1 Introduction

In London, 2.2 million people were in relative poverty according to the latest data for 2020/21-2022/23 – meaning that 24% of the capital’s residents lived in households with less than 60% of UK median income.[1] This was higher than the UK-wide rate of 22%. At that time, many labour market indicators in London were performing strongly, with low unemployment rates and high employment rates.

The contrast between the apparent health of the London labour market and the prevalence of poverty suggests that to better understand poverty in London and tackle it, we need to know how it relates to labour market participation. More generally, in-work poverty has risen over recent decades in the UK,[2] and has become an issue in many OECD economies,[3] so it is important to have up-to-date evidence for London specifically about this increasingly relevant phenomenon.

This paper looks at in-work poverty in London between 1996 and 2023. It provides an overview of the links between working-age poverty and employment participation at the individual and household levels in the capital. It focuses on the current standard definition of relative poverty, with people in poverty defined as those whose household net equivalised (i.e., adjusted for household size) income is below 60% of median income[4] after housing costs[5]. The in-work poverty rate is defined as the share of working-age people in employment living in a household in poverty.

Unless otherwise stated, the focus is on London’s working age population (16-64). The data used is from the Department of Work and Pensions’ Household Below Average Income (HBAI) dataset.[6] All the data shown in the graphs or text of this paper can be found in an accompanying Excel data file.

2 Key findings

• Poverty has changed. There are now more people in poverty in London who are in a working family than in a workless family. The reverse was true 30 years ago.
• Insecure forms of work such as part-time work and self-employment are linked to poverty.
• Ethnic minority workers and those with no educational qualifications are most likely to be working and in poverty.
• Working families with three or more children have seen their poverty rate increase sharply since the introduction of the two-child benefit cap on Universal Credit.
• Single parents have a high likelihood of in-work poverty, though the number of people in work and in poverty has also risen sharply among Londoners living as a couple.

3 Mapping in-work poverty in London

The overall share of working-age Londoners in poverty has fluctuated between 22% and 28%[7] since 1996(Figure 3.1). However, there has been a big shift in the composition of the population in poverty.

In 1996/97-1998/99, it could be broken down between 15% of working-age Londoners in poverty and in a workless family[8], and 9% in poverty and in a family with at least one person employed.

In 2020/21-2022/23, those figures were reversed: 14% of working-age Londoners were in poverty and in a family with at least one person employed, and 8% were in poverty and in a workless family.

Poverty in London has shifted from being a phenomenon primarily associated to family worklessness, to primarily affecting adults living in working families.

Figure 3.1. Poverty rate in London

The number of Londoners aged 16-64 who are employed and in poverty rose from around 320,000 in 1996/97-1998/99 to about 660,000 in 2020/21-2022/23 (Figure 3.2).

This corresponds to an in-work poverty rate in London of 15% in the latest data (Figure 3.2). This is up from 11% at the start of the period studied. However, both the count of workers in poverty, and the share of all workers that they represent, have been falling since a peak in the second half of the 2010s.[9]

Figure 3.2. In-work poverty in London

4 In-work poverty and type of employment

In-work poverty rates are much higher for part-time employees and the self-employed compared to full-time employees (Figure 4.1). Note that the data here refers to people’s main job.[10]

The rise of in-work poverty since 1996/97-1998/99 happened alongside both a rise in these forms of work in absolute numbers (Figure 4.2), and an increase in the poverty rates associated with them (Figure 4.1).

The poverty rate among full-time employees has been around 10% over the last decade in London. This is about one third of the poverty rate for part-time employees, which was 29% in 2020/21-2022/23, up from 21% in 1996/97-1998/99 (Figure 4.1).

The poverty rate for self-employed workers was 24% in the latest data. It has been broadly stable in the long run, with shorter-term fluctuations.

Figure 4.1. In-work poverty and type of employment

Figure 4.2. Type of employment

The biggest group of workers, by far, in London over the period studied, has been full-time employees. But since 1996, the number of full-time employees and the number of workers in all other forms of work – part-time employee, full and part-time self-employment – have risen in equal proportions (by 50% each).

