Research and analysis

Attitudes and Awareness before State Pension age

Published 7 October 2024

Findings from a survey of 2,535 British adults aged 54 to 64.

October 2024

Author: Lyndon Clews Pensions & Later Life Analysis Division, DWP

Research Report 1067

Crown copyright 2024.

You may re-use this information (not including logos) free of charge in any format or medium, under the terms of the Open Government Licence. To view this licence, visit www.nationalarchives.gov.uk or write to:

The Information Policy Team,
The National Archives,
Kew,
London
TW9 4DU

or email psi@nationalarchives.gsi.gov.uk.

This document/publication is also available on our website

If you would like to know more about DWP research, email socialresearch@dwp.gsi.gov.uk

First published October 2024. ISBN 978-1-78659-719-9

Views expressed in this report are not necessarily those of the Department for Work and Pensions or any other government department.

Summary

Background

The Department for Work and Pensions (DWP) is responsible for ensuring the financial security of current and future pensioners. The State Pension is core to achieving this objective and provides a regular income to almost 13 million people.

Demographic trends have necessitated reforms to the UK State Pension system; the new State Pension was introduced in 2016, the State Pension age is being increased and State Pension deferral policy has been revised.

This report presents the results of a survey commissioned by DWP to understand awareness of the State Pension system and attitudes toward retirement among a cohort approaching State Pension age. A random sample of 2,535 adults aged 54 to 64 in Great Britain participated in the survey between March and May 2018.

Key findings

• Retirement expectations are changing. More non-retired respondents expected to retire at 66 than any other age. Most still expect to retire before State Pension age, the average expected age of retirement was 64.5 years.

• Over three quarters (77%) expected to become eligible for the State Pension within 12 months of their actual State Pension age. 42% knew when they would reach State Pension age exactly.

• There is broad awareness the State Pension age is changing. Less than 1-in-5 men expected to reach State Pension age at 65 and only 6% of women expected to receive the State Pension before 65.

• Most (60%) were aware it is possible to defer the State Pension but there was less awareness of the specific details of the policy. Just under half correctly identified they could receive a higher weekly payment (increment) if they chose to defer their State Pension.

• Almost four-fifths (78%) correctly identified that the income from the State Pension was taken into account when working out whether they have to pay income tax. Half of the respondents knew that National Insurance did not apply to earnings past State Pension age.

• 10% intended to defer their State Pension. Expecting to still be working was the main reason for this decision.

• Faced with a hypothetical choice between deferring for a higher weekly payment or a lump sum, there was a stronger preference for the lump sum regardless of the generosity of the increment.

Glossary of terms

New State Pension	The new flat-rate, or single-tier, state pension that applies to everyone reaching State Pension age after 6 April 2016.
State Pension	The State Pension is a regular income paid by the UK Government to people who have reached State Pension age. Eligibility and amount payable depend on an individual’s record of National Insurance contributions/ credits.
State Pension age	State Pension age is the earliest age an individual can start receiving their State Pension.
State Pension deferral	State Pension deferral refers to delaying the payment of State Pension beyond an individual’s State Pension age.
Workplace Pension	Any pension scheme provided as part of an arrangement made for the employees of a particular employer.

Background

The United Kingdom (UK) population is ageing. The Office for National Statistics estimate 18.2% of the population were aged 65 years or older in 2017, with this figure projected to grow to 20.7% by 2027.^{[footnote 1]}

The Department for Work and Pensions (DWP) is responsible for ensuring the financial security of current and future pensioners. The State Pension is a benefit that provides eligible individuals with a regular income after they reach State Pension age. In 2018/2019, 12.7 million people received the State Pension and it accounted for over half of total benefit expenditure.^{[footnote 2]}

The foundation of a universal contribution-related State Pension was laid in the 1940s. There have been recent reforms to the State Pension system, including:

• Equalisation of State Pension age (SPa). Since 2010, Women’s SPa has gradually increased from 60 to equal the male SPa (65 years) in November 2018.

• Increases to State Pension age. Between December 2018 and October 2020 the SPa for both males and females will increase to age 66 before increasing to age 67 between 2026 and 2028.

• Introduction of new State Pension (nSP). Everyone reaching SPa after 6 April 2016 will receive the nSP. This provides a flat-rate amount of benefit for each qualifying year of paid or credited National Insurance contributions.

• Revised State Pension deferral arrangements. When an individual reaches SPa they can delay (defer) receiving the State Pension in return for a higher weekly payment when they decide to claim. The options available following deferral were revised when nSP was introduced in April 2016.

• Independent review of State Pension age. A framework establishing independent review of SPa was legislated for in 2014. The first review reported in 2017 and included recommendations relating to State Pension age and deferral.^{[footnote 3]} DWP has made a commitment to consider these recommendations.^{[footnote 4]}

The government has also set out measures to support people to remain economically active in later life. The Default Retirement Age was removed in 2012, prohibiting compulsory retirement, unless objectively justified, and providing greater opportunity for people to participate in the labour market at age 65 and beyond.^{[footnote 5]} In 2017, the government published Fuller Working Lives: A Partnership Approach outlining the business case for employers to recruit, retain and retrain worked aged 50 and over.^{[footnote 6]} In response to these reforms DWP commissioned Kantar TNS to conduct a survey among members of the public approaching SPa in order to understand attitudes towards retirement and awareness of the UK State Pension system. An additional aim was to explore attitudes towards the deferral recommendations made in the independent review of SPa.

This report complements previous research commissioned by DWP on State Pension Deferral,^{[footnote 7]} Attitudes to Pensions^{[footnote 8]} and Fuller Working Lives.^{[footnote 9]}

Methods

In order to measure awareness of aspects of the State Pension system and to assess the attitudes of the population towards State Pension deferral, Kantar TNS were commissioned by the Department for Work and Pensions (DWP) to include a module of bespoke survey questions in a pre-existing nationally representative omnibus survey.

The questionnaire was administered to an achieved sample of 2,535 British adults aged 54 to 64 years old. Survey interviews were conducted face-to-face at the respondent’s home using computer assisted personal interviewing (CAPI). Fieldwork was conducted between Wednesday 14 March and Friday 11 May 2018. Weights were applied to ensure that the data is representative of the population of Great Britain aged 54 to 64 years by; sex, age, social grade, region and working status.

Kantar TNS Omnibus use a random, nationally representative and robust sample to ensure data is collected from a wide cross section of the target population. The sample design involves randomly selecting sampling points and randomly selecting blocks of 250 addresses within these areas. Finally, interviewers select addresses to fulfil quota requirements.

In strict terms, the application of standard errors and statistical significance are applicable only to probability samples rather than the random location design adopted for this survey. Significance testing is commonly carried out for quota surveys like this one, with the assumption being that the variance found in a random location sample is broadly similar to that of an equally specified probability sample. In such cases, the significance testing is intended to provide a rough guide to the margins of error governing the survey.

This data has been analysed here as though it were a probability sample, in line with how other government institutions have published analysis of this omnibus survey.^{[footnote 10]} The analysis contained in this report has been prepared by staff at DWP.

A pilot study with 305 respondents was conducted in February 2018, in order to inform the final design of the questionnaire. The pilot provided valuable insight into question comprehension and informed alterations that improved the reliability of the data collected. The pilot data was for design purposes only and does not feature in the analysis of findings presented in this report.

Reporting conventions

The survey respondents remain a sample of the total population of 54 to 64 year olds in Great Britain. There is therefore likely to be a difference between the answers given by the sample and the true figure only achievable if census data was collected from all GB adults aged 54 to 64. Further information about the statistical reliability of the sample is provided in Appendix A.

This report examines differences in the data between different groups of respondents within the total surveyed sample. All of the sub-group differences discussed have been tested at the 95% significance level and are statistically significant unless specifically stated. This means that the differences present are unlikely to be the result of random variation between the sample and the population from which it was drawn.

In addition to being statistically significant, only sub-group differences which are relevant in the context of the question being analysed are commented on in the report.

Percentages in charts do not always add to 100% due to rounding, the exclusion of ‘don’t know/not applicable’ categories, or relate to survey questions that allow survey participants to give more than a single response. In instances where findings presented are based on responses of fewer than 100 respondents, reference is made to the small base sizes in the text.

Findings

This chapter presents the findings from the research. The questionnaire addressed 4 main topics:

1. retirement expectations

2. awareness of State Pension age

3. awareness of working past State Pension age and State Pension deferral

4. attitudes toward the deferral reforms recommended by the Independent Review of State Pension age, published in 2017.

Retirement Expectations

The average age of exit from the labour market has increased over the past two decades, although it has failed to keep up with increases in life expectancy.^{[footnote 11]} As the UK has an ageing population and workforce, it is increasingly important to have a good understanding of the working intentions of individuals who are approaching State Pension age (SPa).

