Skip to main content

Main navigation

Report
Public attitudes
Care
Housing
Savings, debt and assets
Social security
Work

Addressing key voters' economic insecurity is vital for all parties

People aged 35–59 are a pivotal swing voter group that all political parties need to appeal to. They're also the most economically insecure in Britain. What policies will throw them a lifeline?

Written by:
Nuffield Politics Research Centre
Date published:
Reading time:
50 minutes

1. Introduction

Using new, purposely designed survey data in 2024, we show how economic insecurity peaks in mid-life (between 35 and 59) and is much lower in later life, due to a combination of economic ‘stressors’: mortgage payments, debts, insufficient income, low savings and childcare responsibilities, even for those in full-time work. This group of mid-life insecure adults should be a focus for policy-makers. It is also of critical electoral importance because age is so important to vote choice. Adults in mid-life, especially those who feel economically insecure, are today’s ‘swing voters’, and they are already deserting Labour in greater numbers.

2. A focus on economic security is needed in addition to a focus on poverty

Economic insecurity is the subjective feeling about one’s broad economic circumstances. It reflects assessments of the balance of money going out and money coming in, financial commitments, the degree to which people have a safety net and expectations about what may happen in the future.

Having a low income does not always make someone economically insecure. Some low-income households and individuals have significant sources of wider financial security, such as assets and/or savings, or good economic prospects that can reduce feelings of economic insecurity. Likewise, a higher income does not always make someone feel economically secure. Many people on higher incomes – especially those in mid-life – lack wider sources of financial security and have much higher demands on their incomes. This means that using income as the only indicator of financial distress can be misleading, as it can conflate those who are on low incomes and experiencing poverty and hardship with those who are on low incomes and not experiencing financial hardship and, crucially, ignore those who are on higher incomes but are still experiencing financial hardship.

The economically insecure are a large section of society who do not have significant sources of financial security to relieve demands bearing on their income. Moreover, the distribution of these wider sources of security and their impact on feelings of economic insecurity varies over the life cycle, making age critical to understanding the distribution of perceived financial hardship and distress.

Even a focus on wealth, in the form of property wealth or financial wealth, will miss the broader experience of financial distress that comes from the balance of incomings and outgoings, prospects, and current and future risks. Financial distress can occur also for those who are not in a formally defined category of relative poverty. This is particularly true if we examine income and economic insecurity over the life cycle.

It is of considerable importance to know who feels economically insecure and why. Economic insecurity matters for the choices people make, because a person will modify their behaviour to take account of current and future risks.

The concept of economic insecurity has been associated with a range of negative life outcomes such as poor mental health (Rohde et al. 2016), poor physical health and obesity (Smith et al. 2009; de Witte et al. 2015; Rohde et al. 2017; Watson 2018), and drug and alcohol abuse (Glei and Weinstein, 2019). It will likely have many kinds of downstream effects on the broader economy. Those who experience economic insecurity have been found to decrease their private spending (Benito, 2006), postpone fertility decisions and alter their labour market decisions (Fiori et al. 2013; Modena et al. 2014; Mansour, 2018; Canto et al. 2020).

People who experience economic insecurity have distinct policy preferences, supporting, for example, greater levels of social policy assistance and political decisions that mitigate or buffer the relevant economic risks (Burgoon and Dekker, 2010; Hacker et al. 2013). Hacker and colleagues (2013) show that in the US, the magnitude of these associations between economic insecurity and policy attitudes rivals those for partisanship and ideology, and almost always exceeds the influence of education and household income.

Despite these insights into the importance of economic insecurity, there has been very little research on its measurement, distribution or implications in Great Britain, and most measures of the concept have been partial – focusing, for example, on specific sources of economic insecurity, such as job insecurity. This report reveals the insights we gain by concentrating specifically on measuring the experience and feelings of economic insecurity in Great Britain, through in-depth purposely designed surveys.

Our surveys about economic security

The NPRC (based at Nuffield College, University of Oxford) and the JRF designed 2 large, nationally representative surveys, each of around 8,000 people, with a third planned later in 2025, as part of an internet survey ‘panel’ in which as many of the same people as possible are interviewed in subsequent survey ‘waves’. Around 75% of the respondents in wave 1 also completed the survey in wave 2. This design will eventually allow us to examine the implications of entering and exiting different types of economic pain or gain and provide insights into the drivers and consequences of economic security and insecurity through analysis of the temporal ordering of experience and outcome. For now, we focus on the distributions of economic insecurity and some of the associated financial experiences, and show the stability of these patterns. This gives us insight into who is experiencing insecurity and why.

We focus on the life cycle because we know that income can be a limited indicator of financial security for certain age groups. For instance, young people with high expectations for the future (for example, students with high earning potential, or those with a wealthy family), or for retirees, for whom mortgage repayments will most likely have ended, savings and assets have accumulated in value, and family financial burdens are much lower. It has also been shown that life satisfaction dips in mid-life, and that stresses – including financial stresses – peak in mid-life: the so-called ‘happiness curve’ (see Blanchflower, 2020; Blanchflower and Graham, 2020). By focusing on the economic experiences of people across the age distribution, we can understand the financial stressors that explain why economic insecurity peaks in mid-life.

To understand economic security in an in-depth and original way, our surveys focus on the following types of measures, or ‘variables’ (in addition to demographics):

  • Feelings of insecurity: we use people’s subjective assessments of their economic security, for themselves and their family, as a key ‘outcome’ variable to understand what is associated with these feelings of economic insecurity.
  • Worries about different sources of security: these questions relate to feelings about job security, housing, savings, income and debts.
  • Economic experience: we ask comprehensive questions about respondents’ financial circumstances, including their savings, homeownership, employment status, occupation (and that of their partner, if relevant), income, income sources, debts, ability to pay for an emergency expense, job security (for them and their partner) and use of any benefits.
  • Economic evaluations: we ask a series of questions about evaluations of aspects of the national economy, including potential job threats.
  • Relevant attitudes: we ask whether respondents think they could rely on the benefit system if they needed it and ask about their attitudes to some policy outcomes such as redistribution, and their attitudes towards other social groups.
  • Political choices: we include vote intention and likelihood of voting, and link data to respondents also interviewed as part of the British Election Study internet panel, providing a wider set of political variables for respondents interviewed in both.

