage algorithms based around Arabic naming conventions, and used associative matching methods to enable additional links to be made based on information from people within the same family unit. The pilot study linked data for refugees resettled between 2015 and 2020 under the Vulnerable Persons Resettlement Scheme (VPRS) and Vulnerable Children’s Resettlement Scheme (VCRS) to Home Office Borders and Immigration (HOBI) data, and NHS Personal Demographics Service (PDS) data. For more information, see Refugee Integration Outcomes (RIO) data linkage pilot.

In June 2022, we extended the RIO dataset to include refugees living in England and Wales who were granted asylum and refugee status between 2015 and 2020 and included linkage to the ONS Census 2021 data. ONS further published Early integration outcomes for refugees resettled in England and Wales: 2015 to 2021 analysis article covering self-reported health, access to health services, illness, disability and death, housing, and migration moves after initial accommodation (to within England and Wales and outside) in June 2023. Further information, including methodology and quality details of the linked administrative datasets, can be found on the Refugee integration outcomes data-linkage pilot: Census 2021 linkage methodology.

Innovative analysis, promising results and close collaboration between HO, ONS and MHCLG on this study enabled HO to secure an additional £2.9m of funding through the Shared Outcomes Fund Round Three between 2023 and 2025 to further extend the RIO dataset. This enabled work on improved linkage methodology and linkage to additional administrative data sources, including Birth Notifications data (NHS Digital), Birth Registrations and Death Registrations (General Register Office), and updated HOBI data and NHS PDS data. We also included data for approximately 500 additional refugees who either arrived in 2014, or who were scheduled to arrive in 2020 via the VPRS/VCRS schemes but arrived in early 2021 due to the COVID-19 pandemic travel restrictions.

1.2 The VPRS/VCRS schemes

The UK government established the Vulnerable Persons and Vulnerable Children’s Resettlement Schemes to resettle vulnerable adult and child refugees. The VPRS was launched in 2014 for refugees displaced by the Syrian conflict and aimed at those in greatest need, including people requiring urgent medical treatment, survivors of violence and torture, and women and children at risk. The scheme aimed to resettle 20,000 people fleeing the conflict in Syria by March 2020.

The VCRS was launched in 2016 with the aim of resettling up to 3,000 at-risk children and their families from the Middle East and North Africa (MENA) region. The region comprises 19 countries, including Iraq, Syria and Yemen.

By the time the schemes closed in February 2021, 20,319 individuals were resettled in the UK through the VPRS and a further 1,838 were resettled through the VCRS. Both schemes closed to new arrivals in February 2021 following lower numbers of new arrivals because of COVID-19-related travel restrictions. A new UK Resettlement Scheme (UKRS) was subsequently launched through which vulnerable refugees of any nationality could be resettled to the UK.

By design, refugees resettled under the VPRS were primarily of Syrian nationality (91%). There were slightly higher proportions of males (52%) than females (48%), and half (50%) were under the age of 18 on arrival. Further data for resettled refugees is available in regularly published Home Office (HO) Asylum and resettlement datasets.

We continued to focus on this scheme beyond the initial pilot since there was confidence that the linkage rate will be high when new datasets are linked, and we could make comparisons with asylum route refugees. We refer to refugees resettled under the VPRS and VCRS as resettled refugees (RR) in this report.

1.3 Asylum route refugees

Asylum route refugees are individuals who arrived in the UK and subsequently applied for and were granted asylum and refugee status. Asylum seekers may have arrived in the UK via an irregular route without permission to enter, or via a regular route such as a visa, and subsequently claimed asylum. We refer to asylum route refugees as ARR in this report.

The RIO dataset contains data on approximately 97,000 individuals who were granted asylum between 2015 and 2020 and living in England and Wales. This cohort excludes those still awaiting a decision on their asylum claim, or those who were denied asylum. Many asylum refugees are young adults (see HO Asylum and resettlement datasets), with approximately one-third being female and two-thirds male. As noted in the Refugee integration outcomes data-linkage pilot: Census 2021 linkage methodology, ARR data are more challenging to link due to the refugees arriving via this route tending to be more transient and dispersed across England and Wales, being less likely to arrive in familial groups, and comprising a higher proportion of younger males who, regardless of refugee status, are in general less likely to engage with official surveys and government surveys (as reported in the Patient Register: quality assurance of administrative data used in population statistics, Dec 2016 article). The majority of the ARR population in this study is from Iran, Eritrea, Sudan, Syria and Afghanistan. Further data for asylum refugees is available in regularly published HO Asylum and resettlement datasets.

1.4 Data and analysis presented in this report

The analysis presented in this report is exploratory and focuses on a subset of topics chosen for their relevance to policy priorities, the availability of data, and the potential to showcase the dataset’s analytical capabilities. Further information on data and methods is provided in Appendix 1: Data quality, methodology and limitations.

This report examines the differences in outcomes between the RR and ARR groups, comparing them, where possible, to the usual resident population of England and Wales. It also explores the distinct patterns of characteristics and behaviours between cohorts that had been in the UK for different lengths of time. We report analysis on numbers of refugees who remained in or left the UK since arrival, their economic activity status, labour market participation, and household-level housing analysis. The employment and housing-related outcomes presented in this analysis are considered fundamental indicators of refugee integration.

These are not official statistics of the refugee population, and the numbers quoted here may differ from official statistics. The analysis and results may change as we continue to update linkage methods. The linkage rates for the ARR group were lower and therefore analysis of this group should be used and interpreted with more caution. These groups represent a subset of refugees in England and Wales and conclusions from this study should not be used to make inferences about the wider refugee population. Official statistics on asylum and resettled refugees are available in regularly published HO Asylum and resettlement datasets.

2. Analysis

2.1 Refugee departure from the UK

There has been limited information on whether refugees remain in the UK long term or leave shortly after their arrival. By linking RIO cohort data to the HOBI data, we looked at how many refugees who originally resettled or were granted asylum living in England and Wales between 2015 and 2020 departed or remained in the UK at the time of the HOBI data extract (31 December 2022). Where refugees had departed, we determined how long they spent in the country before departing. Embark status is susceptible to future change, especially for those who departed close to the data extraction date, as some may subsequently return.

Further information about the definition of embarkation, methodology used and more explanatory notes for aiding the interpretation of the following data and analysis can be found in Appendix 1: Data quality, methodology and limitations.  