This section includes findings relating to economic activity, expected age of retirement, retirement duration and perceived adequacy of future retirement income.

Economic Activity

The most common economic status (reported by 43% of respondents across the whole sample) was - employed full time - defined as 30 hours or more a week. Just 15% of the sample of 54 to 64 year olds considered themselves retired.

Figure 1 shows how economic activity declines with age. The proportion working full time decreased substantially from approximately half of the respondents aged 54 to 57 (51%) to less than a third of respondents aged 61 and above (31%).

Figure 1: Economic Activity by Age

Q. Which of the following best describes your current economic situation?

	Age 54 to 57	Age 58 to 60	Age 61 and above
Employed full time	51%	42%	31%
Employed part time	18%	18%	19%
Retired from paid work	7%	15%	27%

Base: All participants: 54 to 57 (n=912), 58-60 (n=773), 61 and above (n=850)

This decrease in full time employment was matched by an increase in the proportion who defined themselves as ‘retired from paid work’, which increased to over a quarter (27%) of respondents aged 61 and over.

Respondents who were retired from paid work (15% of the whole sample) were asked to identify their main reason for retiring. The most reported reason for being retired was ‘because of own ill-health’, mentioned by 29% of retirees. Other prominent reasons for being retired included ‘wanting to give up work’ (18%) and being ‘offered reasonable financial terms to retire early or take voluntary redundancy’ (14%).

Once out of the labour market, the vast majority of retired respondents did not expect to return to employment. Only 16% of retired respondents expected to do any paid work, either full time or part time, in the future.

These findings serve to highlight that a sizeable proportion of this cohort were not in paid employment in the years before their State Pension age (SPa). While some factors could be said to involuntarily ‘push’ respondents to leave the labour market, such as ill-health - other factors could be seen as ‘pulling’ respondents out of the labour market such as being offered attractive financial terms to retire.

Expected retirement age

The vast majority (85%) of the sample were not retired. These respondents were asked about their plans and expectations for working in later life. Specifically, those currently in work were asked at what age they expect to retire or will consider themselves retired.

Figure 2 shows that more respondents expected to retire at 66 than at any other age, with 22% of non-retired respondents expecting to retire at this age. Age 66 is the most common SPa among this sample of 54 to 64 year olds, indicating the continued signal effect that SPa has on retirement decisions. Analysis by the Institute for Fiscal Studies has suggested that recent SPa increases have had an effect on employment rates for women, potentially through providing a signal of an appropriate age to retire.^{[footnote 12]}

Figure 2: Expected age of retirement among non-retired respondents

Q. At what age do you expect you will retire or will consider yourself to be retired?

Age	Percentage of people
Before 60	10%
60 to 64	19%
65	20%
66	22%
67	17%
After 67	12%

Base: Non-retired respondents excluding ‘Don’t Know’ responses (n=1912)

Although age 66 is the most common response among this sample, the median expected age of retirement is 65 and average (mean) expected age of retirement is 64.5 as substantially more respondents expect to retire before age 66 than after (49% compared to 29%).^{[footnote 13]}

There is a small difference in expected age of retirement between males (64.7 years) and females (64.3 years), equivalent to approximately 5 months. Respondents who were self-employed were more likely to state that they expected to retire later, on average (65.4 years) while those who described themselves as sick or disabled expected to retire the earliest (63.5 years).

Overall 38% of all respondents were either already retired or expected to retire before age 65, indicating that a substantial number expect to or have already exited labour market before 65.

Revised retirement age once State Pension age is known

In an effort to better understand the association between SPa and expected age of retirement, those respondents who did not know their SPa exactly were asked if they still intended to retire at the age they originally reported after they were informed of their SPa.

36% of those who did not know their SPa chose to revise their expected age of retirement. This may reflect that most respondents estimate of their SPa were very close to their actual SPa (see Figure 7).

Figure 3 shows that once individual SPa was known there was a six percentage point increase in the proportion who expected to retire at 66 and a decrease in the proportion who expected to retire at 65, in their early 60s and before age 60.

Figure 3: Original and revised expected age of retirement

Q. At what age do you now expect to retire or will you consider yourself to be retired?

Retirement age	Revised	Original
After 67	12%	12%
67	20%	16%
66	28%	22%
65	16%	20%
60 to 64	17%	18%
Before 60	7%	10%

Base: All non-retired respondents excluding Don’t Knows (n=1874)

This indicates that SPa is an important factor influencing expected age of retirement and in making retirement decisions.

Retirement duration

The period of time - in years - that an individual expects to be retired, is likely to influence the amount of wealth that they accumulate for their retirement and the rate at which this wealth is drawdown to provide income in retirement.

Non-retired respondents were first asked if they had thought about how many years of retirement they may need to fund. Among non-retired respondents, 39% reported they had considered how many years of retirement they might need to fund.

Respondents from professional occupational groups were more likely to report that they had considered how long they expect to spend in retirement. 63% of respondents in occupational Group A (senior professionals) and 58% in occupational Group B (middle management executives) had considered their retirement duration compared to 27% in Group D (semi-skilled and unskilled manual workers) and 15% in Group E (casual workers and those who are economically inactive).

Next, respondents were asked how many years they expect to be retired. Figure 4 shows those who had thought about their retirement duration were more likely to expect to be retired for between 20 and 24 years and 25 and 29 years than respondents who had not.

Figure 4: Expected duration of retirement

Q. For how many years do you think you will be retired?

Expected retirement duration (years)	Considered retirement duration	Not considered retirement duration
Don’t know	5%	23%
Below 15	20%	30%
15 to 19	12%	11%
20 to 24	29%	17%
25 to 29	18%	7%
30 to 34	10%	5%
35 plus	6%	6%

Base: Non-retired respondents who have considered retirement duration (n=934). Non-retired respondents who have not considered retirement duration (n=1601)

This trend could reflect that respondents from higher occupational classifications expect to be retired for longer, a proxy for life expectancy. However, this could also signal that some individuals in this cohort may have unrealistic expectations regarding how long they will be retired. Analysis by the Institute for Fiscal Studies using the English Longitudinal Study of Ageing, found on average that older individuals tend to underestimate their chances of survival to ages 75, 80 and 85.^{[footnote 14]} The implications being that some respondents may be at risk of not accumulating sufficient resources for retirement or do not make optimal decisions in order to ensure they have a secure income in retirement.

Perceived adequacy of future retirement income

Non-retired respondents were asked to consider whether the income that they expect to receive during retirement would be, in their opinion, adequate or sufficient to meet their needs. It was left to each respondent to determine what constitutes a sufficient level of income.

60% of all respondents expected to have just about enough income to ‘meet their needs in retirement’ and 17% expected to have ‘more than enough income to meet their needs’ in retirement.

Figure 5 shows that female respondents were less likely to report an expectation to have ‘more than enough income in retirement to meet their needs’ than males (14% compared to 21%).

Figure 5: Perceived adequacy of future retirement income

Q. Looking to the future, which of the following best describes how you expect your income during retirement to be?

	Male	Female
More than enough to meet your needs	21%	14%
Just about enough to meet your needs	60%	64%
Less than enough to meet your needs	19%	22%

Base: Non-retired respondents excluding Don’t Know: Male (n=985) Female (n=973)

The differences observed for the other two responses to this question were not statistically significant.

Awareness of the State Pension age

This section presents findings relating to the knowledge respondents have of their own State Pension age (SPa), the earliest age at which they can personally expect to receive the State Pension. This cohort of 54 to 64 year olds is distinct, as it will be affected by legislated increases in the SPa from 65 to 66 and from 66 to 67. A self- assed and objective measure of State Pension knowledge is presented.

Self-assessed knowledge of the UK State Pension System

Perceived awareness of the State Pension is an important factor as it is likely to influence behaviours, such as whether an individual seeks additional information about their State Pension entitlement or makes other arrangements to plan and prepare for later life.

To evaluate self-assessed knowledge of the UK State Pension system, respondents were asked how knowledgeable they feel about the UK State Pension. The majority of respondents reported having either a ‘good’ (23%) or ‘reasonable basic’ (38%) understanding of the UK State Pension system.

Figure 6 shows females were more likely to report knowing ‘little or nothing about the UK State Pension’ than men (19% compared to 14%).

Figure 6: Self-assessed knowledge of the UK State Pension

Q. Which of the following statements best describes how knowledgeable you feel about the UK State Pension?

Reported knowledge	Female	Male
Know little or nothing	19%	14%
Very patchy knowledge	21%	22%
Reasonable, basic knowledge	38%	38%
Good knowledge	21%	25%

Base: All respondents: Male (n=1284), Female (n=1251). ‘Don’t know’ and ‘Refused’ not shown.