Our key economic insecurity measure is the response to the question:

How worried are you about your and your family’s economic security?2

We code those who gave a score of

  • 7–10 as ‘insecure’,
  • 0–3 as ‘secure’,
  • and those in the middle (and ‘don’t knows’)3 as neither secure nor insecure.

This is a validated item shown to vary in response to information about a respondent’s income, savings, homeownership and occupational status in the ways we would expect, as well as being strongly associated with vote intention (see Green and de Geus, 2022).

Overall, 35% of our respondents report feeling insecure (they score 7–10 on the economic insecurity question). This suggests that around 18.5 million adults in Great Britain feel economically insecure,4 equivalent to around 35% of the eligible British voting population. The first illustration of the importance of understanding economic insecurity comes from comparing its distribution to household incomes.

Figure 1a for March 2024 and 1b for October 2024 show the pattern of reported feelings of economic insecurity by gross household income, equivalised for household size, divided by deciles. Figure 1 demonstrates – via box-and-whisker plots – the extent to which each decile of income tends to feel more or less economically insecure. The ‘boxes’ indicate the location of around 50% of the sample at that decile of income, with the thick black horizontal line indicating the median for that decile. That is, the point at which half of the respondents in that decile have a higher level of insecurity, and half have a lower level. ‘Don’t know’ responses to the insecurity and income questions of our survey were excluded.

Figure 1 confirms the importance of understanding economic insecurity in addition to income. Economic insecurity is related to income, but income is only part of the story as it affects feelings of security or insecurity primarily at the extremes of the income distribution. The pattern is consistent across the 2 time periods: those on higher incomes, and especially those at the top of the income distribution, are generally more secure, whereas those on lower incomes, at the bottom of the distribution, are more insecure. A large proportion of the income distribution, however, is neither insecure nor secure.

In October 2024, over half (55.2%) of those in the bottom 2 deciles of earners had an insecurity score of 7 or more (our threshold for being ‘economically insecure’ in this report) once ‘don’t knows’ are removed. This figure more than halves (to 23.2%) among those in the top 2 deciles. However, the fact that almost 1 in 4 high earners still feel economically insecure, and a large proportion of those on the lowest incomes do not feel insecure, underlines the importance of focusing on economic insecurity as a wider indicator of financial hardship and distress than income alone. The two values in Figure 1 of gross equivalised household income decile and reported feelings of economic insecurity have a correlation of less than 0.3.

3. Economic insecurity is highest in mid-life

The distribution of economic insecurity over the life cycle stands in contrast to the distribution of being on a low income.

We can see the divergence in Figures 2a and 2b, which show the patterns of reported economic insecurity by age in our NPRC/JRF surveys in March 2024 and October 2024 to demonstrate their stability, alongside the percentage having a ‘low’ household income (defined as a gross equivalised household income equal to less than 60% of the sample median). These graphs plot the data using smoothing and showing the statistical confidence intervals around the estimates, in the shading around each line. The differences are significant where the shading diverges from the bold line in the comparison group, and the wider confidence intervals represent samples for whom the data is sparser, which happens above the age of around 75.

Both graphs show a sharp divergence in the average age at which someone experiences peak economic insecurity and the average age of being in a low-earning household. Namely, it is in mid-life, approximately between a person’s mid-30s and late-50s, when the highest levels of economic insecurity are reported. This is when the same respondents are the least likely to be on the lowest household incomes.

Figure 2c then compares average economic insecurity by age across both waves of our study (March and October) to age-related rates of relative poverty and material hardship, taken from the UK Government’s Department for Work and Pensions’ ‘Households Below Average Income’ (HBAI) dataset for 2022/23 (the most recent year for which data is available). Poverty is indicated by having an equivalised net household income that is less than 60% of the national median. Material deprivation is not defined by income but by an inability to afford or access given items or activities deemed necessary for an acceptable standard of living (for instance, a damp-free and adequately warmed home, regular dental appointments, and reliable access to the internet).5 This measure is only available for those of working age.

While the ‘official’ income-based relative poverty indicator shows – like our own sample of low-income respondents – no pronounced increase in mid-life, the material deprivation measures do, albeit one that is not as pronounced as the peak in economic insecurity. This suggests that the mid-life peak in economic insecurity could be rooted in real economic experience. However, the peak in the material deprivation scores is by no means as clear; we need to understand a wider range of economic experiences that account for the age-based distribution in economic insecurity.

The peak in economic insecurity in mid-life is not a recent phenomenon driven by the timing of our surveys. The same question about subjective economic insecurity has been asked by us and colleagues in the British Election Study (BES) Internet Panel since 2018; this shows the same pattern across the life cycle in each of the 6 times it has been asked. This can be seen in Figure 3. Note that the final 2 graphs show the distributions using the NPRC/JRF dataset and they also compare the distributions for 2 questions (one asking about the respondent and their family, the other, only the respondent) to show that the question wording does not alter the general pattern.

There is a consistent problem of economic insecurity in Britain in mid-life.

To begin to understand these patterns, consider those in the youngest age groups in the above graphs. While these individuals may be going without some of the essentials and may be in relative poverty as suggested in Figure 2c, they report below-average levels of economic insecurity and noticeably less economic security than the group immediately older than them, starting from the mid-30s. Younger adults may be considering their future prospects optimistically, their ability to borrow from family or being directly supported by family, either at the time of the survey or in the future.