Table 1 shows that most refugees in the RIO dataset remained living in the UK. Although few refugees left the UK overall, there was a higher percentage of people leaving the UK from the ARR group (6%) compared with RR group (2%). This difference may indicate greater geographical mobility within the ARR cohort, which is known to be more transient, potentially due to the ARR cohort including a higher proportion of younger males without families compared with the RR cohort. Refugees in the RR cohort were usually resettled as family units and receive more substantial support from local authorities or the state in areas such as housing and healthcare upon arrival.

Table 1: Number and percentage of refugees who remained in or left the UK by RIO cohort

Arrival route	Number (percentage) of people who remained	Number (percentage) of people who left
RR	12,598 (98.3%)	216 (1.7%)
ARR	71,007 (93.9%)	4,598 (6.1%)

Source: Linked VPRS, VCRS, and asylum refugee data to Home Office Borders and Immigration data

We also investigated how long refugees spent in the UK before their latest recorded departure, based on either the year asylum was granted for the ARR cohort, or the arrival year for the RR cohort. Figure 1 shows that of those who have left, most did so after staying in the UK for 3 years or longer.

Figure 1: Length of time in years before leaving the UK by RIO cohort

Source: Linked VPRS, VCRS, and asylum refugee data to Home Office Borders and Immigration data

Notes:

1. Due to statistical disclosure control rules, some values are suppressed and may not appear on this graph.
2. Percentages are based on a very small number of refugees leaving each year, with a median of 680 for ARR and 50 for RR. The RR series represents 1.7% of all RR and the ARR series represents 6.1% of all ARR.

There is no notable difference in the amount of time spent in the UK before leaving between ARR and RR groups. Although the reasons for leaving the UK are unknown, some may include movement to another country to join family or, in some cases, possibly a return to their home country.

2.2 Labour market integration

Participation in the labour market is widely seen as a positive integration outcome within both refugee communities and receiving communities. Work is valuable in establishing social roles, developing language and broader cultural competence, and establishing social connections. Work is not, therefore, solely an integration outcome but a vital facilitator of further integration as noted in the Home Office Indicators of Integration framework.

We explored the economic activity outcomes for the ARR and RR refugees aged 16 and over and compared these with outcomes for the usual resident population of England and Wales from Census 2021. Further information about the definition of economic activity, the methodology used and more explanatory notes for aiding the interpretation of the following data and analysis can be found in Appendix 1: Data quality, methodology and limitations.

2.2.1 Economic activity status by RIO cohort

Compared with the working-age population of England and Wales – where in 2021, 57% of the population was in employment and 3% were unemployed – we see much smaller percentages of RIO refugees in employment, and larger percentages being unemployed (Figure 2).

Figure 2: Economic activity status (Census 2021, classification 4a) by RIO cohort compared with the usual resident population of England and Wales

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

In the RR cohort, a similar percentage of refugees were in employment (12%) as were unemployed (11%). In the ARR cohort, around 3 times as many refugees were in employment (37%) compared with unemployment (13%).

The large difference in the percentages of the RR cohort and the ARR cohort in employment might be partly explained by the demographics of each cohort. A majority of ARR are single, working-age males who may have greater availability and flexibility to take up employment. In contrast, RR refugees were resettled under the VPRS/VCRS based on the degree of identified vulnerabilities and needs. They are also more commonly resettled as family units, which may impact their ability to enter the labour market as quickly as ARR refugees.

2.2.2 Economic activity status by sex

As shown in Figure 3, within both RIO cohorts, a larger percentage of men were economically active compared to women. In the RR cohort, similar percentages of men were in employment and unemployed (19% and 15%, respectively). By comparison, in the ARR cohort, a much larger percentage of men were in employment (44%) compared with unemployed (15%).

Overall, we see much lower percentages of women refugees being economically active than men in both RIO cohorts. While in the ARR cohort, nearly a quarter of women were in employment (24%), in the RR cohort there were fewer than 5% of women in employment. In Section 2.2.4, we explore reasons for differences in economic inactivity between men and women for the RIO cohorts to further understand potential barriers to employment.

Figure 3: Economic activity status (Census 2021, classification 4a) by sex and RIO cohort

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

2.2.3 Economic activity status by length of time in England and Wales

Within each RIO cohort, we analysed how economic activity varied with length of time after arrival (for people in the RR cohort), or length of time after being granted asylum (for people in the ARR cohort). Note that RR, being accepted as refugees on arrival, immediately have the right to work in the UK; however, ARR do not have the right to work in the UK until their asylum application has been decided and their claim granted. The time lag between ARR arrival and receiving refugee status may give this cohort the opportunity to become more familiar with the labour market and enter employment more quickly once they receive refugee status.

Figure 4: RR economic activity status (Census 2021, classification 4a) by year of resettlement

Source: Linked VPRS and VCRS refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. Due to statistical disclosure control rules, some values are suppressed and may not appear on this graph.
2. The additional 2021 data for the RR cohort results from some people scheduled to arrive in 2020 but were delayed until early 2021 due to COVID-19 travel restrictions.

Figure 5: ARR economic activity status (Census 2021, classification 4a) by year granted asylum

Source: Linked asylum refugee data to Census 2021 data from the Office for National Statistics

In both the RR and ARR cohorts, the percentage of people in employment was greater amongst refugees who had been resettled or granted asylum in earlier years (Figure 4 and Figure 5). The data suggests that the longer refugees have been in the country, the more likely they are to be economically active and employed, and the less likely they are to be economically inactive.

For the ARR cohort, there was a 27 percentage point gain in people in employment between the most recently arrived cohort from 2020 (20%) and the earliest arriving cohort in this data from 2015 (47%). The RR group followed a similar trend, although the earliest arriving RR cohort reached a lower employment rate (19%) around 5 years after arrival. In both the 2020 and 2021 cohorts, the percentage of RR individuals in employment was extremely low. It is important to note here that the COVID-19 pandemic may have contributed to the extremely low economic activity levels observed in the 2020 and 2021 cohorts, particularly for those in the RR category. Individuals resettled in 2020 or 2021 faced an exceptionally unusual economic environment, besides the other challenges new refugees face. In contrast, those granted asylum status in 2020 likely arrived before the pandemic began and so may have had slightly more time and opportunities to find employment.

Looking specifically at the 2015 ARR cohort, the percentage of economically active refugees nearly matched the levels seen in the usual resident population of England and Wales aged 16 years and over (58% versus 61%). Although the employment rate was still 10 percentage points below that of the usual resident population (47% versus 57%), 5 to 6 years after being granted asylum, the economic activity levels of the ARR cohort were nearing those of the overall usual resident population. This demonstrates refugees’ broadly successful integration into the labour market over the medium term.