Among these other forms of work, the rise was sharpest for part-time self-employment (+120%). Nonetheless, it remains the least common form of work in absolute numbers.  

So, employment growth over the past 30 years in London has been even between full-time employee work, which generally is associated with lower poverty rates, and the other forms of work, associated with higher poverty rates.

Academic studies show that mechanisms making these “other” forms of work more prone to poverty include shorter hours (in the case of part-time work) or instability in earnings (for self-employed work).[11] They might also suffer from lower hourly earnings, especially if more prevalent in lower-paid occupations.[12]^,[13]

But what also matters is the household context in which these forms of work are found (Figure 4.3).[14]

Figure 4.3. Family employment participation

Since 1996/97-1998/99, the number of working-age Londoners living in a workless family has fallen. They represented about 930,000 people in 2020/21-2022/23.

The number of those living in families where all adults are in full-time employee work has increased by 57%. There were about 2.6m working-age Londoners in such families in the latest data. However, there have also been large increases in forms of families relying in part or in full on more precarious forms of employment.

In particular, the number of working-age Londoners living in families for which part-time work is the only source of labour income has risen by 72%. They represented around 480,000 people in 2020/21-2022/23. The rise in Londoners living in families with at least one self-employed worker was 52%. Around 780,000 Londoners lived in such families in the latest data.

So, this suggests that the relevant labour market divide across families is becoming less about “workless versus working”, and more about the type of work found within families. We saw that job growth in London since 1997 was evenly fuelled by full-time employee work and all other forms of work. It also seems that the distribution of these different forms of jobs has been unequal across families.

5 In-work poverty rates by demographic characteristics

Table 5.1 shows the in-work poverty rates of Londoners by different socio-demographic characteristics.

The rate of in-work poverty varies by highest level of education. The in-work poverty rates of employed Londoners with a degree have been consistently lower than that of Londoners with qualifications below degree level, or without any qualifications.

Further, this gap has been increasing compared to 25 years ago. In the latest available data[15], workers with no qualifications were roughly three times more likely to be in poverty than workers with a degree. Workers with a qualification below degree level were roughly twice as likely to be in poverty than workers with a degree. At the start of the period, Table 5.1 shows that these differences were smaller. In-work poverty rates have risen across the three groups, but much more so among employed Londoners without a degree.

Table 5.1 also shows that workers from a White ethnic background in London have been less likely than workers from a Black ethnic background and workers from an Asian ethnic background to be in in-work poverty (data limitations prevent us from looking at other ethnic groups).

Again, in-work poverty rates have risen over time in all ethnic groups, but the rise seems to have been much bigger for ethnic minorities.

Educational level and ethnicity are the two characteristics most clearly associated with in-work poverty in our data. Gaps also exist with regards to other characteristics, but they are less clearcut.

A key finding in terms of age is that in-work poverty rates for employees rose across all age groups between the start of the 2000s and the middle of the last decade. Since then, there has been a decline or stagnation, but usually to levels that are still higher than 25 years ago. 

With regards to gender, male workers have been slightly more likely to be in poverty than female workers for most of the period studied. However, that gap seems to have closed very recently, and has always been rather narrow.

Workers with a disability have also had a higher rate of poverty than workers without a disability throughout the period, especially around the middle of the 2010s. That gap has since largely closed. It is difficult to interpret these fluctuations. It could reflect, for instance, changes in disability benefits. But smaller sample sizes when looking at subgroup levels may also induce some data reliability issues that mean caution is needed when interpreting sharp fluctuations.

Also note that we looked at each characteristic separately, but they may be linked. For instance, a certain demographic group may have higher levels of education. Our analyses here are descriptive and do not allow to disentangle these potential interactions.

Table 5.1. In-work poverty by demographic characteristics

6 Family structures

Family structures affect how many adults in the family might be available to work and how much. We concentrate on two key dimensions.