Actual State Pension age knowledge

The actual SPa of each respondent, as set out in current legislation, was calculated by combining information about the respondent’s age with their interview date. In some cases a full date of birth was requested for this calculation. The SPa profile of the sample is shown in Table 1.

Table 1: State Pension age profile of the sample

State Pension age	Number of respondents
65	55
65 to 66 transition	194
66	1334
66 to 67 transition	198
67	754
Total	2,535

It is important to note that the SPa of respondents in the 65 to 66 transition group will be a combination of years, months and days. In these cases, SPa was rounded up or down to the nearest month. As there was only a small number of respondents in the two transitional groups in the sample, any findings relating specifically to these groups should be interpreted with caution.

It is possible to estimate how many people accurately know their SPa by comparing when respondents expect to reach SPa with the actual SPa calculated for each individual. Respondents were given the option to estimate both the year and month they will reach SPa, even though for the vast majority (85%) of the sample their SPa will not contain a month element.

Across the whole sample 42% knew their State Pension age exactly. This figure is likely to represent a lower bound estimate as respondents affected by transitional arrangements may have rounded their estimate to the nearest whole year, in which case they will be classified here as not knowing their SPa exactly.

To account for respondents potentially rounding their estimates, Figure 7 shows how close respondents’ estimates were to their actual SPa. Over three quarters (77%) expected to receive their State Pension within 12 months of their actual SPa.

Figure 7: Difference between estimated and actual State Pension age

Q. Based on current UK State Pension arrangements, at what age will you become eligible to receive a UK State Pension?

Proximity to actual SPa	Percentage
Don’t know State Pension age	6%
Later between 1 and 5 years	5%
Later up to 1 year	20%
Accurate	42%
Earlier up to 1 year	15%
Earlier between 1 and 5 years	9%
Earlier more than 5 years	3%

Base: All respondents (n=2535). Overestimated by more than 5 years not shown.

The Attitudes to Pensions 2012 survey found that respondents tended to expect to receive their State Pension at an earlier age than was legislated for, with this expectation especially prominent among females.^{[footnote 15]} In the research presented here, the proportion expecting to receive their State Pension both at earlier age and at a later age than currently legislated for was approximately equal at 26% either way.^{[footnote 16]}

How does knowledge of State Pension age vary?

This section looks at how actual SPa knowledge varies across key groups in the sample.

Actual State Pension age

The most substantial difference in exact awareness of SPa is whether the respondent was affected by transitional SPa arrangements. In part this is likely to reflect the requirements of the question, for this group to know their SPa exactly respondents needed to report accurately both the year and month they will reach SPa. It is possible some respondents in these transitional groups may have rounded their answer to the nearest year rather than specify a month in their answer.

SPa awareness is higher for respondents whose SPa is a single year. Figure 8 shows exact awareness is 45% among respondents with an actual SPa of 66 or 67. Exact awareness is lower among transitional groups, for those affected by 65 to 66 transition it is 35% where among the 66 to 67 transition it is just 19%.

Figure 8: Expected State Pension age by actual State Pension age

Q. Based on current UK State Pension arrangements, at what age will you become eligible to receive a UK State Pension?

Actual SPa	Before 65	65	65 to 66	66	66 to 67	67	67 plus	Don’t know
65	0%	78%	7%	5%	-	-	-	6%
65 to 66	-	23%	35%	24%	3%	9%	4%	-
66	4%	13%	-	45%	4%	21%	5%	4%
66 to 67	5%	-	12%	17%	19%	29%	12%	3%
67	7%	17%	-	7%	-	45%	11%	8%

Base: All respondents approaching State Pension age: SPa 65 (n=55), SPa 65-66 (n=194), SPa 66 (n=1,334), SPa 66-67 (n=198), SPa 67 (n=754)

Note: small base sizes for those reaching SPa at 65, 65 to 66 and 66 to 67 mean these results should be treated with caution.

Figure 8 highlights that among this cohort, there was a general awareness that SPa is changing, as relatively few individuals expected to become eligible at the traditional SPa. The male SPa has remained at age 65 for a period of over 70 years and started to increase from age 65 to age 66 from December 2018. Less than 1-in-5 males (18%) expected to reach SPa at 65. Before 2010 the female SPa was 60, only 6% of females in this cohort expected to reach their SPa before age 65.

Sex

Figure 9 shows how exact awareness of SPa varies between males and females. Although men have slightly higher exact awareness (43% compared to 40%) this difference is not statistically significant.

Figure 9: State Pension age awareness by sex

Q. Based on current UK State Pension arrangements, at what age will you become eligible to receive a UK State Pension?

	Earlier more than 1 year	Earlier by up to 1 year	Accurate	Later by up to 1 year	Later more than 1 year	Don’t know when will reach SPa
Female	13%	15%	40%	21%	4%	6%
Male	10%	15%	43%	19%	6%	6%

Base: All respondents: Female (n=1251), Male (n=1284)

Variation by other demographic and socio-economic characteristics

To further understand how awareness of SPa varies, a Logistic Regression analysis was carried out. This analysis is useful to identify associations between one or more explanatory variables and a binary outcome (see Appendix B for more detail). In this case it is used to show the characteristics associated with a respondent knowing their exact SPa (see Appendix C for full results).

The results from the Logistic Regression analysis support the findings reported in Figure 8 and Figure 9. The analysis shows that there is a significant and large difference in exact knowledge of SPa based on whether a respondent is affected by SPa transitional arrangements. However, there is no statistically significant difference based on sex.

The analysis also shows that the following characteristics were associated with a respondent knowing their SPa exactly:

• Ethnicity. Respondents who identify as Asian or Asian British, relative to White British, were less likely to know their SPa exactly. Black, African or Caribbean respondents were 20 percentage points less likely to know their SPa exactly than the White British reference group.

• Tenure. Respondents who were renting (either privately or from a Local Authority/housing association) were less likely to know their SPa exactly.

• Age. Awareness increases with age, a 1-year increase in age was associated with a 2.9% increases in the probability of knowing the SPa exactly.

• Private Pension wealth. Respondents who expected to receive a substantial income from a private or personal pension were more likely to know their SPa than those who did not.

• Self-assessed State Pension age knowledge. Respondents who reported greater ‘subjective’ knowledge of the UK Pension system were more likely to know their exact SPa.

• Occupational group. Respondents in Occupational Group A and E were less likely to know their SPa than the C1 reference group.

• State Pension information. Each additional reported source of information consulted in the last 12 months increased the probability of a respondent knowing their State Pension age exactly by 3.7%.

State Pension information

The Department for Work and Pensions (DWP) and a range of agencies, businesses and charities provide information about the State and Workplace Pensions.

Respondents were asked which sources, if any, they had used in the last 12 months to find out anything about the UK State Pension (Table 2).

Table 2: Sources of State Pension information consulted

Q. Have you used any of the following sources of information in the last 12 months to find out about anything to do with the UK State Pension?

Sources of State Pension information	% of respondents*
No searching done in the last 12 months	41%
Online with DWP, Pension Service, GOV.UK or Check your State Pension	32%
Internet searches (e.g. Google, Yahoo)	11%
Friends and family	7%
Employer or in the workplace	7%
A State Pension statement through the post	7%
Professional financial adviser	6%
Free Advisory Services (e.g., Citizens Advice)	6%
Face to face or over telephone with Jobcentre Plus, DWP or the Pensions Service	6%
Press or social media	3%

Base: All respondents (n=2580) yielded 3247 responses

*Percentages do not add up to 100% due to participants being able to select more than one response

Most respondents reported that they had used one source of information in the last 12 months to find out something about the UK State Pension. Online sources were the most frequently mentioned, 32% reported using an online service provided by DWP, GOV.UK and/or Pension Service and 11% mentioned broad internet searches (Google, Yahoo).

Seeking advice directly, either through DWP, Free Advisory Services and/or Financial Advisers was mentioned by just 1 in 20 respondents (6%).

Awareness of working past State Pension age and State Pension Deferral

This section examines the extent respondents understood the options currently available to them when they reach State Pension age (SPa). Specifically, it looks at knowledge of the option to continue working past SPa and to defer the State Pension.

Working past State Pension age

Prior to 5 October 2012, the Default Retirement Age provided employers with a legal basis to enforce compulsory retirement at age 65. In 2012 the Default Retirement Age was removed giving employees increased control over labour market participation decisions.

Respondents were asked if they believed that they could continue working upon reaching their SPa. The vast majority (86%) were aware that they could continue working when they reach their SPa. Just 4% were of the opinion that they must stop working at SPa.