It is plausible that the lower economic activity levels observed in the RR cohort could be attributed to greater difficulties in entering employment due to the more extensive physical and mental health challenges. The VPRS and VCRS aimed to help those in the greatest need, including people requiring urgent medical treatment, survivors of violence and torture, and women and children at risk. These findings may not represent other RR cohorts. We consider the reasons for economic inactivity in more detail in Section 2.2.4.

Further analysis by sex, found in our published data table (Table I) on the landing page, showed that ARR men saw the greatest percentage point increase (30 percentage points) of those in employment over the 5-year period. For the 2015 cohort, the percentage of ARR men who were economically active was 7 percentage points greater than the usual resident population, with the percentage in employment being similar. RR women saw the smallest percentage point increase (8 percentage points) of those in employment over the 5-year period.

In all groups, however, the percentage point difference in employment rates by year of resettlement or year granted asylum was always larger than the percentage point difference between economic inactivity and economic activity rates over a 5-year period.

In their 2018 article Triple Disadvantage? A first overview of the integration of refugee women, Liebig and Tronstad found that employment outcomes for refugee men stabilise after approximately 8 years; however, for women, this can take longer than 12 years (Liebig & Tronstad, 2018). As noted in the cited report, this is due to refugee women facing challenges relating to poorer health and lower education, but also additional childcare or family responsibilities.

2.2.4 Reasons for economic inactivity

Analysis in Section 2.2.1 showed that larger proportions of RR and ARR refugees were economically inactive compared with the usual resident population of England and Wales. Figure 6 to Figure 9 show the reasons stated for economic inactivity in Census 2021, split by RIO cohort, sex, and year of resettlement or year granted asylum.

In both RIO cohorts, regardless of the year of resettlement or year granted asylum, most women were economically inactive because they were looking after the home or family. However, looking across cohorts, a much higher percentage of RR women reported looking after family as their reason for economic inactivity compared to ARR women. The distribution of refugees between economic inactivity categories did not appear to fluctuate much by year of resettlement or year granted asylum.

ARR women who arrived earlier were slightly more likely to cite looking after home or family as their reason for being economically inactive than those who arrived later, while the opposite is true for RR women. Although the changes were rather small, it might suggest that ARR women who had been resettled for a longer period were more likely to have family caring responsibilities, compared with ARR women who had more recently been granted asylum.

Figure 6: Economic inactivity category (Census 2021, classification 9a) of RR women by year of resettlement

Source: Linked VPRS and VCRS refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. Due to statistical disclosure control rules, some values are suppressed and may not appear on this graph.
2. The additional 2021 data for the RR cohort results from some people scheduled to arrive in 2020 but were delayed until early 2021 due to COVID-19 travel restrictions.

Figure 7: Economic inactivity category (Census 2021, classification 9a) of ARR women by year granted asylum

Source: Linked asylum refugee data to Census 2021 data from the Office for National Statistics

Figure 8: Economic inactivity category (Census 2021, classification 9a) of RR men by year of resettlement

Source: Linked VPRS and VCRS refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. Due to statistical disclosure control rules, some values are suppressed and may not appear on this graph.
2. The additional 2021 data for the RR cohort results from some people scheduled to arrive in 2020 but were delayed until early 2021 due to COVID-19 travel restrictions.

Figure 9: Economic inactivity category (Census 2021, classification 9a) of ARR men by year granted asylum

Source: Linked asylum refugee data to Census 2021 data from the Office for National Statistics

The greatest proportion reported other reasons for economic inactivity; however, we cannot speculate on what these reasons might be. ‘Other’ was a tick-box option on the Census 2021 form and respondents did not have the option to provide further information.

A much smaller proportion of men cited looking after home or family as their reason for economic inactivity compared with women in both cohorts. Instead, we saw men reporting studying as a more frequent reason compared with women. However, when examining yearly cohort data, we saw economically inactive men, but also women, who arrived in later years reported studying more frequently than men and women who arrived in earlier years – a pattern observed in both the ARR and RR cohorts. The difference was generally much larger for men than women in both cohorts, and perhaps reflective of the men’s greater success in obtaining employment. When comparing cohorts across time, there was little difference in the proportions of RR men and ARR men, as well as RR women and ARR women, reporting studying as their reason for economic inactivity. Similar percentages of both RR and ARR men, and RR and ARR women respectively, cited studying as the reason for their economic inactivity.

When looking at the percentages of refugees who cited long-term sickness or disability as a reason for economic inactivity, we found that this reason was reported more frequently by men in both cohorts than by women – although only a relatively small proportion of refugees. More RR men reported this compared to ARR men, with it being the second largest reason for economic inactivity among RR men settled in 2015. Male and female refugees who had been resettled or granted asylum in earlier years reported long-term sickness or disability more frequently than those in later years, regardless of cohort.

Overall, very few refugees were economically inactive due to retirement.

There might also be a cultural element associated with these results, specifically the perceived role of women primarily as caregivers and homemakers and established norms of men going to work. Studies show that immigrants from more patriarchal cultures, particularly from parts of Africa and the Middle East, often exhibit stronger traditional gender role patterns, and that gender differences do play a role in labour market outcomes for these immigrants in Europe (Adsera & Chiswick, 2007).

Overall, the findings are consistent with observations in the 2018 article Triple Disadvantage? A first overview of the integration of refugee women, where Liebig and Tronstad found that, in several other countries, women refugees have a lower starting point in terms of employment rates, and their initial progress in terms of improvements in employment rates is slower. However, the data cannot tell us whether caregiving, a lack of skills, or other factors are the primary barriers to women’s employment.

Further breakdown of reasons for economic inactivity by age category is provided in the published data table (Table L) accompanying this report. 

2.2.5 Working patterns – hours worked per week

We also analysed the working patterns of RIO refugees in employment at the time of Census 2021, specifically looking at the number of hours they reported working per week. In both RIO cohorts, more than half of refugees in employment worked part-time (up to 30 hours a week) compared with full-time (31 or more hours a week). This differs from the working patterns for all usual residents in England and Wales, where 70% work full-time and 30% part-time. However, there was a more even split observed for the ARR cohort in working patterns compared to the RR cohort, who predominantly reported working part-time (Figure 10).

Full results and figures can be found in our published data table (Table M to Table O) on the landing page.

Figure 10: Number of hours worked per week (Census 2021) by RIO cohort compared with the usual resident population of England and Wales

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

We further analysed how working patterns of RIO refugees varied between men and women as well as with length of time after arrival (for people in the RR cohort), or length of time after being granted asylum (for people in the ARR cohort).