First, single-parenthood may induce a higher likelihood of in-work poverty, as time taken by caring responsibilities may restrict the ability of single parents to work full-time.[16]^,[17] They may also be in lower-paid occupations, often due to gender inequality, as most single parents are single mothers.[18]

Figure 6.1 shows that employees (the data was not reliable enough to include the self-employed) in London have a much higher poverty rate when they are single parents, and that this rate has risen over time.

This higher in-work poverty rate for single parents came as their employment participation increased over time and is tightly linked to part-time work.

In the latest data, 44% of employees who were single parents in London were in part-time work, compared to 18% for employees living as a couple with children, 14% for single employees without children, and 10% for employees living as a couple without children[19].

Single parents are clearly vulnerable but only have a limited role in explaining rising in-work poverty in London because they are a small group. In the latest data, 34% of employees who were single parents were in poverty, corresponding to just under 50,000 people.

By contrast, while the poverty rate of employees living as a couple with children was much lower at 15%, it corresponded to about 180,000 people. This is generally consistent with the fact that only a small minority of working-age Londoners live as single parents (240,000 in the latest data compared to 3.6m living as a couple).[20] In-work poverty has risen across different family structures.

Second, Figure 6.1 shows that in-work poverty rates are higher for single parents than single people without children, and for couples with children than couples without children. In turn, among working families, poverty rates are much higher when there are three children or more (Figure 6.2).

In the latest data, the poverty rate of Londoners in such large working families (40%) was more than double that of those living in a working family with one or two children (19%), and three times that of those living in families without children (13%).

Figure 6.1. In-work poverty by family type

Figure 6.2. Poverty and work by number of children

There has always been a gap, but it was much smaller before the introduction of the two-child Universal Credit limit and has widened since. The poverty risk of adults living in large working families increased sharply since 2017, while that of adults living in working families with fewer than three children fell or stayed broadly stable.

This is not causal evidence for the role of the two-child limit, because other things may have changed in the economy in 2017, and there might be socio-economic characteristics that affect both the likelihood that someone is in poverty and lives in a large family. But the temporal coincidence and the divergence in trends between family sizes makes it a likely factor, especially as research emphasises the role of this limit in increasing poverty.[21]

The two-child limit is relevant to working families because it is possible to receive Universal Credit while working. In our latest data, more than 70% of working-age Londoners living in a family with three children or more and receiving any kind of state benefits had at least one adult in work in the family.[22]

Further, our data shows that large families were more likely than smaller families to receive state benefits. They were also comparatively more likely to follow a model in which one but not all adults work, with a lower likelihood that all adults would be in work compared to other families.[23]

The combination of paid work and benefits is therefore important for larger families. But there are well-known issues in the British welfare system regarding the incentives for someone on low-pay and receiving benefits to increase their work earnings, for instance by working more, because the amount of means-tested benefits received tends to decrease as work earnings increase, leading to high effective marginal tax rates.[24]

This could be an issue for any family combining work and benefits. Families where the only source of employment is part-time employee work only represent around 10% of the families in which working-age Londoners live, overall. But they have been representing 40%-50% of those families in poverty despite receiving benefits and there being employment in the family.[25]

7 Limitations and scope for further research

While this paper provides an overview of the trends and drivers of in-work poverty in London, it has many limitations.

It focuses on relative poverty, i.e., how people at the bottom of the income distribution compare to the median, rather than on absolute deprivation, which may capture destitution better. It does not tackle material deprivation, nor does it analyse the depth of poverty – how deep below the poverty line people in poverty are. It may be that in-work poverty is a phenomenon more relevant to people below, but close to, the poverty line, while worklessness could still be more relevant to people at the very bottom of the income distribution.

The analysis also excludes the informal economy, which potentially employs vulnerable people, and it does not look at wealth. It also focuses on poverty at a given point in time. However, for some people and families, poverty might be temporary, while others may remain in poverty much longer. The extent to which Londoners employed and in poverty transition out of in-work poverty is also related to the wider research debate on whether precarious forms of work constitute a “stepping-stone” towards more stable employment, or a “dead-end”.