Of our survey respondents under the age of 35, 32% report still living rent-free in a family or friend's home or in university accommodation. Of those under 35, 24.4% report ‘parents or grandparents’ as a current source of financial support. This rises to 27.1% if ‘other family members’ (non-spouse) are included. Of those under 35, 26.4% report being likely to inherit a substantial amount of money or property in the future, either ‘probably within the next 10 years’ (6%) or ‘probably over 10 years from now’ (20.4%).

Now consider those in the older age groups, after their mid-50s, where the feeling of economic insecurity begins to decline very markedly. These people are significantly more likely to have lower incomes (as we saw in Figure 2) and yet they report significantly higher levels of economic security than people in their mid-life.

Figure 4 uses the October 2024 NPRC/JRF Economic Insecurity survey to plot the percentage of those who are subject to 5 economic ‘stressors’, by age. These economic stressors are: having low savings (being in the bottom quintile, calculated as equivalised by household size); having unsecured or ‘bad’ debts (for example, unsecured loans, credit card or payday and overdraft debt, but excluding student loans and mortgages); currently paying off a mortgage; having childcare responsibilities; and being unable to come up with £300 for an emergency expense.

Figure 4a shows that the peak emerges in each of these financial experiences at slightly different age points, but each can help us understand the mid-life peak in economic insecurity through the addition of a range of financial difficulties which are all at their highest levels in mid-life. Older people are – on average – far more likely to have savings, lower levels of mortgage or other debts, and lower financial responsibilities for family members. As a result of this, their disposable income or ability to draw on savings (for example, if they needed to come up with £300 for an emergency expense) is substantially greater. Figure 4 also suggests that younger adults’ higher economic security is due to the absence of outgoings such as mortgages, other types of debt, and caring responsibilities for children.

Figure 4 shows the importance of mortgage debt, childcare duties and unsecured debts in particular. Importantly, the pattern we currently identify in mid-life insecurity and the above economic experiences is happening for people who are in full-time secure work or who have a partner in full-time secure work, as we see in Figure 4b, as well as for those whose household lacks a secure full-time worker (4c).6

We note, in addition to the patterns shown in Figure 4, that substantial proportions of those in mid-life are not only giving care to children aged under-18, but also to adult children, and older relatives. Of those aged 35–59, 34% report having to provide financial/care assistance to a child aged under 18, 13% report having to provide this support to an adult child and 17% report having to provide for a parent or grandparent. Overall, 49.2% report having to provide assistance to at least one of these three groups. Moreover, 8.4% of the midlife age group is giving care to both younger and older relatives; a problem known to befall those in a ‘sandwich generation’ of having children and older relatives to care for simultaneously (Miller 1981; Vlachantoni et al. 2020).

The above graph demonstrates the slight ambiguity in a hard cut-off age point for when economic difficulties peak in mid-life, since some measures show an increase earlier on (repaying a mortgage), whereas having unsecured debt is higher at a slightly later average age, and the experience of having low savings tends to persist to a slightly later age. Nevertheless, it is clear that these experiences accumulate during mid-life and are generally lower among the young and among those over 60.

These patterns reflect the age distribution of these economic experiences at the time of our October 2024 survey. They are consistent in our March 2024 survey and are likely to be consistent since 2018 (given Figure 3 above), but they of course differ for some people and could alter if we were able to go back in time (or could go forward in time). For example, the decision to have children has become something that adults do later in life, on average, compared with earlier generations, and this delays the settling down stage of purchasing a home and taking on a mortgage. Younger adults are also now less able to afford to purchase a home until a later stage in life than earlier generations were able to (Broome et al. 2023). 

It is, furthermore, possible that today’s mid-life adults feel more insecure than earlier generations did at the same age because they lack prospects of a good retirement in the future (as we show later in this report) and because earlier generations may have experienced an earlier peak in insecurity, when they bought homes and started families at earlier ages. There is no guarantee, then, that this distribution will look the same for future generations. It could look worse if current mid-life adults fail to achieve the comfortable retirements many pensioners enjoy today because they were unable to accrue savings or similarly comfortable pensions. Similarly, today’s younger generations may accrue greater mortgage debts later in life, have lower levels of savings and have to take out higher amounts of debt for longer as a result. Alternatively, of course, larger numbers may simply never manage to get on the housing ladder in the first place, and could remain concentrated in the insecure rental market into mid-life and beyond.

There are important variations in the extent of mid-life economic insecurity across different demographic sub-groups.

Figure 5 shows the distribution of the percentages that feel economically insecure by age and by gender, educational attainment level, homeownership, having a long-term serious illness or disability,7 being single or living with a partner or spouse, and having childcare responsibilities. As before, we plot the smoothed estimates. The shaded areas around each line represent the statistical confidence intervals. Differences are statistically significant where the shading diverges from the bold line in the other group. To aid interpretation of their relative importance, the dotted lines in these graphs display the proportion of one of the groups in the whole sample.

Figure 5 shows that women are more likely than men to report feeling economically insecure from around the age of 40 and are significantly more likely to feel economically insecure in old age. This does not mean that future generations of older women will do so: this may reflect the experiences of current women in retirement, for example, who were less likely to participate in the labour market in highly paid jobs during their working age, and the experiences of women in mid-life who are less likely to have low savings and be unable to afford an emergency expense of £300.

The differences between men and women on these economic indicators can be seen in Figure 6, which shows the different economic experiences of men and women; specifically, the higher likelihood among women of having low savings and an inability to pay an emergency expense.

Returning to Figure 5, those with a degree are more likely to report feeling secure, though the differences are quite small. Graduates tend to reach peak mid-life insecurity at slightly older ages (around 45–50) than non-graduates, who reach peak mid-life insecurity at around 35–40. This may reflect the delay in having children that typically happens among graduates, and the accumulation of other financial responsibilities that also typically start later.