Overall, a higher proportion of refugee women were working part-time (up to 30 hours a week) compared with full-time (31 or more hours a week) in both RIO cohorts (Figure 11). For refugee men, we observed roughly equal proportions of ARR men working part-time and full-time, whereas for RR men there was a much higher proportion of men working part-time than full-time.

Figure 11: Number of hours worked per week (Census 2021) by sex and RIO cohort

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. As seen in Section 2.2.2, RR women had very low employment rates. Analysis of the number of hours worked by RR women was based on a very small population and should be interpreted with caution.

No substantial difference was observed in reported working patterns between RIO men and women with length of time after arrival. We saw only a minor trend for RR men potentially transitioning from part-time hours to full-time hours with increasing time in the UK, but the differences in proportions were overall low. Generally, there was no distinct trend of any increase or decrease in proportions reporting working part-time or full-time with length of time in the UK for most of the RIO refugees.

Full results and figures can be found in our published data table (Table M to Table O) on the landing page.

2.2.6 Occupation types

We also looked at the types of occupations of RIO refugees in employment. Information on the types of occupations included in each broad category is provided in Appendix 1: Data quality, methodology and limitations.

Figure 12 shows that working refugees in the RR cohort were most likely to be employed in the elementary occupations (29%) although around a quarter were also employed in skilled trades occupations (24%). Employment in other occupation categories ranged from 2% employed in administrative and secretarial occupations to 13% employed as process, plant and machine operatives.

Similarly, ARR refugees were also most likely to be employed in the elementary occupations (23%). The remaining people in employment were relatively evenly distributed between the other occupation categories, with percentages ranging from 5% employed in administrative and secretarial occupations to 15% employed in sales and customer service occupations.

Figure 12: Occupation type (Census 2021) by RIO cohort

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

2.3 English proficiency

Proficiency in speaking the English language is an important facilitator of other integration outcomes, including securing employment and forging social connections. Here we analyse how the self-reported language proficiency of refugees changed with length of time in the UK and how it correlated with economic activity.

In Census 2021, people in England were asked whether their main language was English and people in Wales were asked whether their main language was English or Welsh. Approximately 4% of those of working age in the RR cohort and 20% in the ARR cohort reported their main language was English (or Welsh in Wales).

Generally, we observed the baseline level of English to be better for the ARR than for RR cohorts. This could indicate their countries of origin, particularly if a proportion of the ARR cohort come from countries where English is commonly spoken (see below). For both cohorts, English proficiency levels showed an increase with more time spent in the country, and this was particularly more pronounced in the RR cohort. This could indicate that the English language support programmes available to RRs upon arrival are likely to have a positive overall impact.

Looking at overall differences between men and women, in both RIO cohorts, we did not observe large differences in the levels of language proficiency. We see a trend of a slightly higher percentage of men reporting speaking English ‘very well’ than women, but overall, we did not see a large difference.

We have further explored the language proficiency of the refugees by age category and sex. Across both cohorts and sex categories, language proficiency was higher for younger age categories, being the strongest for the 16 to 24 age category. We also saw that across both cohorts, slightly higher proportions of younger men compared to women reported overall better English.

We saw the lowest levels of English proficiency in the 65 and over age category. For this age category, in both RIO cohorts, we saw higher percentages of women reporting not speaking English at all compared to men in the same age category.

There are several potential reasons for the differences seen between the age categories and the sexes. As noted, the RR cohort mainly comprises nationals from Syria and other middle eastern countries where English might not have been as prevalent. The ARR cohort contains more people from Eritrea, Pakistan and Sri Lanka, where English is often widely spoken. For instance, in Eritrea, English is considered a working language and in Pakistan, it serves as an official language and is used as the key language in education and government.

The younger age categories, especially among the ARR groups, may be more likely to have had higher exposure to English beforehand given the increased prevalence of English-speaking media across the wider world in recent years (Sundqvist & Sylvén, 2016). This may have been something that the older age categories were not exposed to. Additionally, those in the younger age categories are more likely to be in education or employment, which provides more opportunities for development of their language skills, both in formal and informal situations (Dustmann & Fabbri, 2003). This is especially true for ARR men, as we saw in Section 2.2.

The above considerations are most likely to affect older RR women. We saw in Figure 2 that RR women were the group with the smallest proportion in employment. We also see from Table G in the published data tables that only a small proportion of women over 65 years was in employment. They are less likely to have had prior exposure to English before arrival in the UK, and cultural differences with more traditional values of women not working or acting primarily as caregivers are most likely to be observed in older members of the population than younger (Charrad, 2011).

Plausibly, there is a link between English language proficiency and wider wellbeing, as well as integration. We know from the previous ONS article Early integration outcomes for refugees resettled in England and Wales: 2015 to 2021 that the percentage of those who reported very good health tends to decrease from their mid-40s. Further planned linkage of health data to RIO will enable us to explore the interplay between language proficiency and wellbeing more fully.

Full results for English language proficiency can be found in our published data tables (Table Q to Table U) on the landing page.

2.3.1 Economic activity status by language proficiency

To understand potential influences of English language levels on economic activity trends, we looked at RIO refugees’ language proficiency in the context of their economic status.

Figure 13 and Figure 14 show that across both RIO cohorts, refugees who reported better English language proficiency were more likely to be economically active. Simultaneously, lower levels of economic inactivity were reported among refugees with greater English language proficiency. This pattern is much more evident for the ARR cohort than the RR cohort, where we see much higher levels of employment amongst those refugees reporting English as their main language or speaking it ‘very well’ and ‘well’ compared with those not speaking English well or at all.

For the RR cohort, there was a 15 to 20 percentage point difference in employment rates between refugees reporting that they cannot speak English at all and those who reported speaking English well or better. In contrast, for the ARR cohort, the difference in employment rates is much higher between refugees reporting that they cannot speak English at all and those who reported speaking English well or better, ranging between 32 to 41 percentage points difference.

It is interesting to note, that amongst both cohorts, those who reported English as their main language, have slightly lower rates of employment than those whose main language is not English and speak English very well (therefore falling into the next highest level of English language proficiency), with trends being more evident for the RR cohort. Although we cannot easily identify the reason for this with RIO data, this might indicate that there are other factors at play and English language proficiency is potentially not the strongest or lone factor in facilitating employment.