Claiming the State Pension

When an individual reaches SPa they are required to make a claim in order to receive the State Pension. If an individual is eligible for the State Pension but does not make a claim they will automatically defer their State Pension, accruing additional State Pension entitlement until the point at which the State Pension is received.

The requirement for a claim to be made in order to receive the State Pension, means that it is possible that some people may ‘accidentally defer’ their State Pension.^{[footnote 17]} To explore the likelihood of this possibility, respondents were asked if they think that the UK State Pension has to be actively claimed or if they think that it is paid automatically when people reach SPa.

Figure 10 shows responses to this question adjacent to the responses to a similar question which asks whether Pension Credit - a means tested benefit for pensioners on low income- is paid automatically or whether recipients are required to actively make a claim. DWP estimates show that 6 out of 10 of those entitled to Pension Credit claimed the benefit in 2016/2017.^{[footnote 18]}

Figure 10: Understanding of claiming State Pension and Pension Credit

Q. Could you tell me whether you think [the UK State Pension/ Pension Credit, additional money the government gives to low income pensioners] is paid automatically to eligible people [when they reach State Pension age] or does it have to be claimed?

	Pension Credit	State Pension
Needs to be claimed	72%	41%
Paid automatically	7%	49%
Don’t know	11%	9%

Base: All respondents (n=2535)

A clear difference emerges between the way that people believe the two benefits are accessed. While the vast majority (72%) of respondents correctly identified that active claims are required in order for those entitled to receive Pension Credit, more respondents expressed the view that the State Pension is paid automatically upon reaching SPa (49%) than those who are aware that in order to receive State Pension it needs to be claimed (41%).

Those who are currently in receipt of benefit were less likely to say that Pension Credit needs to be claimed (64%) compared individuals who were not currently in receipt of a benefit (73%), potentially indicating current benefit recipients may expect their eligibility for Pension Credit to be automatically assessed.

State Pension Deferral knowledge

With a few exceptions, upon reaching SPa individuals are able to claim State Pension immediately or to defer (delay) receiving the State Pension. This choice is available to people regardless of whether they are employed when they reach SPa. Deferral is a longstanding feature of the State Pension system and has been reformed in recent years, as summarised in Table 3.

Table 3: Summary of recent changes to State Pension Deferral

Applicable to people who reach SPa	Minimum period	Increment accrued	Lump sum
After 6 April 2016	9 weeks	1% every 9 weeks (5.8% per year)	No
Between 6 April 2005 and 5 April 2016	5 weeks	1% every 5 weeks (10.4% per year)	Yes- after 12 months
Prior to 6 April 2005	7 weeks	1% every 7 weeks (7.4% per year)	No

Under the new State Pension (nSP), applying to all people reaching SPa after 6 April 2016, if an individual defers they receive an additional amount (increment) in their weekly State Pension payment for the duration of their claim. The weekly payment increases by 1% for every 9 week period of deferral, equivalent to a 5.8% increase over a year.

Between 2006 and the introduction of the nSP if an individual deferred their State Pension for at least 12 months they could chose to receive a lump sum payment. The value of the lump sum was equivalent to the amount of State Pension forgone with an interest rate 2% higher than the Bank of England base rate.

Alternatively, individuals could choose to receive an additional increment to their State Pension, with the weekly payment increasing by 1% for every 5 weeks of deferral (equivalent to 10.4% a year).

To assess broad awareness of State Pension deferral all respondents were asked their view of whether, if they were to continue working beyond SPa, they could receive the State Pension at the same time, put off claiming State Pension until a later date (defer), both of these options or neither. Figure 11 shows that 42% identified the correct answer that both of these options were available, only 3% thought that neither option was available.

Figure 11: Awareness State Pension can be deferred or combined with earnings

Q. Which of the following is your understanding of what people can do if they continue working beyond their State Pension age?

	Percentage
Receive SP at the same time	29%
Claim at a later date	18%
Both of these	42%
Neither	3%
Don’t know	8%

Base: All respondents (n=2535)

A smaller proportion chose one of the two correct options, either receiving State Pension at the same time (29%) or putting off claiming until a later date (18%). In total 60% knew that they could defer their State Pension.

To assess knowledge of more specific elements of deferral policy, respondents were asked a subset of four true or false questions taken from the Attitudes to Pensions Knowledge test.^{[footnote 19]}

The first two statements refer to deferral and asked whether it was true or false that if an individual delayed receiving their State Pension they could receive:

1. Extra State Pension in their regular payments when they do start to take it.

2. An additional one-off cash lump sum.

All the sample in this research fall under nSP deferral rules, which means a weekly increment is the only deferral option available, not the lump sum.

Figure 12: Knowledge of lump sum and increment

Q. Do you think the following statement is true or false?

• If you delay receiving your State Pension, you can get extra State Pension in your regular payments when you do start to take it?

• If you delay receiving your State Pension you can get an additional one-off cash lump sum when you do start to take it?

	True	False	Don’t know
If you defer you can get an additional one-off cash lump sum	36%	42%	22%
If you defer you can get extra State Pension in your regular payments	48%	36%	16%

Base: All respondents (n=2535)

Figure 12 shows that 42% correctly identified that the lump sum statement was false, where 48% knew the increment statement was true. Respondents were more accurate when answering the statement about the increment than the lump sum, potentially reflecting the recent removal of the lump sum option.

Although more respondents answered these statements correctly than incorrectly, the large proportion of ‘don’t know’ answers illustrate that less than half of the total sample identified the correct answer. If respondents were to guess at random, approximately 50% would correctly answer each statement.

Awareness of the availability of the increment increased with age, the proportion who identified the statement as false decreased from 40% among 54 to 57 year olds to 31% among the oldest group of respondents (61 to 64 years old).

Figure 13 shows this increase in awareness is driven solely by increases among males, with a 12 percentage point gap emerging between males and females aged 61 and above.

Figure 13: Awareness of deferral increments by age and sex

Q. If you delay receiving your State Pension, you can get extra State Pension in your regular payments when you do start to take it?

	Male	Female
Age 54 to 57	46%	44%
Age 58 to 60	51%	45%
Age 61 and above	58%	46%

Base: All respondents; Male (54-57 n=444, 58-60 n=376, 61 and above n=464), female (54-57 n=468, 58-60 n=397, 61 and above n=386)

Respondents were asked whether a further two statements about the State Pension system were true or false:

3. People who continue working after State Pension Age do not have to pay National Insurance contributions? 4. The income a person receives from the State Pension is taken into account when working out whether they have to pay income tax?

The correct answer for both statements is true. Figure 14 shows that awareness of both these topics was higher than the two deferral statements reported in Figure 12.

Figure 14: Awareness of National Insurance exemption and Tax

	True	False	Don’t know
The income received from State Pension is taken into account when working out whether they have to pay income tax	78%	13%	9%
People who continue working after SPa do not have to pay National Insurance contributions	50%	36%	14%

Base: All respondents (n= 2535)

Almost four out of five respondents (78%) correctly identified that the income received from the State Pension is subject to income tax. This high degree of awareness could be because respondents in this age group are more likely to know somebody who is claiming their State Pension than someone who has deferred.

Half of respondents correctly identified that if they worked past SPa they would be exempt from paying National Insurance contributions, which are only levied on people below SPa. Awareness of the National Insurance exemption was seen to increase as individuals approached SPa. 45% of 54 to 57 year olds identified the statement was true, compared to 53% among respondents aged 58 to 60 and 61 to 64.

Deferral intentions and preferences

The final section of this chapter examines deferral intentions and hypothetical preferences. First it looks at the intentions individuals have for when they reach State Pension age (SPa). The second section explores preferences people may have when presented with deferral reform options based on the recommendations raised in the 2017 independent review of SPa.

As the sample in this study were below SPa when they were interviewed in 2018 these findings relate to stated preferences and it cannot be assumed they will necessarily translate into behaviours and actions when the respondents reach SPa.

Intentions at State Pension age

Once respondents were made aware of their individual State Pension age they were asked whether they intend to claim immediately or delay (defer) their State Pension. Figure 15 shows that the vast majority (81%) intend to claim their State Pension immediately at SPa, 1 in 10 (10%) intend to defer.

Figure 15: Intention at State Pension age

Q. Currently when you become eligible to receive the State Pension at approximately [SPa], which of the following do you intend to do?

	Percentage
Claim State Pension immediately	81%
Delay (defer) claiming your State Pension	10%
Don’t know	9%

Base: All respondents (n= 2535).

The main reason for intending to claim State Pension immediately are reported in Table 4. The most frequently mentioned reasons were related to income, either needing the money or income (mentioned by 33%) or to top up the income they expect to be getting at the time (16%). Reaching SPa and becoming entitled to receive the State Pension was mentioned by 28% of respondents and a desire to stop working by 16%.