Homeownership shows the largest gap in reported economic insecurity, along with the gap between those with long-term illness and/or disabilities. Focusing first on homeownership, note the dotted line showing the proportion in the sample who do not own their own home, which drops from the youngest ages onwards. Among the remaining minority who are not homeowners, economic insecurity peaks at pre-retirement age, but is very high from around the age of 40. The pre-retirement peak may reflect the awareness of one’s income prospects decreasing, but it is notable that older renters, post-retirement, still exhibit a marked decline in feelings of economic insecurity.

People with long-term illnesses and/or disabilities are much more likely to feel economically insecure. This likely arises through being unable to work or to progress in work as they would like, as well as the additional costs faced in adjusting to their conditions. Among the disabled or long-term sick of working age (18–65), for example, we note that only 39% are in full-time work (compared to 62% in the remaining sample). Of those disabled individuals who are in work, 19% think their job prospects will get worse in the next 12 months (compared to 11% of non-disabled workers). Furthermore, 27% do not have any savings and 46% are not homeowners (compared to 15% and 35% of non-disabled adults, respectively). This group also reports higher insecurity for a prolonged age span: between 35 and 60. Part of this higher rate of economic insecurity could arise from some long-term ill or disabled respondents having difficulties with aspects of mental health that make them more insecure in general, but the economic patterns we see in the data suggest this is rooted in economic experience too.

Turning to those who are in cohabiting relationships (either married or with a partner), who we can assume share economic resources to some degree, the peak in economic insecurity starts at slightly younger years, around 40, but insecurity is higher for those who are single and peaks for a prolonged period in comparison to other groups except those with long-term illnesses and/or disabilities; between 40 and 60.

Finally, having children – an additional financial burden – also increases feelings of economic insecurity, though the differences between those with and without childcare responsibilities are not large. There may be some ‘selection effects’ here because a decision to have a family is also in part an outcome of having the economic means to support one. The dotted graph in Figure 5f shows how having childcare responsibilities peaks in mid-life, and so the proportions who have children and feel economically insecure will be greater in this age range.

We can examine the relationships between demographics, economic experiences and economic insecurity using statistical models. This allows us to assess the relative strength of statistical association between different demographics, economic experiences and feelings of economic insecurity. The following plots show the statistical ‘coefficients’ (the strength of statistical association) between the demographics mentioned above, and the economic circumstances measured in the survey, and repeat the model using the March 2024 data and the October 2024 data, to examine the stability of these relationships over time.

Each ‘whisker plot’ shows the association of each variable with economic insecurity if it is analysed on its own (shown with a circle in the middle of a whisker) and then the association when other variables are also taken into account (shown with a diamond). The latter provides an estimate of the additional statistical relationship to feelings of economic insecurity, net of all the additional factors included. Black whiskers denote a significant relationship (statistically) and red whiskers denote a non-significant relationship.

Figure 7 shows that the demographics and economic factors associated with economic insecurity are stable over the 2 time periods, with very minor exceptions (the relationship of being in a cohabiting partnership is stronger in October 2024 than in March 2024, and the effects of education are only significant in October, without controls). The whiskers to the right of the vertical line at 0 indicate factors that increase feelings of economic insecurity and are statistically significant if the ‘whisker’ bar does not cross 0. The factors to the left of the vertical line at 0 indicate factors that decrease feelings of economic insecurity (the higher they are, the lower someone’s score on the economic insecurity scale). The estimates are transformed so they equate to percentage point increases or decreases, shown on the horizontal x-axis.

The influences we identified above: the life-stage a person is at, their age, gender, disability, education, owning a home and having childcare responsibilities, are all in the expected direction and all have a statistically significant effect. Having a higher income, having savings, and being in a cohabiting couple each contributes to someone feeling more economically secure, and debts and not being in secure full-time work contribute significantly to their feeling more insecure. Being in secure full-time work is an especially strong predictor, but all other factors contribute similarly. This can be concluded because these patterns remain the case if we include the other different variables as ‘controls’, so we can say, for example, that owning a home outright tends to decrease feelings of economic insecurity by around 10 percentage points, even when we take into account all the other demographics and economic factors described above (including a person’s income, savings, debts, whether they have childcare duties and likely share financial resources).

Having a disability or long-term illness also increases feelings of economic insecurity by around 10 percentage points, even accounting for income, homeownership, savings and so on. We note that the ‘classic’ indicators of economic hardship, namely income, property (homeownership) and financial wealth (the balance of income and savings), are certainly not the whole story. Indeed, income has, if anything, a slightly smaller effect in comparison to other factors, and outgoings (such as having children) and financial burden sharing (in the form of cohabitation) are equally important.

The statistical association between economic insecurity and life-stage, which we code as those between 18 and 34, those between 35 and 59, and those above 60, remains significant controlling for the other demographics and economic experiences that vary with the life-stage, though not for the comparison of the mid-life and over-60s groups, where the distinction is accounted for by the other variables.

One difficulty with assessing economic insecurity through economic stressors – through people’s reported levels of savings, debts, their home ownership and so on – is that some people will accept or feel sanguine about their levels of financial precarity and will not find them as much a source of ‘stress’ as other people may do. After all, these levels are a consequence to some degree of decisions made by the individual about what level of debt, savings or mortgage they can manage, comfortably or otherwise. Other people will be more worried about the associated financial risks. We would expect such worries to explain reported feelings of economic insecurity. We therefore focus next on understanding people’s feelings about their economic circumstances to help us interpret the experience of economic insecurity in Great Britain in 2024.