Figure 13: Economic activity status (Census 2021, classification 4a) of RR by English language proficiency (Census 2021)

Source: Linked VPRS and VCRS refugee data to Census 2021 data from the Office for National Statistics

Figure 14: Economic activity status (Census 2021, classification 4a) of ARR by English language proficiency (Census 2021)

Source: Linked asylum refugee data to Census 2021 data from the Office for National Statistics

We see higher proportions of refugees being economically active within those reporting higher English levels. However, even when accounting for potential differences in English language skills, ARR were more likely to be in employment than RR. It is possible that the length of time awaiting an asylum application decision influenced this. Although ARR cannot legally work in England and Wales until asylum has been granted, they may be more aware of the job market and opportunities. However, we know that the RR cohort includes a greater proportion of women and dependent children (see Household composition data in Table AE), and people with potentially greater vulnerabilities whose exposure to English may be more limited compared with the ARR cohort.

It is important to note that English language proficiency is an outcome that not only facilitates other outcomes but is also itself influenced by other outcomes. Language proficiency may help improve the prospect of acquiring employment, but employment also helps the development of language proficiency. It is not possible to know to what extent one affects the other, but the steady improvement we see in English language proficiency with longer time living in the UK is clear in the data. Improved language proficiency early on may be expected to positively influence refugee integration into society and facilitate both their short-term and longer term outcomes.

2.4 Housing and living conditions

Housing is an important integration outcome. Housing influences a community’s sense of security and stability, opportunities for social connection, and access to healthcare, education and employment. Like economic integration, housing is not solely an outcome, but an important facilitator of further integration.

2.4.1 Residence type

We used the Residence type variable: Census 2021 to group people according to whether they reported living in a household or a communal establishment (CE) at the time of Census 2021. Examples of a CE include university halls of residence, boarding schools, armed forces bases, hospitals, care homes, prisons and more. More information about households and CEs can be found in Appendix 1: Data quality, methodology and limitations.

Nearly all RIO refugees (99%) reported living in households, with only 1% living in CEs. Broken down by arrival route, we found that those in the ARR cohort were slightly more likely to be in a CE, but the proportions were still very low (2% for ARR compared to <1% for the RR cohort).

Figure 15: Communal establishment (CE) types (Census 2021) for RIO refugees who reported living or staying temporarily in a CE

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. Analysis only includes those who reported living or staying temporarily within the CE, thereby excluding those who indicated that they worked and lived there. We used the ONS methodology described in the article People experiencing homelessness, England and Wales: Census 2021 to identify members of our cohort who were at hostels or temporary shelters for homeless people in Census 2021. This does not cover all people experiencing homelessness, as others may have been in other types of accommodation, or rough sleeping, at the time of the census, so they may not have been captured in or responded to the Census 2021.

As shown in Figure 15, 42% of refugees flagged as living or staying temporarily in a CE were living in a hostel or temporary shelter for the homeless. This corresponds to approximately 0.5% of the entire RIO refugee population at the time of Census 2021.

Given that the percentage of refugees living in communal establishments was low for both cohorts, we do not report further analysis for those refugees.

2.4.2 Accommodation type

We investigated the accommodation type of those that live in households. The ONS previously published data on refugee housing outcomes outcomes in their article Early integration outcomes for refugees resettled in England and Wales: 2015 to 2021, but this was at an individual level. The data here are reported at a household level. Please refer to Appendix 1: Data quality, methodology and limitations for further information on household analysis presented here.

Figure 16 shows that most ARR households (59%) reported living in a flat, maisonette or apartment. Although the most frequently reported category for RR households was also flat, maisonette or apartment (29%), a higher proportion of households reported living in detached, semi-detached and terraced houses compared with ARR households. The proportion of refugee households living in flats, apartments or maisonettes is also higher than the 22% reported for the usual resident population of England and Wales in Census 2021. These differences could, in part, reflect household composition, with RR refugees more likely to be resettled as a family unit and therefore requiring larger properties. The breakdown of housing by accommodation type in the RR cohort was much more closely aligned with the usual resident population of England and Wales.

Figure 16: Household accommodation type (Census 2021, classification 5a) by RIO household route compared with the usual resident population of England and Wales

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. Due to statistical disclosure control rules, some values are suppressed and may not appear on this graph.

When analysing accommodation types by the year of resettlement or the year asylum was granted, we found no notable changes in patterns across different cohort years for both RR and ARR.

Full results and figures for household accommodation type can be found in our published data table (Table Y and Table Z). Further detail on how year of arrival was determined for households is available in Appendix 1: Data quality, methodology and limitations.

2.4.3 Occupancy rating for bedroom (overcrowding) by year of arrival

The Census 2021 occupancy rating for bedrooms variable is a household-level variable that tells us whether an accommodation is overcrowded, ideally occupied (or ‘occupied-to-standard’) or under-occupied.

Please refer to Appendix 1 for further information about the definition of overcrowding, methodology used, and more explanatory notes to aid interpretation of the following data and analysis. 

Figure 17 shows the majority of both ARR and RR households were occupied-to-standard, with most properties having the required number of bedrooms for the household size and composition. The percentage of refugee households living in overcrowded accommodation was similar for both ARR and RR households, at 19% each. This is notably higher than the overall proportion of households reported by Census 2021 as living in overcrowded accommodation in England and Wales, which was 4%.

Figure 17: Household occupancy rating (bedroom standard) by RIO household route compared with the usual resident population of England and Wales

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

Looking at occupancy rating by year of resettlement or asylum granted and comparing Figure 18 and Figure 19 we can see that there was a very little difference in overcrowding for RR and ARR households who had been in the UK for several years. In RR households for the 2015 cohort, 28% were overcrowded compared with 24% of ARR households of the 2015 cohort.

Figure 18: RR household occupancy rating (bedroom standard) by earliest year of resettlement

Source: Linked VPRS and VCRS refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. Due to statistical disclosure control rules, some values are suppressed and may not appear on this graph.

Figure 19: ARR household occupancy rating (bedroom standard) by earliest year granted asylum

Source: Linked asylum refugee data to Census 2021 data from the Office for National Statistics

There does, however, appear to be a slight trend where overcrowding is higher for refugee households who have been in the UK for longer for both RIO cohorts. Potential reasons for an increased percentage of overcrowding with length of time in the UK might include further family reunification or having more children.

2.4.4 Household tenure by year of arrival

Tenure is a Census 2021 household-level variable that indicates whether a household owns or rents the accommodation that it occupies.

The analysis and results reported below were created following a slightly different methodology compared to the rest of the housing analysis. Please refer to Appendix 1: Data quality, methodology and limitations for further information about the definition of tenure, methodology used and more explanatory notes for aiding the interpretation of the following data and analysis.

As shown in Figure 20, most refugee households live in rented accommodation, with most ARR households living in private rented accommodation and most RR households living in social rented accommodation. The percentage of households living in owner-occupied accommodation was very low for both cohorts (1% for RR households, and slightly higher at 4% for ARR households). These patterns are notably different to the Census 2021 figures for the usual resident population of England and Wales, where most households (63%) live in owner-occupied accommodation, and roughly an equal proportion of the rest of the population live in private rented accommodation (20%), and in social rented accommodation (17%).