All these are the subjects of intense research agenda and may be issues to explore in the future.

---

[1] Economic Fairness – Population in Poverty – London Datastore.

[2] Why has in-work poverty risen in Britain?, Institute for Fiscal Studies (2019).

[3] Living wages in practice, OECD (2023).

[4] The median income is the income level such that 50% of people earn below that level, and 50% earn more. The median household disposable income in the UK in 2023 was £34,500 (ONS).

[5] UK Poverty 2024, Joseph Rowntree Foundation (2024).

[6] Available at: https://stat-xplore.dwp.gov.uk/webapi/jsf/login.xhtml. The latest data is for the years 2021/22 to 2022/23: HBAI requires pooling data from multiple years for analysis at regional, instead of national, level. Department for Work and Pensions (DWP), released 21 March 2024, GOV.UK website, statistical release, Households below average income: for financial years ending 1995 to 2023.

[7] Note that this rate, focusing on people aged 16-64, is different from the one mentioned in the introduction, which applied to the entire population of London.

[8] Note that throughout, we use the concept of “family”, defined by DWP as “a single adult or a married or cohabiting couple and any dependent children”. A family is a subset of the broader concept of households. For instance, multiple families can live in the same household (like in the case of a multi-generational household).

[9] Note that there could be many different reasons for that, for instance linked to earnings, benefits, or workers in poverty leaving London. It is beyond the scope of this note, focused on longer-run trends, to explore them.

[10] DWP guidance states that: “Several respondents have more than one job. The FRS [the source household survey behind HBAI] identifies which of these is their 'main job'. This is the job which the respondent says is the dominant activity. Where they cannot decide, the number of hours worked will determine which is the main job. This process of categorisation also applies to respondents who are employees in one job but self-employed in another; whilst the survey will capture information on both of these jobs, only one can be their main job.” This can induce some limitations to our ability to compute poverty by type of employment: for instance, if someone is in poverty and has part-time employee work as the main job and self-employment as the second job, it will count towards the part-time employee poverty statistics but not the self-employed poverty statistics.

[11] Horemans, J. (2018). Atypical employment and in-work poverty. Handbook on In-Work Poverty

[12] Manning, A. and B. Petrongolo. (2008). The Part-Time Pay Penalty for Women in Britain. The Review of Economics and Statistics.

[13] Horemans, J. and I. Marx. (2024). Poverty, inequality and material deprivation among the self-employed in Europe. Research Handbook on Self-Employment and Public Policy.

[14] Salverda, W. (2018). Low earnings and their drivers in relation to in-work poverty. Handbook on In-Work Poverty.

[15] Latest available data for all three groups is 18-19/20-21, with in-work poverty rates of 11%, 21% and 30% for employed Londoners with, respectively, a degree, a qualification with no degree, and no qualifications. See sheet 6 in accompanying data file.

[16] Taylor-Gooby, P. and T.P. Larsen (2004). The UK – A Test Case for the Liberal Welfare State? New Risks, New Welfare: The Transformation of the European Welfare State.

[17] Nieuwenhuis, R. and L.C. Maldonado. (2018). Single-parent families and in-work poverty. Handbook on In-Work Poverty.

[18] Single parents facts and figures, Gingerbread.

[19] See sheet 7 in accompanying data file.

[20] See sheet 7 in accompanying data file.

[21] For instance:

Patrick, R., Andersen, K., Reader, M., Reeves, A. and K. Stewart. (2023). Needs and entitlements: welfare reform and larger families.

Chzhen, Y. and J. Bradshaw. (2025). The two-child limit and child poverty in the United Kingdom, International Journal of Social Welfare.

[22] See sheet 9 in accompanying data file.

[23] See sheets 9 and 10 in accompanying data file.

[24] For instance:

Institute for Fiscal Studies.
(2024). Universal credit: incomes, incentives and the remaining roll-out.

Griffiths, R. and M. Wood. (2024). Cliff edges and precipitous inclines. IPPR Policy Brief.

[25] See sheets 5 and 11 in accompanying data file.