The meaning of economic insecurity: financial worries

The patterns we report above suggest that economic insecurity arises through the balance of how good a person’s ‘buffers’ are, or their economic ‘insurance’ against hard times, and the demands on their finances, in the form of mortgage debts and outgoings, which make it all the more important that a person can insure themselves effectively.

We can see that having a partner to share resources and risks, having a home, good long-term prospects (through educational attainment) and having savings as sources of financial protection are sources of economic security. Being ill or having a disability, not owning a home and having debts and low (or no) savings are all associated with a greater risk of economic insecurity. Having children increases those risks.

The economic circumstances we report above, across the lifespan, are associated with the accumulation of greater economic psychological burdens in mid-life. This can be seen using different questions in the surveys, which asked about whether respondents were worried about their income, housing tenure, job prospects (if in work or unemployed), their level of debt (if they had debts) and whether they were satisfied with their level of savings.

Specifically, we asked respondents: ‘How much do you agree or disagree with the following statements?’ (1–5, strongly agree – strongly disagree, or ‘don’t know’):

  • I am very worried about my level of income in the next 12 months
  • I am very worried that I might have to move out of my current home in the next 12 months, not by personal choice
  • I am very worried about keeping my job (If employed)
  • I am very worried about my level of debt

For each of the above, we plot – in Figure 8 – the proportion who agreed or agreed strongly that they were ‘very worried’ in contrast to other responses and the respondents for whom the question did not apply. Slightly different were 2 questions about savings. Respondents were asked to agree or disagree with the following 2 statements:

  • I am generally satisfied with the amount of savings I have
  • I feel very confident that I will have a decent standard of living in old age

These 2 items were reverse coded so that the percentage of the sample who ‘disagreed’ with them was classified as having saving worries.

We can see in Figure 8 that all the measured worries tend to decline among the oldest respondents, while employment worries, debt, savings satisfaction and ‘long-term worries’ (being confident of having a decent standard of living in old age) all peak in mid-life. Income worries and housing worries tend to be high among young people and those in mid-life, but a considerable amount of psychological distress (measured as reported worries) peaks in mid-life through the accumulation of different financial worries. This is consistent with the patterns we demonstrated earlier for feelings of economic insecurity. These worries also suggest that part of the explanation for mid-life feelings of insecurity arises because people think they might not be able to improve on their problems; they expect to experience financial distress in the future too. For example, current satisfaction with the amount of one’s savings is lowest in mid-life, but so is the assessment that the respondent will not have a decent standard of living in old age.

If we take an average score across all the ‘worry’ responses from those in Figure 8a–f, we see that this peaks in mid-life, at around 2 worries per person, as shown in Figure 9a. Younger respondents have a greater average score than older respondents, and those in mid-life have the highest scores. This does not mean that each of these worries has equal importance, but it is a way of depicting how worries accumulate in mid-life, and – for current generations of older people – significantly decline in later life.

Figure 9b displays how many ‘worries’ a person reported agreeing with: 1, 2, 3, 4 or more, or none of the questions asking how worried a respondent was about their income, housing costs, job prospects, debts, savings or living standards in old age. It shows that the group with the greatest likelihood of reporting ‘no’ worries were those respondents over the age of 60. This blue line dips in mid-life, as we would expect. The proportion with ‘one’ or more worries, in contrast, rises for those in mid-life, with numbers generally peaking around 40. Nevertheless, the largest proportion of over-60s reported no worries whatsoever in relation to the topics we covered (income, housing, job prospects, debts and savings).

Each of these worries is associated, statistically, with feelings of economic insecurity. In Figure 10 we use a coefficient plot to demonstrate the statistical relationships between each ‘worry’ and feelings of economic insecurity. We find that they are broadly of equal ‘weighting’ when assessed this way, with worries about income being slightly more strongly associated with feelings of economic insecurity. We also show the proportions agreeing or disagreeing with the ‘worry’ (or satisfaction) questions on the y-axis. While each of these worries is strongly associated with feelings of economic insecurity, those who are worried about their income (36%) and dissatisfied with their levels of savings (43%) represent the largest proportions in the sample overall.

Of course, in reality, we know that people do not think about these worries in isolation. Most crucially, as we demonstrated in Figure 9b, those in mid-life will be more likely to have a combination of different worries – likely compounding their levels of insecurity – whereas those of retirement age will more likely have none or only one of these types of ‘worry’.

4. The electoral risks of mid-life economic insecurity to political parties

In the final section of this report, we consider the electoral implications of our focus on those in mid-life, especially those in mid-life who feel economically insecure.

Age has become a critical dividing line in British elections, particularly since the Brexit referendum in 2016 (Fieldhouse et al. 2020). The Labour Party is seen much more as a party representing the young (see Grant, Green and Evans, 2022), and this has been strongly mirrored in voting behaviour, with older voters coalescing to parties on the right and younger voters supporting parties on the left. As reported by Griffiths and colleagues (2025), the most recent UK general election saw the age divide still very firmly in evidence, especially separating younger voters into a left-liberal ‘bloc’ (including Labour, the Liberal Democrats, the Greens and nationalist parties), and older voters into a right-socially conservative ‘bloc’ (including the Conservatives and Reform UK).

This suggests that those in mid-life may now be Britain’s current ‘floating voters’, less anchored by either of the 2 ideological blocs and feeling less well-represented by them. Given our evidence of economic insecurity peaking in mid-life, and of the electoral relevance of feelings of economic insecurity, any such greater tendency to be politically moveable in mid-life would suggest that mid-life economic insecurity could be an extremely important focus for any political party hoping to retain or increase its support.

We can show this electoral potential among those in mid-life, and those in mid-life who feel economically insecure, in 4 ways: through the patterns of support for all political parties by age, post-July 2024, the volatility in support over the age distribution since July 2024, the degree to which people are politically ‘undecided’ across the age distribution, and through Labour’s vote losses by October 2024 compared to July 2024.