Figure 20: Household tenure (Census 2021) by RIO household route compared with the usual resident population of England and Wales

Source: Linked VPRS, VCRS, and asylum refugee data to Census 2021 data from the Office for National Statistics

Figure 21 and Figure 22 show a similar trend in the tenure by year of resettlement or asylum granted for RR and ARR, respectively. A higher percentage of refugees who had been in England and Wales longer were living in social rented accommodation than later arriving refugees. Approximately 69% of RR households who arrived in 2015 lived in social rented accommodation compared with 44% who arrived in 2021, and 52% of ARR who arrived in 2015 lived in social rented accommodation compared with 33% of those who arrived in 2020. The exception to this trend was the RR cohort who arrived in 2020, with relatively high percentage (59%) living in social rented accommodation.^{[footnote 4]} Overall, RR were more likely to be in social housing than ARR. This is to be expected because RR are supported into housing by local authorities on arrival, whereas ARR would have to present themselves to local authorities for support.

Home ownership among RR was very low, and due to very low counts, we cannot comment on any trends by year of resettlement. However, we observe a modest increase in the percentage of ARR who own their own home with time in England and Wales. Close to 3% of ARR households who arrived in 2020 were identified as owner-occupied households on the Census 2021 compared with 4.4% for those who arrived in 2015. The previous ONS report 2011 Census Analysis: Social and Economic Characteristics by Length of Residence of Migrant Populations in England and Wales indicated that home ownership among all migrants (not exclusively refugees) generally increases over time, eventually reaching or surpassing the average for other households. However, this process tends to be slower, with the high levels of home ownership noted in the report observed mainly among those who have been in the UK for over 30 years, reflecting the substantial investment required.

Figure 21: RR household tenure (Census 2021) by earliest year of resettlement

Source: Linked VPRS and VCRS refugee data to Census 2021 data from the Office for National Statistics

Notes:

1. Due to statistical disclosure control rules, some values are suppressed and may not appear on this graph.

Figure 22: ARR household tenure (Census 2021) by earliest year granted asylum

Source: Linked asylum refugee data to Census 2021 data from the Office for National Statistics

Overall, data shows refugees were more likely to be in social rented accommodation the longer they had been in the UK. This may be due to the limited availability of social housing, requiring them to rely on private rental accommodation until social housing becomes available.

Appendix 1: Data quality, methodology and limitations

Data and analysis presented in this report

The Refugee Integration Outcomes (RIO) analytical dataset used in this report included data for 16,757 refugees who arrived in England and Wales via the Vulnerable Persons Resettlement Scheme (VPRS) and Vulnerable Children’s Resettlement Scheme (VCRS) routes between 2014 and 2021, and 96,967 refugees who were granted asylum between 2015 and 2020 and were residing in England and Wales. The findings from these groups represent only a subset of refugees and should not be generalised to other refugee populations who arrived in the UK through different schemes or routes.

To date, the RIO analytical dataset includes Home Office (HO) refugee data linked to ONS Census 2021 data for England and Wales, NHS Digital Birth Notifications data, General Register Office Birth Registrations and Death Registrations data, the NHS Personal Demographic Service (PDS), and Home Office Borders and Immigration (HOBI) data. More information about these data sources can be found in the Sources of data article on the Office for National Statistics (ONS) website. More information on the linkage methodology can be found in the previously published RIO report Refugee integration outcomes data-linkage pilot: Census 2021 linkage methodology on the ONS website. Most of the outcomes presented in this report are derived from the linked Census 2021, with additional embarkation insights sourced from the linked HOBI data.

The census is a rich data source that includes numerous integration indicators for a point in time, such as families and households, housing, education, English and Welsh language proficiency, employment and health.

Census 2021 linkage rates, although improved from when last published in June 2023, were 91% for the RR group and 72% for the ARR group. There are several reasons why the remaining refugee records may not have linked to Census 2021 data, for example:

• we may have missed links (false negative linkage errors)
• refugees may have moved from England and Wales to Scotland or Northern Ireland before or soon after 21 March 2021 (Census Day), so would not have completed a Census 2021 questionnaire for England and Wales
• refugees may have left the UK before or soon after 21 March 2021 – an embark event
• refugees may have died before or soon after 21 March 2021
• refugees may have failed to complete a Census 2021 questionnaire (census non-response) – coverage adjustment methods account for census non-response when producing census estimates; however, for linkage purposes, we have used census records before any coverage adjustments were applied

In the RR group, linkage rates were similar across age, sex and nationality, apart from a noticeably lower linkage rate for Eritrean refugees. In the ARR group, linkage rates varied more, with lower linkage rates for males compared with females, and lower linkage rates for those in the 25 to 44 age group compared with those aged zero to 24 and 45 and over groups. Looking at rates for the most common nationalities (where there were 1,000 records or more) showed that linkage rates were lowest for Eritrean and Sudanese refugees.

The analysis presented in this report, including the differences in linkage rates, emphasises the demographic and social differences between the RR and ARR groups. It is therefore advisable to conduct analysis and explore integration outcomes separately for the 2 refugee cohorts. 

We removed all duplicate records where more than one refugee was linked to the same Census 2021 record, and only present analysis for those refugees flagged as a usual resident, according to Census 2021. The final dataset used for analysis included approximately 81,880 linked records, 15,090 from the RR route and 66,790 from the ARR route.

Embark analysis

HOBI data are derived from a linked database that combines data from HO systems to build travel histories that comprise an individual’s travel into or out of the UK, together with data relating to their immigration status. The RIO analytical dataset was linked to the HOBI data with a match rate of 78% for ARR and 76% for RR. The linkage rate for the RR cohort is lower than that reported in the RIO pilot study due to the removal of false positive matches that were identified during an additional manual review. We have also applied additional filtering, cleaning and de-duplication to the data. Linkage rates may change across different iterations of the RIO dataset, as new data and methods are incorporated into the project. The analysis of the embark flag in Section 2.1 uses only linked records, excluding those that could not be linked to these datasets.

Embark flag

We used the linked HOBI data to create an embark flag, which is a person-level status indicating whether an individual in the RIO analytical dataset departed or remained in the UK at the time of the HOBI data extract (31 December 2022). We defined individuals who embarked as those having a departure date within the HOBI reference period without a subsequent arrival back into the UK at the time of data extract, while those remaining were those whose records did not have a departure date or a departure date and a later re-arrival date. The status was derived from the latest arrival and departure dates from HOBI travel histories dataset.