Table 4: Reasons for claiming State Pension immediately

Q: Why do you intend to claim your State Pension as soon as you can?

Reason	% of respondents
I will need the money or income	33%
I will be entitled to the State Pension	28%
To top up the income I will be getting at the time	16%
I want to retire and stop working	16%
No reason or incentive to delay	15%
Health issues	11%
Circumstances may change in the future	10%
Other	1%

Base: respondents who intend to claim State Pension at SPa (n=2086). Yielded 2704 responses

Only 231 respondents in the sample of 2,535 intended to defer the State Pension. Among these individuals the central reason, mentioned by 106 respondents (42% of respondents who intended to defer) was because ‘I will be or I intend to continue working’. Due to the small number of respondents who intended to defer the other reasons are not reported.

The link between employment and deferral is explored further in Figure 16. This analysis shows that respondents who are already retired or expect to retire in their early 60s were substantially more likely than over 65s to say they will claim their State Pension immediately.

Figure 16: Intention to claim State Pension immediately by expected age of retirement

Expected retirement age once once SPa is known	% intending to claim immediately at SPa
Already retired	94%
Before 65	90%
65	86%
66	85%
67	77%
After 67	58%

Base: Respondents who provide an expected age of retirement (n= 2290): Already retired (n=496), Before 65 (n=395), 65 (n=311), 66 (n=570), 67 (n=388), After 67 (n=210).

The intention to claim State Pension immediately upon reaching SPa declines from 94% of those who are already retired to 58% among those who intend to retire after 67. The intention to defer is much higher among those who intend to retire at 67 (16%) and after 67 (29%) than those who intend to retire at earlier ages.

A logistic regression analysis was carried out to examine the association between expected age of retirement and deferral intentions. This analysis, presented in Appendix D, shows that those who expected to retire after 67 are 17 percentage points more likely to they intend to defer the State Pension than those who were already retired. Other notable findings include:

• State Pension age. All respondents reaching SPa after 65 years were less likely to intend to defer. Respondents with a SPa of 66 were 9 percentage points less likely to report intending to defer than those who will reach SPa at 65. Among those with an actual SPa of 67 this effect was 16 percentage points

• Self-reported health. Respondents reporting good, fair and poor health were less likely to report intending to defer compared to respondents who reported excellent health. For those with poor health this difference was 10 percentage points

• Time preferences. Respondents who said they would rather receive £1,100 in a year’s time rather receive £1000 immediately were more likely to say they intend to defer upon reaching SPa

• Tenure. Respondents who own their home outright were less likely to intend to defer, once controlling for other variables. This may be because these respondents already have sufficient wealth for retirement

• Private Pension wealth. The proportion intending to defer their State Pension was 6.5 percentage points lower among respondents who expect to receive a substantial income from a private or personal pension that those who do not

No relationship was found in this analysis between expected length of retirement, a proxy for life expectancy, and deferral intentions. As weekly increments are paid from the point the State Pension is claimed until death, it would be more worthwhile to defer if an individual expected to live for a long period of time as the additional increments would be paid for longer.

Analysis by the Institute for Fiscal Studies using the English Longitudinal Study of Ageing found that individuals were twice as likely to defer if doing so represented a fair deal given their survival expectations, although deferral take up remained surprisingly low.^{[footnote 20]} The English Longitudinal Study of Ageing used a different measure of life expectancy than the one used here.

Preferences for the increment and lump sum

To support people to remain economically active past SPa the 2017 independent review of SPa made two recommendations in relation to State Pension deferral.^{[footnote 21]}

The first was that people who defer should have the option to be rewarded through a lump sum once they start drawing their State Pension. The lump sum was previously available to people reaching SPa between 2006 and 2016.

In ideal conditions behavioural evidence would be used to understand whether this cohort exhibit a stronger preference for deferral to be paid as a lump sum or a weekly increment. As weekly increments are the only option currently available, attitudinal evidence pertaining to hypothetical options put to respondents was collected instead. It is not permissible in law to trial differential State Pension benefit offers with different groups for the purposes of research.

Preferences were assessed by asking respondents scenario (vignette) questions where they were invited to choose between:

1. Claiming State Pension immediately at [respondents State Pension age] and receive a weekly State Pension of £160.

And two options:

2. Claiming State Pension later at [respondents State Pension age + 1 year ] and receive a weekly State Pension of [A: £169, B: £172, C: £180].

3. Claim State Pension later at [respondents State Pension age + 1 year] and receive a one off lump sum of [A: £8,355, B: £8,380, C: £8,430], in addition to a weekly State Pension of £160.

There were three scenarios (A, B or C) of the two deferral options, each with a different value attached to the weekly payment (increment) or lump sum.

Respondents were randomly allocated to either version A, B or C for both questions. It was explained to respondents that some of the choices were hypothetical.

Figure 17: Preferences for the lump sum and increment

Q: Upon reaching State Pension age, which one of the following would you choose to do?

	Percentage
Indicated would defer to receive a £8,355 cash lump sum	52%
Indicated would defer to receive a higher State Pension of £169 per week	22%
Originally reported intending to defer the State Pension	10%

Base: Respondents offered the hypothetical choice between claiming immediately, a £8,355 lump sum and £169 increment (n=856)

Figure 17 shows that approximately half (52%) chose the deferral option when it was presented as a lump sum equivalent to the value of the new State Pension plus 1.1% interest, but less than one quarter (22%) chose the £169 weekly payment (roughly equivalent to the current 5.8% increment).

The number selecting the increment option is substantially higher than the 10% who indicated they intended to defer the State Pension when they were asked earlier in the survey (see Figure 15).

As respondents were randomly allocated to three different deferral scenarios, each with a different weekly payment (increment rate) and lump sum amount it is possible to explore how preferences were influenced by these changes.

The effect is most noticeable when looking at the three increment scenarios, presented in Figure 18. As expected, the higher the value of the weekly payment following deferral the more attractive this option became to respondents.

Figure 18: Proportion selecting the weekly increment

Q. Upon reaching State Pension age, which one of the following would you choose to do?

Hypothetical deferral scenario	% indicating they would defer
£180 (13%)	39%
£172 (8%)	31%
£169 (5.8%)	22%

Base: Respondents who were assessed as having understood the vignette questions; £169 (n=813), £172 (n=804), £180 (n=783)

More specifically the proportion who indicated they would consider deferral increased from 22% when the weekly payment was £169, to 31% when it £172 and reached 39% when it was £180.

The lump sum preferences ranged from 53% to 54% and did not vary significantly between the three scenarios. This may reflect the fact that the value of the lump sum was modelled over a one-year period, making the differences between the three scenarios minimal.

Limitations

It is important to emphasise that there are clearly limitations with these hypothetical vignette questions and behavioural inferences should not be made from them as the responses may bear no resemblance to actual future behaviour. They do not provide a reliable estimate of the potential behavioural response to changes in deferral policy.

In order to achieve viable social survey questions the scenarios required simplification of a complex policy area. For instance, it was not possible to explain during the survey whether the deferral lump sum would be subject to income tax or could be inherited in the event of death.

The findings from this survey experiment contrast with previous behavioural evidence when both the lump sum and weekly increment were available prior to 2016, as the majority of people who deferred their State Pension opted to receive the increment instead of the lump sum, although value of the increment was higher than under the current deferral arrangements.

Potential explanations

The apparent popularity of the lump sum option may be related to the return from deferral being more immediate when received as a lump sum rather than a weekly increment. The lump sum would be less subject to the process of discounting, where income received today is valued more than income received in the future, as the lump sum would be paid after just one year where the weekly increment would be paid over a longer period of time.

The lump sum may also appeal to individuals who are more risk averse or who do not expect to live for a long enough to be able to benefit from a deferred pension paid through a higher weekly payment. Some research has identified the ‘certainty effect’^{[footnote 22]}, a preference for certain gains rather than potential risky gains of a higher value, is more prevalent among older age groups.^{[footnote 23]}

Partial drawdown

The second recommendation made in the independent review of SPa was that people over SPa should be able to part drawdown their State Pension.

This would mean individuals could claim a proportion of their State Pension at SPa, deferring the rest for a later date. Unlike the lump sum partial drawdown has never been a feature of the UK State Pension system.

Potential interest in this option was explored by asking respondents if they would be interested in delaying part of their State Pension, while claiming the rest of it. Figure 19 shows that around a fifth (19%) were definitely interested in partial drawdown, and 42% said that they were possibly interested in partial deferral.

Figure 19: Interest in Partial Deferral

Q. Another possibility is that you may be able to delay taking up part of your State Pension, while you claim the rest of it. When you eventually claim the delayed part you would be able to receive either a higher weekly rate or an additional one-off lump sum payment. Would you consider this, if it was available?