Party support by age

British politics exhibited an especially high degree of fragmentation in the form of greater proportions of votes for smaller parties in the 2024 general election, which followed a trend towards greater fragmentation that had continued up to 2015 (Miori and Green 2025; Griffiths et al. 2025). Voters showed a strong age-alignment on the left-liberal side of the party system, supported by younger voters on average, and on the right-socially conservative side of the party system, supported by older voters. It is therefore relevant to consider whether there is less alignment by age in mid-life, and whether this differs according to feelings of economic insecurity.

In Figure 11, we show vote intention for the left-liberal and right-socially conservative sides of the current party choices in British politics by age. This shows how younger respondents remain (in October 2024) more likely to support left-liberal parties – Labour, the Green Party and the Liberal Democrats – and older respondents are more likely to support right-socially conservative parties – Reform UK and the Conservatives. It also shows how, due to the strong relationship of party support and age, the likelihood of voting Labour, Conservative or Reform UK crosses in the age distribution at the end of our mid-life classification: as voters approach age 60. This, then, is the more competitive part of the age distribution for these 3 political parties, when people are equally likely to vote in 3 different ways and between the left-liberal and right-socially conservative party blocs, as opposed to being strongly left-liberal or right-conservative.

The age gradient is particularly pronounced for the Conservatives, and it is notable that Reform UK’s support shows its peak just past our mid-life definition; at around 60 years of age. The Conservatives are retaining, as of October 2024, their support among the oldest voters in the British electorate, but the anchoring nature of this age effect weakens among those around 60, where the Conservatives are losing support to Reform UK: the largest proportion of Reform UK’s voters were from the Conservative Party in July 2024 (Miori and Green, 2025).

Figure 12 reveals the differences in the age-vote intention relationship among those who are economically secure (Figure 12a) and economically insecure (Figure 12b). This comparison shows some important differences, which we note, again, are happening at a very early part of the electoral cycle (by October 2024).

Among the secure, the Labour and Conservative age relationship remains almost unchanged, but support for Reform UK is slightly lower than for the whole sample (shown in Figure 11). Among the insecure, however, we can see that Reform UK’s support is considerably higher, and especially increases among those in mid-life to peak around 60 years of age. Support for the Green Party is also higher, particularly among the youngest in the sample. While the age gradient for Conservative support is unchanged among the insecure, there is greater contestation between Labour and Reform UK in the mid-life range of the sample, crossing at around 50–60 years of age. This provides some evidence that there is greater electoral competition between the left-liberal and right-socially conservative party blocs in the mid-life part of the age distribution, particularly among those who report feeling economically insecure.

Electoral volatility and ‘undecided’ voters, by age and economic insecurity

Traditionally, we might expect younger voters to be especially volatile in their political support, due to their lower likelihood of forming partisan attachments which tend to increase political loyalty (Fieldhouse et al. 2020). We would also expect younger people to be less decided about their vote choice on average; they are still going through a period of political socialisation, finding out about politics and deciding which party may best represent their interests.

However, as we noted above, this assumption changes with the perception of greater representation of younger generations by left-liberal parties, at least among those younger voters who have made a political choice. If there is now greater political contestation among the mid-life group, and if this increases among those who feel economically insecure, we would expect those in mid-life to be more undecided between the parties, but for the insecure group to be exhibiting the greatest likelihood of switching their political support in response to their economic insecurity. These individuals have a greater policy-based political grievance.

We can explore these patterns by age and insecurity by examining the proportions who have switched their support between any political party, between July 2024 and October 2024, and the proportion who say they ‘don’t know’ who they would vote for in a future general election, which we display in Figure 13. In each case, the horizontal gold (switching) and grey lines (don’t know) show the average across the whole age distribution, for each group and outcome.

Figure 13 shows that electoral volatility – in the form of switching party support since the July 2024 general election – is higher among those in mid-life, but only if those respondents also report feeling economically insecure. Figure 13 also shows that the average proportion of those saying ‘don’t know’ is highest among those who are insecure, compared to the secure, and that this political uncertainty peaks in mid-life. 

The parallel nature of these patterns with the evidence we have shown throughout this report is striking. These mid-life patterns, at least in the early stage of the Labour Government’s term in office, suggest there may be a new ‘swing voter’ group in British politics that has emerged due to the political anchoring effect of youth and old age in recent British elections and, crucially, is amplified by economic insecurity.

We emphasise that these patterns could change. The fact that the political parties are currently seen as representing the young (on the left-liberal side) and older generations (on the right-socially conservative side) arises because of the nature of political competition, through ideology, image, rhetoric and so on. Were the political parties to significantly change the groups to which they are appealing – and to which they are seen to appeal – we would not necessarily expect the greatest competition to take place among voters in the mid-life part of the age distribution. However, for the reasons given throughout this report, we do expect the age distribution of economic insecurity to continue, and we consider the experience of economic insecurity to be very important at the ballot box. We provide evidence for this next.

Labour’s early vote losses among mid-life economically insecure voters

Labour had already lost support by October 2024, and subsequent opinion polls show that this has dropped further, up to the time of writing. In total, 40% of Labour’s July 2024 voters had moved by October 2024 to either ‘undecided’ or a vote intention for one of the other parties, according to our data. By usual standards, this is a very short period in which to observe defections for incoming governments, but the large amount of vote losses allows us to draw early conclusions about systematic patterns and explanations that might be expected to continue.

Furthermore, by analysing change over time, we can observe change within the same individuals who voted Labour in July but would no longer intend to vote Labour if there were ‘a general election tomorrow’, which improves the confidence we draw from observing change, as opposed to just correlation. The above evidence of greater switching and greater uncertainty among those in mid-life and among those economically insecure respondents in mid-life suggest that this is likely where Labour’s greater electoral losses have already been.