Embark status is susceptible to future change, especially for those who departed close to the extract date of 31 December 2022, as they may be more likely to return, resulting in a change in their status to remaining. Also, the longer-term movement of the more recent arrivals is not yet known. Future HOBI data linkage with more recent RIO cohorts will tell us more about the length of stay in the UK.

Labour market analysis

The analysis of labour market outcomes in Sections 2.2, 2.3, and 2.4 is presented for England and Wales only and is based on linked data. It excludes those who did not link to Census 2021 data.

Census 2021 took place during the COVID-19 pandemic, which was a period of unparalleled and rapid change. The national lockdown, associated guidance and furlough measures may have affected responses in certain areas of analysis. Please consider this context carefully when interpreting the results.

We present findings for members of the RIO cohort who, at the time of Census 2021, were:

• usual residents (see definition below)
• aged 16 years and over
• not students or schoolchildren living away during term-time (who would have been recorded at their term-time address) (see details below)

Census 2021 defines a usual resident as “anyone who on Census Day, 21 March 2021, was in the UK and had stayed or intended to stay in the UK for a period of 12 months or more or had a permanent UK address and was outside the UK and intended to be outside the UK for less than 12 months.” It is worth noting that the Census 2021 considers schoolchildren and students in full-time education who are studying away from home to be usually resident at their term-time address. Further detail can be found on the ONS webpage Schoolchild or full-time student indicator variable: Census 2021.

Further detail about measurements can be found on the ONS webpage Measurements used in Census 2021 data. Further Census 2021 information about the labour market for the England and Wales population, including publications, data and supporting information, can be found in the Labour market and travel to work: Census 2021 in England and Wales release on the ONS website.

Economic activity status

The Census 2021 economic activity status variable considers people as economically active if they were aged 16 years and over and, between 15 and 21 March 2021, were:

• in employment (an employee or self-employed)
• unemployed, but looking for work and could start within 2 weeks
• unemployed, but waiting to start a job that had been offered and accepted

Census 2021 considered people to be economically inactive if they were aged 16 years and over and did not have a job between 15 to 21 March 2021, had not looked for work between 22 February to 21 March 2021, or could not start work within 2 weeks. For more information about economic activity status, see the ONS webpage Economic activity status variable: Census 2021.

For the analysis presented in this report, we aligned with the analysis presented in Section 2 of the Economic activity status, England and Wales: Census 2021 statistical bulletin and used the economic activity status classification 4a. The census definition differs from the International Labour Organization definition used on the Labour Force Survey. Further information on how the census definitions differ from those used by the Labour Force Survey can be found in the ONS article Comparing Census 2021 and Labour Force Survey estimates of the labour market, England and Wales: March 2021.

The economically active population included people who were put on furlough at the time of Census 2021 and considered to be temporarily away from work. In England and Wales, HM Revenue and Customs statistics state that 3.8 million employees were enrolled on the Coronavirus Job Retention Scheme on Census Day and 1.8 million self-employed people made a claim to the fourth grant of the Self-Employed Income Support Scheme.

Because of the impact of the COVID-19 pandemic on the UK labour market, the estimated number of people who were economically inactive might be higher than anticipated in some areas. It is possible that some people on furlough may have self-identified as economically inactive, instead of temporarily away from work. Guidance was provided to help people on furlough answer the census questions about work. However, we are unable to determine how furloughed people followed the guidance. Please consider this context carefully when interpreting the results. Read more about specific quality considerations in the Labour Market Quality Information for Census 2021 methodology article.

The analysis in Section 2.2.4 focuses on the economic inactivity status of refugees who were 16 and over and economically inactive at the time of Census 2021. We used the economic inactivity categories that appear in the economic activity status classification 9a of the Economic activity status classifications: Census 2021.

Working hours

The analysis in Section 2.2.5 focuses on the number of hours worked by refugees who were over 16 and over and in employment at the time of Census 2021, using the primary and aggregate Hours worked classifications: Census 2021.

The Census 2021 definition of ‘hours worked’ is the number of hours worked per week (including paid and unpaid overtime) by an individual in the week before the census (between 15 and 21 March 2021). This covered the main job of anyone aged 16 years and over. See the ONS webpage Hours Worked Variable: Census 2021 for more information.

Occupation types

Census 2021 defines occupation as “what people aged 16 years and over do as their main job. Their job title or details of activities they do in their job and any supervisory or management responsibilities form this classification. This information is used to code responses to an occupation using the Standard Occupational Classification (SOC) 2020”. For more information about the census variable, see the ONS webpage Occupation (current) variable: Census 2021.

The analysis in Section 2.2.6 groups occupations of people in employment into 9 high-level occupation categories. Further information about the high-level occupation categories is provided in Section 2 of the SOC 2020 Volume 1: structure and descriptions of unit groups.

Language proficiency

In Census 2021, people in England were asked whether their main language was English and people in Wales were asked whether their main language was English or Welsh. Language proficiency is defined by Census 2021 as “how well people whose main language is not English (English or Welsh in Wales) speak English”. For more information, see the ONS webpage Proficiency in English language variable: Census 2021.

Housing and living conditions analysis

Residence type

We used the Residence type variable: Census 2021 to group people according to whether they reported living in a household or a communal establishment (CE) at the time of Census 2021. Census 2021 defines a household as “one person living alone, or a group of people (not necessarily related) living at the same address who share cooking facilities and share a living room or sitting room, or dining area”, and a CE as “a place that provides managed full-time or part-time supervision of residential accommodation”. Examples of a CE include university halls of residence, boarding schools, armed forces bases, hospitals, care homes, prisons and more. For more information on the other types of establishments included, see Communal establishment management and type variable: Census 2021.

Further detail about households and CEs, including definitions, can be found on the ONS website Measurements used in Census 2021 data.

To understand more about the type of CEs that the refugee cohorts were living in around the time of Census 2021, we filtered the dataset to only include those who reported living or staying temporarily within the CE, thereby excluding those who indicated that they worked or owned the CE and lived there (or were a family member of staff or owners of the CE and lived there). We did this to align with the methodology used in the ONS article People experiencing homelessness, England and Wales: Census 2021, and this removed approximately 20 refugees.

Note: CE analysis is carried on an individual basis, and not a household basis.

Household analysis

The concept of a household is important as Census 2021 variables relevant to housing and living conditions (such as overcrowding and tenure) are typically collected and presented at a household-level. More information about the definition of a household can be found on the ONS website Measurements used in Census 2021 data. The RIO analytical dataset, however, is a person-level dataset, so for the analysis in Section 2.4.2, 2.4.3 and 2.4.4, we have grouped refugees into households according to those flagged as being in the same household in Census 2021.