Level of interest in partial	Percentage
Yes - definitely	19%
Yes - possibly	42%
No	36%
Don’t know	3%

Base: All respondents (n=2,535). Note: respondents who said they do not understand the question or that the question is too complicated are not shown.

This question explores general sentiment towards partial deferral. As no details about how a partial deferral could work in practice were provided, inferences regarding how many people would potentially use partial drawdown cannot be made from this finding.

Conclusion

This report has presented results of a survey commissioned by DWP to better understand awareness of the State Pension system and attitudes toward retirement among a cohort approaching State Pension age (SPa). The main conclusions to draw from this report are:

• Retirement expectations are changing in line with increases to State Pension age. More non-retired respondents expected to retire at 66 than any other age. The SPa may provide an important signalling function indicating when to retire. 36% of respondents revised their expected age of retirement once they were informed of their SPa.

• Early exit from the labour market remains an issue. Most people continue to expect to retire before reaching SPa. The average expected age of retirement in this study was 64.5 years despite most the survey sample reaching SPa at age 66 or later. Respondents were aware that they could legally work past SPa, but there was little appetite to do so with only 16% of retired respondents reporting that they expected to do any paid work again in the future.

• This cohort expects to have enough income to ensure needs are met in retirement. Over three quarters of respondents expected to have just enough income to ‘meet their needs in retirement’ (60%) or to have ‘more than enough income to meet their needs’ (17%). However, on average people underestimated how long they were likely to be retired.

• People were generally aware the State Pension age is increasing. Less than 1-in-5 men expected to reach SPa at 65 and only 6% of women expected to receive the State Pension before 65.

• Although people know in general that the State Pension age is changing, awareness of how these changes will affect individuals is high but not yet universal. 42% of the sample knew when they would SPa exactly. Over three quarters (77%) were accurate to within one year of the age at which they will receive the State Pension. An equal proportion (26%) expect to receive the State Pension at an earlier or later age than currently legislated for.

• Awareness of deferral is mixed. Most people knew it was possible to defer the State Pension, but awareness of specific details such as the weekly increment and lump sum was lower. Awareness of these details was similar to awareness of the National Insurance exemption for those working past State Pension age.

• Expecting to remain in work at State Pension age was the main factor driving the intentions of those who expect to defer. Among the 10% of the sample of respondents who indicated that they intend to defer their State Pension, the main reason given was expecting to still be in work. Respondents who expected to retire after 67 were 17 percentage points more likely to report intended to defer than those who were already retired.

• In a hypothetical scenario test a lump sum payment appeared the be more attractive than a weekly increment. When faced with a hypothetical choice between deferring State Pension for a higher weekly payment or a lump sum, there was a stronger preference for the lump sum regardless of the generosity of the increment. However, this hypothetical approach is limited and does not constitute a certain indication of the likely behavioural response should the lump sum be reintroduced. Further research would be required to make conclusive statements about public preferences for the lump sum and weekly increment.

These findings are limited to a unique cohort of 54 to 64 year olds included in the research. Further research would be required to identify whether these findings apply to the broader population. A nationally representative survey of attitudes to later life examining how a broader sample of ages plan and prepare for later life, along the lines of an expanded and enhanced Attitudes to Pension 2012 survey, would be required in order address this requirement.

Appendices

Appendix A: Technical details of the survey

This appendix summarises the methodology, questionnaire design, sampling and weighting used by the Kantar TNS face-to-face Omnibus survey.

Methodology

Fieldwork took place across ten waves of the Omnibus from the 14 March to the 11 May 2018. The sample consisted of a representative sample of 2,535 respondents in Great Britain who satisfied the eligibility criteria.

The research was conducted through the regular Omnibus survey which uses face-to-face interviews, employing face-to-face Computer Assisted Personal Interviewing (CAPI), and selects respondents using a random location sampling method (described below).

Interviewing is conducted by professional interviewers who work exclusively for Kantar TNS. Assignments are conducted over two days of fieldwork and are carried out weekday 2pm to 8pm and at the weekend.

Questionnaire

A module of 45 questions were included with questions from existing surveys, where the questions are likely to have been extensively developed and piloted, used where possible. Questions were taken from Understanding Society, the Wealth and Assets Survey, English Longitudinal of Ageing, Attitudes to Pensions Survey and a previous State Pension deferral survey commissioned by the Department for Work Pensions.

A pilot study ran from the 21 to 25 February 2018 and included 305 respondents aged 54 to 64 years. Interviewers who carried out the pilot survey were asked to complete a feedback questionnaire and telephone interviews with a selection of interviewers were conducted. The questionnaire was altered to reduce participant burden, simplify questions, and ensure exhaustive and mutually exclusive answers.

As an Omnibus is a multi-client survey they consist of questions commissioned by more than one client within the same survey. Respondents were made aware that this module of questions were asked on behalf of the Department for Work and Pensions. An information card to find additional information at the UK State Pension was provided to respondents at the end of the interview.

Sampling

The Kantar TNS Omnibus sample was designed to be representative of the adult population in Great Britain in terms of various demographic characteristics and geographical location.

The sample design involves dividing Great Britain into 600 sampling points of similar population sizes, using the 2001 Census small area statistics and Postcode Address File. A total of 415 sample points (Primary Sample Units [PSU]) were selected for use by the omnibus after stratification by Region and Social Grade.

The sample was drawn in two stages: at the first stage sample points (Primary Sampling Units [PSUs]) were selected probability proportionate to size to the population; at the second stage, blocks of addresses were selected at random within the sampled PSUs.

The number of sampling points issued varies between 143 and 208 each week, depending upon the length of the questionnaire. 10 to 15 interviews are conducted in each area, depending upon the questionnaire length.

Interviewers aim to fill quotas within the randomly selected blocks of addresses. Quotas are set by gender; for females’ presence of children and working status are collected and for males working status is collected to ensure a balanced sample of adults within effective contacted addresses. All interviewers must leave 3 doors between each successful interview.

Due to the quality provided by its representative coverage of the population Kantar TNS Capi Omnibus have conducted similar surveys for regulatory bodies and government departments and agencies. Examples include the Food Standard Agency (the FSA) Public Attitudes and Food Hygiene Rating System.^{[footnote 24]}

Response rates

The Omnibus uses a Random Location sample and as such response rates are not calculated for the survey.

Weighting

Results were weighted at the analysis stage to be representative of the Great Britain adult population aged 54 to 64 years. Data were weighted for the following characteristics: sex, age, social grade, region and working status.

The weighting matrix targets were sourced from October to December 2017 Target Group Index Survey (TGI). TGI is the longest established marketing and media survey in GB and surveys around 25,000 adults annually.

Kantar TNS calculated the weighting efficiency of the sample, which is a percentage giving indication of the amount of skewing that had to be done to get the weights to converge. The closer this figure is to 100% the less the underlying data needed to be altered.

The weighting efficiency for this study was 88%. This can be considered high, meaning the results can be viewed as more reliable; a higher weighting efficiency means a lower margin of error due to the accuracy of the sampling method and low sampling error.

Non-response weights cannot be produced due to the random location sampling design used in the survey. However differentiating response rates within different segments of the population are taken account of by weighting the sample back to the population.

Appendix B: Statistical methods and terms used in the survey

A basic outline of the statistical techniques used in the report is presented here.

Statistical significance

All samples are a subset of a larger ‘population’. No sample precisely reflects the population it represents. Statistical significance is a technical concept that states whether or not an estimated value is likely to have arisen only from random variations in the sampling. It is most often used when talking about a change or a difference: a significant change or difference is one that is not likely to be due only to the sampling, and therefore is likely to reflect a real change or difference in the population. All of the changes have been tested at the 95% significance level and are statistically significant unless specifically stated.

Strictly speaking standard errors and statistical significance are only applicable to probability samples and are not applicable to the random location design adopted for this survey. However simple random sampling, where each individual in the population has an equal chance of selection, is rarely used in practice. Most government funded surveys have more complex designs, which typically involve sampling clusters (groups) of the population and then then taking a random sample of individuals within the selected clusters.

It is widely assumed that the variance of a random location sample is similar to that of an equally specified probability sample. As such the data has been analysed as if it were a probability sample, in line with how other public agencies have analysed this data.^{[footnote 25]} In such cases, the significance testing is intended to provide a rough guide to the margins of error governing the survey.

Binary Logistic Regression

Regression analysis estimates the associated between an outcome (dependent variable) and one or more explanatory independent variables. These independent variables can include categories (such as male or female) or numerical such as age.