The first 100 days of the Labour Government was a period when the Government was particularly pessimistic about the UK’s economic prospects, the potential for spending on public services and the prospect of a long horizon before improvements would be noticed. The likelihood of increased measures (ostensibly taxation) was claimed to be necessary to meet the size of the deficit left by the outgoing Conservative Government, which was argued to have been larger than expected. The period included measures that elicited considerable attention in the media, such as the withdrawal of the winter fuel allowance.

As expected, Labour lost votes between July and October 2024 at a greater rate among those who felt economically insecure. The ‘flow of vote’ graphs displayed in Figure 14 show that while economically insecure voters made up a smaller proportion of Labour’s July 2024 vote, compared to those who felt secure (the starting proportion is smaller on the right-hand-side graph, compared to the left-hand-side graph), the rate of ‘defecting’ from a Labour vote was higher among the insecure: 45.6% compared to 30.8%, including moving from Labour to other parties, undecided and ‘will not vote’.

The Labour Party had already lost 39% of its July 2024 voters by October 2024 to other parties and ‘undecided’, and feelings of economic insecurity are important for understanding this. Whereas the party only lost 3 in 10 (31%) of its economically ‘secure’ supporters, Labour lost almost half (46%) of its more numerous economically ‘insecure’ voters. Put differently, had Labour’s ‘insecure’ supporters defected at the same rate as its ‘secure’ ones, and everything else remained the same, Labour’s total losses would have been 5 percentage points (13%) lower in total. Had all of Labour’s non-secure July 2024 voters (including those who were neither secure nor insecure) defected at only the rate of secure supporters, total losses would have been around 8 percentage points (21%) lower.

Labour’s July 2024 vote was primarily a vote among progressive, liberal voters who were younger and with higher levels of educational attainment (see Griffiths et al. 2025), which is reflected in their reported vote losses since. Figure 14b shows that there were proportionately more defections from a Labour vote to the Liberal Democrat or Green Party vote intention compared to those to Reform UK or the Conservatives, as of October 2024, though note that the small sample sizes mean we cannot discriminate in detail.

Figure 15 shows the likelihood of switching away from Labour in this period by age; the left-hand-side graph shows the pattern by age, and the right-hand-side graph shows the pattern by economic insecurity and age. Recall that where the confidence intervals overlap the bold line in the comparison group, the patterns become statistically indistinguishable. This means that we need to focus on the differences between the shaded areas and the bold lines.

Figure 15a (defections by age) shows that defections were more likely amongst respondents aged 60 to 70. Labour had fewer over-70-year-olds among its July 2024 vote (due to its younger vote base in the election), and so the confidence intervals become wider and less reliable at the older ages. Figure 15b shows that feelings of economic insecurity are a discriminating factor for those who defect and those who do not – again, crucially, among those in mid-life. The red line shows the likelihood of defecting from Labour among the insecure, by age, and the green line shows the much lower likelihood of defecting among the secure, again by age. These lines diverge at the start of the age period we have focused on throughout this report: those in ‘mid-life’.

The importance of economic insecurity to defecting from the Labour Government is confirmed in a full statistical model (Figure 16) that takes into account a wide variety of other potentially competing factors. Here again we use a whisker plot to show the statistical effects, and add additional variables to the statistical model to see if the relationship between feelings of economic insecurity and vote defections can be explained by other factors. The relationship holds up very well. The results are reported from 4 models in turn: the first (denoted with a cross) is the bivariate relationship; just the association between feelings of economic insecurity and Labour defection. The second (denoted with a diamond) reports the relationship by adding demographic controls (age, education, gender, and household income). The third (with a triangle) reports the relationship controlling for respondents’ personal assessment of the national economy. The fourth (with a circle) adds respondents’ preferences regarding immigration and the redistribution of incomes from the well-off to the less well-off. See the note underneath Figure 16 for details.

Figure 16 confirms that an increase in the level of economic insecurity by a one standard deviation change on the scale equates to around a 5-percentage point greater likelihood of changing one’s vote intention from voting for Labour in July 2024 to no longer doing so by October 2024. This switching remains the case irrespective of the addition of statistical controls.

Moreover, we can compare the relationship between feelings of economic insecurity and voting against the Labour Government between July 2024 and October 2024 to the relationship between economic insecurity and Conservative vote losses between the December 2019 and July 2024 general elections. This is presented in Figure 17. Here, the addition of control variables reduces the size of the relationship a little, but not significantly so, and the results confirm the importance of economic insecurity to explaining Conservative defections between 2019 and 2024, as they do so far for defections from Labour since they took office.

The parallel patterns indicate that the vote losses that Labour has incurred are best understood as anti-incumbent vote losses, rather than those driven, for example, by particular types of ideological persuasion. Economic insecurity helps us understand the demise of the Conservative Government, and – at an early stage – helps us understand the electoral risks being incurred by the incoming Labour Government.

It is likely that economic insecurity is as, if not more, important for understanding Labour’s vote losses at an early stage compared to a series of other political issues. We asked our respondents for their evaluation of how well Labour was handling ‘the economic security of households’, ‘the economy’ (in general), reducing poverty, immigration, the NHS, and ‘reducing carbon emissions’, a set of issues designed to be broad for comparison, each measured on 5-point scales where 1 = very badly and 5 = very well.

The statistical association between each of these handling scores and whether a respondent voted Labour in July 2024 but no longer intended to vote Labour in October 2024 can be seen in Figure 18. Each of these is a simple bivariate association, not controlling for other factors, so these coefficients should not be seen as isolating the unique effect or relationship of any one of these handling evaluations net of the others. Nevertheless, ‘the economic insecurity of households’ comes out as the strongest statistical association, along with the economy in general.