The 80,840 refugees in the RIO analytical dataset were grouped into 48,900 households. To enable household-level analysis by refugee route, we have derived a ‘household route’ variable which indicates whether all members of the household arrived via the asylum route (ARR household), the resettled route (RR household) or a mixture of routes (mixed households). There were very few mixed households that included refugees from both routes in the same household (<1%), with most households comprising refugees from just the ARR route (92%), or just the RR route (8%). No further analysis is presented for mixed households due to small counts. It is important to note the RIO analytical dataset only includes data on specific refugee cohorts, and we are only analysing data for those who we successfully linked to Census 2021.The derived household route variable does not therefore consider these households may also include refugees in our cohort who were not successfully linked to Census 2021 and/or other individuals in Census 2021 who are not part of the RIO cohorts.

Accommodation type

We used the Accommodation type variable: Census 2021 to determine the type of building or structure used by or available to the RIO households.

According to Census 2021, this could be ‘a whole house or bungalow’, ‘a flat, maisonette or apartment’ or ‘a temporary or mobile structure, such as a caravan’.

Whole house or bungalow describes a property type that is not divided into flats or other living accommodation. There are 3 types of whole houses or bungalows.

• detached, where none of the living accommodation is attached to another property but can be attached to a garage
• semi-detached, where the living accommodation is joined to another house or bungalow by a common wall that they share
• terraced, where a mid-terraced house is located between 2 other houses and shares 2 common walls; and an end-of-terrace house is part of a terraced development but only shares one common wall

Flat, maisonette or apartment describes an apartment or flat in a block of flats. A maisonette is a 2-storey flat.

Occupancy rating for bedrooms (overcrowding)

The Census 2021 occupancy rating for bedrooms variable is a household-level variable that tells us whether a household’s accommodation is overcrowded, ideally occupied (or ‘occupied-to-standard’) or under-occupied. The occupancy rating for bedrooms is obtained by subtracting a hypothetical number of bedrooms recommended for a household from the actual number of bedrooms it has available. A household is considered overcrowded if it has fewer bedrooms available than recommended (a negative occupancy rating), ideally occupied (or occupied-to-standard) if it has the correct number of bedrooms recommended (occupancy rating of zero), or under-occupied if it has more bedrooms than needed (positive occupancy rating).

The number of bedrooms the household requires is calculated according to the Bedroom Standard, which specifies who, based on age, sex and relationship, can share a bedroom (see the ONS webpage Occupancy rating for bedrooms variable: Census 2021 for more information). In 2012, the Department for Communities and Local Government (DCLG, now Ministry of Housing, Communities and Local Government) issued guidance, Allocation of accommodation: Guidance for local housing authorities in England, which recommended that local authorities should use the Bedroom Standard when assessing whether households are overcrowded for the purposes of housing allocation.

Tenure

We used classification 4a of the Tenure of household classifications: Census 2021 which groups households into ‘Owned or shared ownership’ (which includes owned outright, owned with a mortgage, part-owned or shared ownership), ‘private rented or lives rent free’, and ‘social rented’. We shorten these category names to ‘owner-occupied’, ‘private rented’ and ‘social rented’ respectively throughout the report.

For more information about Census 2021 tenure, see Tenure of household variable: Census 2021.

Homeownership

The RIO analytical dataset does not currently enable us to fully identify households which also include refugees in our cohort who were not successfully linked to Census 2021 and/or other individuals in Census 2021 who are not part of the RIO cohorts. Living in an owner-occupied household does not necessarily mean that the RIO cohort member is a homeowner themselves; for example, they could be renting a room or living with a host family who own the accommodation. To gain a better understanding of home ownership among our refugee cohort and to ensure that households in our analysis were solely composed of individuals from the RIO cohort, we only looked at households where the RIO household size matched the Census 2021 household size. This approach means we can assume we have captured everyone in that household, which gives us the greatest confidence that a member of our RIO cohort owns the property.

In Section 2.4.4 on tenure by year of arrival, we determined the household cohort year by selecting the earliest year of asylum granted among household members for ARR households and the earliest year of resettlement among household members for RR households.

References

2011 Census, Office for National Statistics. (2014). 2011 Census analysis: Social and economic characteristics by length of residence of migrant populations in England and Wales. The National Archives.

Adsera, A., & Chiswick, B. R. (2007). Are there gender and country of origin differences in immigrant labor market outcomes across European destinations? Journal of Population Economics, 20(3), 495 to 526.

Charrad, M. M. (2011). Central and Local Patrimonialism: State-Building in Kin-Based Societies. Annals of the American Academy of Political and Social Science 636(1), 49 to 68.

Del Rey, A., & Parrado, E. (2012). The fertility of immigrant women: family dynamics, migration and timing of childbearing.

Dustmann, C., & Fabbri, F. (2003). Language proficiency and labour market performance of immigrants in the UK. The Economic Journal, 695 to 717.

Home Office. (2019). Indicators of Integration framework 2019 (3rd ed.). (Home Office Research Report 109). Retrieved from https://www.gov.uk/government/publications/home-office-indicators-of-integration-framework-2019.

Liebig, T., & Tronstad, K. (2018). Triple Disadvantage?: A first overview of the integration of refugee women. OECD Social, Employment and Migration Working Papers, No. 216. Paris: OECD Publishing doi:https://doi.org/10.1787/3f3a9612-en.

Sundqvist, P., & Sylvén, L. K. (2016). Extramural English in Teaching and Learning. From Theory and Research to Practice. Palgrave Macmillan London.  

1. Home Office asylum and resettlement datasets. ↩

2. Home Office Immigration System Statistics showed that, between 2016 and 2020, there were a total of 29,853 Family reunion visas issued to partners and children of those previously granted asylum or humanitarian protection in the UK. Each year within this period, the number of Family reunion visas issued ranged from 5,000 to 7,000, with the highest number of Family Reunion visas issued in the year ending December 2019 (7,083 visas granted). ↩

3. It has been documented in other countries, such as Spain, women’s fertility has proven to be low before migration and elevated shortly after arrival. The extent of disruption relates to family dynamics and spousal separation associated with the migration process (Del Rey & Parrado, 2012). Similarly, other countries, such as Sweden and Belgium, found that migrant women exhibit disproportionately elevated fertility within years of arrival. ↩

4. The counts of RR households for some years are very low, which means percentages should be interpreted with caution. ↩