Binary Logistic regression is a specific form of regression appropriate for when the dependent variable is binary (consists of only two possible values, for instance whether the respondent knows or does not know their State Pension age).

Multiple regression analysis is useful in determining the variables that have the strongest combined relationship with a single variable. While this analysis provides an indication of the association between these variables, this analysis does not support causal inference. Specific experimental research design, rather than observational survey data, would be required to substantiate these claims.

The reported logistic regression output (reported in Appendix C and D) provides three measures of the association between variables, the Logit (log odds), Odds Ratio and Average Marginal Effect.

The odds ratio is derived from the logit and represent the number by which the odds of being in the category of the dependent variable being modelled would multiply for each one-unit increase in the independent variable. An odds ratio greater than 1 indicates that the odds increase for a one-unit increase in the independent variable (a positive association), where an odds ratio less than 1 indicates that the odds of being in the category of the dependent variable decrease for a one-unit increase in the independent variable (a negative association).

As odds ratios have a complex interpretation Average Marginal Effects are presented and reported. This summarises the average effect of a variable on the probability of being in the category of the dependent variable in a single coefficient. The Average Marginal Effect indicates the percentage point difference for a one-unit increase in the independent variable.

Logistic regression also tests for the statistical significance of parameter estimates. The Wald statistic is used to determine if the parameter estimated is statistically significant from zero (no effect). The number of asterisks indicates whether the parameter is statistically significant from zero at the 10%, 5% and 1% level.

The overall explanatory power of the model is provided by two statistics, the adjusted R-Square and the Log Likelihood Ratio. The Log Likelihood ratio is an indicator of how much unexplained variation exists once the model has been fitted. By comparing the Log Likelihood of the fitted model to a baseline model with only the intercept provides an estimate of the overall explanatory power of the model. Each time a variable is added the logistic regression model is calculated with and without the new variable to determine the predictive power added by the new variable.

The Adjusted R-Squared is also reported, which is a measure of the total explanatory power of the model, with a value of 0 indicating the model does not predict the dependent variable and 1 indicating the model perfects predicts the dependent variable.

The Hosmer & Lemeshow goodness of fit test provides an indication of how well the data fits the model. A non-significant results indicates that the models predictions do not differ significantly from the observed data.

Appendix C: Modelling awareness of State Pension age

Dependent variable: Exact Knowledge of State Pension age (1=Yes)

Independent variables	Independent variables	Logit	z	Log Odds	Average Marginal Effect
Actual State Pension age (reference group: 65)	65-66	-2.83	-6.38***	0.059	-0.571
Actual State Pension age (reference group: 65)	66	-0.8	-1.91*	0.451	-0.16
Actual State Pension age (reference group: 65)	66-67	-2.3	-4.6***	0.1	-0.464
Actual State Pension age (reference group: 65)	67	-0.05	-0.1	0.951	-0.010
Self-reported State Pension Knowledge (Reference group: know little or nothing about State Pension)	Good knowledge	0.61	3.69***	1.848	0.124
Self-reported State Pension Knowledge (Reference group: know little or nothing about State Pension)	Reasonable, basic knowledge	0.46	3.02***	1.584	0.093
Self-reported State Pension Knowledge (Reference group: know little or nothing about State Pension)	Very patchy knowledge	0.21	1.28***	1.228	0.041
Number of State Pension information sources consulted in last 12 months	-	0.19	3.83***	1.206	0.038
Expects to receive considerable Private Pension income in retirement	-	0.2	2.07**	1.225	0.041
Age	-	0.15	4.41***	1.161	0.030
Female	-	-0.06	-0.63	0.940	-0.013
Ethnicity (reference group: White)	Asian or Asian British	-0.7	-2.13**	0.498	-0.140
Ethnicity (reference group: White)	Black, African or Caribbean	-1.01	-3.27***	0.363	-0.204
Ethnicity (reference group: White)	Mixed or Multiple Ethnic Groups	-0.04	-0.08	0.957	-0.009
Ethnicity (reference group: White)	Other	-2.04	-1.61	0.13	-0.412
Economic Activity (reference group: Employed full time (>30 hours))	Employed part time (<30 hours)	-0.08	-0.61	0.924	-0.016
Economic Activity (reference group: Employed full time (>30 hours))	Self employed	0.06	0.3	1.058	0.011
Economic Activity (reference group: Employed full time (>30 hours))	Unemployed	0.07	0.31	1.071	0.014
Economic Activity (reference group: Employed full time (>30 hours))	Sick or disabled	-0.23	-0.87	0.796	-0.046
Economic Activity (reference group: Employed full time (>30 hours))	Retired from paid work	-0.04	-0.26	0.964	-0.007
Economic Activity (reference group: Employed full time (>30 hours))	Looking after the home, family or caring	-0.32	-1.2	0.729	-0.064
Social grade (reference group: C1)	A	-0.58	-2.4**	0.560	-0.117
Social grade (reference group: C1)	B	0.05	0.36	1.048	0.009
Social grade (reference group: C1)	C2	0.04	0.31	1.041	0.008
Social grade (reference group: C1)	D	0.06	0.37	1.057	0.011
Social grade (reference group: C1)	E	-0.48	-2.21**	0.226	-0.096
Tenure (reference group: Own outright)	Own- with mortgage or loan	-0.11	-0.91	0.897	-0.022
Tenure (reference group: Own outright)	Rent privately	-0.54	-3.14***	0.584	-0.108
Tenure (reference group: Own outright)	Rent from Local Authority/Housing Association	-0.5	-3.41***	0.606	-0.101
Tenure (reference group: Own outright)	Other	-0.89	-1.61*	0.410	-0.180
Self-reported health status (reference group: Excellent)	Very good	0.15	0.97	1.165	0.031
Self-reported health status (reference group: Excellent)	Good	0.17	1.11	1.186	0.034
Self-reported health status (reference group: Excellent)	Fair	0.04	0.24	1.042	0.008
Self-reported health status (reference group: Excellent)	Poor	-0.38	-1.49	0.684	-0.077
Marital status (reference group: Married / Living as married)	Single	0.51	3.67***	1.664	0.103
Marital status (reference group: Married / Living as married)	Widowed, Divorced or Separated	0.12	0.94	1.128	0.024
Has children	-	0.02	0.09	1.016	0.003
Government Office Region (reference group: London)	North East	0.39	1.56	1.474	0.078
Government Office Region (reference group: London)	North West	0.05	0.27	1.055	0.011
Government Office Region (reference group: London)	Yorkshire & the Humber	-0.06	-0.28	0.943	-0.012
Government Office Region (reference group: London)	East Midlands	0.51	2.33**	1.669	0.103
Government Office Region (reference group: London)	West Midlands	0.33	1.55	1.385	0.066
Government Office Region (reference group: London)	East of England	0.02	0.15	1.024	0.005
Government Office Region (reference group: London)	South East	-0.15	-0.75	0.861	-0.030
Government Office Region (reference group: London)	South West	0.12	0.58	1.122	0.023
Government Office Region (reference group: London)	Wales	0.46*	1.93*	1.585	0.093
Government Office Region (reference group: London)	Scotland	0.12	0.59	1.128	0.024
Intercept		-8.96

NOTE: complete case analysis based on 2,426 observations. Significance: * significant at 0.1 level, ** at 0.05 level, *** at 0.01 level

-2 Log Likelihood: 2,924, x2=422 , df=47, p<.001 Nagelkerke R2 = 21.37%

Hosmer & Lemeshow Test p=0.2686

Classification accuracy: 65.5%

Appendix D: Modelling deferral intentions

Dependent variable: Intends to defer the State Pension upon reaching State Pension Age (1=Deferral)