Explaining Labour’s losses among the economically insecure

Labour is not seen by the economically insecure to be handling household economic security well. This can be seen if we compare Labour’s handling scores for those respondents who reported feeling ‘secure’ or ‘insecure’, as we do in Figure 19. This shows a doubling of those saying Labour is handling the economic security of households very badly among the insecure (41%), compared to the economically secure (20%).

Note that evaluations of Labour’s handling of household economic security are more negative than positive among both groups, with half of the secure thinking Labour is handling household economic security badly, and 70% of the insecure group doing so. Only 9.5% of secure respondents thought Labour was handling household economic security ‘well or very well’, with the figure being 5% among those who reported feeling insecure. This may be related to household economic security being something people think governments cannot handle especially well in general, or might not be expected to ‘handle well’, as opposed to being the responsibility of individuals and households, but if that were the case, it could be argued that economic insecurity should not be related to defections away from the Government at such an early stage or, for the outgoing Conservative Government, as strongly as they were between 2019 and 2024.

Unsurprisingly, there are differences between the insecure and secure in the importance they attach to the economic security of households, although both groups rate the economic security of households as very important, according to our survey. Among those who reported being economically insecure, 84% said that the economic security of households was important (scoring it 6 or above on a 0–10 scale where 0 = least important and 10 = most important), whereas 77% said so if they reported feeling secure. Tellingly, 39% of insecure respondents rated it extremely important (a 9 or 10 on the scale) whereas the same figure was only 20% among those who felt economically secure.

We also see – in Figure 20 – a bigger perceived gap between the importance to the respondent and the perceived importance of economic insecurity for households to the Labour Government. Those who attach equal importance to household economic security represent the same proportion – about one-quarter of respondents – in each group, the secure and insecure. However, those who felt insecure were more likely to rate economic insecurity as more important than they thought the Labour Party gives to the importance of household economic insecurity: 45% of the insecure rate the issue as more important than they believe is true for Labour, and 18% gave at least a 5-point gap between their own prioritisation of the issue and the extent to which they believe that Labour was prioritising it. This compares to 45% and 9%, respectively, among those who are economically secure.

Overall, the greater importance attached by respondents to household economic security than it is perceived as important to Labour suggests that it is politically important for Labour, or any political party, to match the importance attached to economic security by the general public. This is a potentially fruitful strategy both for retaining or winning back the support of those who feel insecure and retaining support among those who currently feel secure.

5. Conclusion

Economic insecurity is a powerful concept for understanding the economic experiences of Britons, particularly in mid-life, and the electoral risks for political parties in not addressing people’s feelings of insecurity and financial worries.

Our analysis has shown how economic insecurity is a more wide-ranging concept in understanding financial hardship and distress than income-based measures. Indeed, a focus on income to understand financial distress may be misleading considering the age distribution of financial worries. The lower income levels of pensioners, for example, are at odds with the much higher levels of economic security felt by pensioners and the over-60s in general. Young adults in Britain report higher levels of insecurity than their elderly counterparts, but the group exhibiting the highest level of economic insecurity and financial worries is by far those in mid-life: that is, those aged approximately between 35 and 59.

Broadly, the experience of insecurity in mid-life is consistent with higher mortgage costs, lower levels of savings, childcare responsibilities, higher levels of unsecured debts and less disposable income. We also noted how younger respondents were more likely to be supported by family and have expectations of future economic well-being, and those in mid-life have substantial financial outgoings, not just for ‘children’ but also for adult children, and – for some – caring for parents and/or grandparents. Economic insecurity in mid-life is even more pronounced among women compared to men, people without university degrees, those who do not own their home, people with long-term illnesses and/or disabilities, and single adults, which we attribute to dealing with financial burdens and risks on their own, without the buffer of an additional income or other sources of financial insurance to support the household.

Our insights into mid-life economic insecurity are important for policy-makers. We noted the negative outcomes associated with economic insecurity, such as poor mental health, obesity, and alcohol and drug abuse, and important economic behaviours that matter for the wider economy, such as fertility decisions, labour market decisions and private expenditure. Understanding the distribution of economic insecurity in mid-life should be a focus for the Government. Using appropriate measures of insecurity, as a complement to established measures of absolute and relative income poverty, is critical to developing policies that mitigate these negative outcomes.

Political parties that ignore electoral groups, fail to focus on household economic insecurity and people’s financial worries in mid-life, and fail to devise policies that improve economic insecurity – such as those mitigating economic risks – will likely suffer at the ballot box. Since age is such a structuring and anchoring factor in contemporary British elections, those in mid-life take on an increased electoral importance. We suggest that they could be a critical ‘swing voter’ group because they are both more moveable between different party choices and have greater levels of economic insecurity.

We have shown that those in mid-life are more likely to be politically ‘undecided’, they are more likely to switch parties at a very early stage of the electoral cycle following the July 2024 general election and they are more likely to have defected from a vote for the incoming Labour Government. Our analysis suggests that Labour is already suffering from an ‘anti-incumbent vote’ among the mid-life economically insecure, just as the outgoing Conservative Government was punished at the ballot box by those who felt economically insecure. Labour’s difficulties among this group appear to be one of ‘importance incongruence’; the majority of our respondents place greater importance on household economic insecurity than they believe is true of the Labour Government. Those who feel economically insecure are especially likely to think Labour is handling household economic insecurity badly, or very badly.

Even in the first 100+ days of a Labour Government, Labour has had a problem among those who feel economically insecure. Our analysis suggests that if Labour wants to do something about this electoral risk, it should focus on policies that provide greater support against economic risks, and it should focus in particular on people who experience these risks most acutely: those in mid-life.

A view of British pedestrains at a road crossing, from directly above.

This report is part of the public attitudes topic.

Find out more about our work in this area.

Discover more about public attitudes