Independent variables	Independent variables	Logit	z	Log Odds	Average Marginal Effect
Expected retirement age once State Pension age is known (reference group: Already retired)	Below 60	-0.21	-0.31	0.81	-0.016
Expected retirement age once State Pension age is known (reference group: Already retired)	60 to 64	-0.16	-0.33	0.85	-0.012
Expected retirement age once State Pension age is known (reference group: Already retired)	65	0.17	0.35	1.18	0.013
Expected retirement age once State Pension age is known (reference group: Already retired)	66	0.39	0.94	1.48	0.029
Expected retirement age once State Pension age is known (reference group: Already retired)	67	1.8	4.3***	6.02	0.134
Expected retirement age once State Pension age is known (reference group: Already retired)	Greater than 67	2.3	5.57***	9.93	0.171
Actual State Pension age (reference group: 65)	65-66	-0.96	-1.28	0.38	-0.072
Actual State Pension age (reference group: 65)	66	-1.19	-1.64*	0.3	-0.089
Actual State Pension age (reference group: 65)	66-67	-2.17	-2.48**	0.11	-0.162
Actual State Pension age (reference group: 65)	67	-2.21	-2.42**	0.11	-0.165
Self-report State Pension knowledge (reference group: know little or nothing about State Pension)	Good knowledge	0.09	0.27	1.1	0.007
Self-report State Pension knowledge (reference group: know little or nothing about State Pension)	Reasonable, basic knowledge	0.32	1	1.38	0.024
Self-report State Pension knowledge (reference group: know little or nothing about State Pension)	Very patchy knowledge	0.66	2.07**	1.94	0.049
Number of State Pension information sources consulted in last 12 months	-	0.01	0.08	1.01	0.001
Has thought about how long will be retired for	-	0.26	1.24	1.3	0.019
Expected retirement duration (reference group: Less than 10 years)	Between 10 and 14 years	-0.03	-0.09	0.97	-0.002
Expected retirement duration (reference group: Less than 10 years)	Between 15 and 19 years	-0.03	-0.09	0.97	0.002
Expected retirement duration (reference group: Less than 10 years)	Between 20 and 24 years	0.35	1.05	1.41	0.026
Expected retirement duration (reference group: Less than 10 years)	Between 25 and 29 years	-0.16	-0.4	0.85	-0.012
Expected retirement duration (reference group: Less than 10 years)	Between 30 and 34 years	-0.64	-1.2	0.53	-0.048
Expected retirement duration (reference group: Less than 10 years)	35 years plus	0.41	0.86	1.51	0.03
Expected retirement duration (reference group: Less than 10 years)	Don’t know	0.02	0.04	1.020	0.001
Expects to receive considerable income from Private Pension	-	-0.87***	-4.24***	0.42	-0.065
Would prefer to receive £1,100 in one year’s time than £1,000 today	-	0.76	3.99***	2.130	0.056
Age	-	-0.14	-1.96**	0.87	-0.010
Female	-	-0.01	-0.07	0.99	-0.001
Ethnicity (reference group: White British)	Asian or Asian British	-0.77	-0.87	0.46	-0.057
Ethnicity (reference group: White British)	Black, African or Caribbean	0.36	0.7	1.43	0.027
Ethnicity (reference group: White British)	Other	3.14	2.54**	23.01	0.234
Social grade (reference group: C1)	A	0.54	1.21	1.72	0.041
Social grade (reference group: C1)	B	-0.21	-0.71	0.81	-0.016
Social grade (reference group: C1)	C2	-0.02	-0.08	0.98	-0.002
Social grade (reference group: C1)	D	-0.02	-0.07	0.98	-0.001
Social grade (reference group: C1)	E	-1.08	-2.44**	0.34	-0.080
Tenure (reference group: Own outright)	Own- with mortgage or loan	0.9	3.68***	2.47	0.067
Tenure (reference group: Own outright)	Rent privately	0.51	1.51	1.55	0.038
Tenure (reference group: Own outright)	Rent from Local Authority/Housing Association	1.25	4.44***	3.5	0.093
Tenure (reference group: Own outright)	Other	0.75	0.83	2.11	0.056
Household income - non equivalised (reference group: £10,000 to £19,999)	Under £10,000	0.78	2.33**	2.18	0.058
Household income - non equivalised (reference group: £10,000 to £19,999)	£20,000 to 29,999	0.22	0.72	1.25	0.017
Household income - non equivalised (reference group: £10,000 to £19,999)	£30,000 to 39,999	0.48	1.43	1.62	0.036
Household income - non equivalised (reference group: £10,000 to £19,999)	£40,000 to 49,999	0.54	1.43	1.72	0.041
Household income - non equivalised (reference group: £10,000 to £19,999)	£50,000+	0.38	1	1.46	0.028
Self-reported health status (reference group: Excellent)	Very good	-0.44	-1.46	0.65	-0.033
Self-reported health status (reference group: Excellent)	Good	-0.85	-2.85***	0.43	-0.063
Self-reported health status (reference group: Excellent)	Fair	-1	-2.84***	0.37	-0.075
Self-reported health status (reference group: Excellent)	Poor	-1.4	-2.88***	0.25	-0.105
Marital status (reference group: Married / Living as married)	Single	-0.23	-0.76	0.8	-0.017
Marital status (reference group: Married / Living as married)	Widowed, Divorced or Separated	0.34	1.33	1.4	0.025
Has children	-	-0.92	-2.29**	0.4	-0.069
Government Office Region (reference group: London)	North East	-1.44	-1.76*	0.24	-0.107
Government Office Region (reference group: London)	North West	0.3	0.82	1.350	0.022
Government Office Region (reference group: London)	Yorkshire & the Humber	-0.15	-0.34	0.86	-0.011
Government Office Region (reference group: London)	East Midlands	-0.91	-1.79*	0.4	-0.068
Government Office Region (reference group: London)	West Midlands	-0.76	-1.59	0.47	-0.057
Government Office Region (reference group: London)	East of England	0.32	0.84	1.380	0.024
Government Office Region (reference group: London)	South East	-0.67	-1.74*	0.510	-0.05
Government Office Region (reference group: London)	South West	0.03	0.09	1.03	0.003
Government Office Region (reference group: London)	Wales	0.15	0.29	1.16	0.011
Government Office Region (reference group: London)	Scotland	-0.38	-0.95	0.68	-0.029

NOTE: complete case analysis based on 1,523 observations. Significance: * significant at 0.1 level, ** at 0.05 level, *** at 0.01 level

-2 Log Likelihood: 2,924, x2=275 , df=61, p<.001 Nagelkerke R2 = 31.79%

Hosmer & Lemeshow Test p=0.4734

Classification accuracy: 89%

1. Office for National Statistics (2018). Overview of the UK population: November 2018. Available at: https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/ articles/overviewoftheukpopulation/november2018 ↩

2. Department for Work and Pensions (2018). Benefit expenditure and caseload tables 2018. Available at: https://www.gov.uk/government/publications/benefit-expenditure-and-caseload-tables-2018 ↩

3. Cridland, J. (2017). Independent Review of the State Pension age - Smoothing the Transition. Available at: https://www.gov.uk/government/publications/state-pension-age-independent-review-final- report ↩

4. Department for Work and Pensions (2018). State Pension age review. Available at: https://www.gov.uk/government/publications/state-pension-age-review-final-report ↩

5. Department for Business, Energy and Industrial Strategy (2018). Removal of the Default Retirement Age, Post Implementation Review. Available at: http://www.legislation.gov.uk/uksi/2011/1069/pdfs/uksiod_20111069_en.pdf ↩

6. Department for Work and Pensions (2017). Fuller Working Lives: a partnership approach. Available at: https://www.gov.uk/government/publications/fuller-working-lives-a-partnership-approach ↩

7. Coleman et al (2008). State Pension deferral: public awareness and attitudes. Available at: nationalarchives.gov.uk ↩

8. MacLeod et al (2012). Attitudes to Pensions: The 2012 survey. Available at: Attitudes to Pensions: The 2012 survey ↩

9. Powell and Bonito (2015). Attitudes of the over 50s to fuller working lives. Available at: https://www.gov.uk/government/publications/attitudes-of-the-over-50s-to-fuller-working-lives ↩

10. Food Standards Agency (2018). Biannual Public Attitudes Tracker Wave 16, May 2018. Available at: https://www.food.gov.uk/sites/default/files/media/document/biannual-public-attitudes-tracker-wave- 16-final-270718_1.pdf ↩

11. Department for Work and Pensions (2018). Economic labour market status of individuals aged 50 and over, tends over time: October 2018. Available at: https://www.gov.uk/government/statistics/economic-labour-market-status-of-individuals-aged-50-and- over-trends-over-time-october-2018 ↩

12. Cribb et al (2013). Incentives, shocks or signals: labour supply effects of increasing the female state pension age in the UK. Available at: https://www.ifs.org.uk/publications/6622 ↩

13. This figure excludes 39 respondents who ‘do not intend to retire’. ↩

14. O’Dea and Sturrock (2018). Subjective expectations of survival and economic behaviour. Available at: https://www.ifs.org.uk/publications/12904 ↩

15. MacLeod et al (2012). Attitudes to Pensions: The 2012 survey. Available at: Attitudes to Pensions: The 2012 survey ↩

16. 25.57% expected to receive the State Pension later than legislated, compared to 26.49% who expected to receive it earlier. ↩

17. Coleman et al (2008). State Pension deferral: public awareness and attitudes. Available at: nationalarchives.gov.uk ↩

18. Department for Work and Pensions (2018). Income-related benefits: estimates of take-up in 2016 to 2017. Available at: https://www.gov.uk/government/statistics/income-related-benefits-estimates-of- take-up-financial-year-2016-to-2017 ↩

19. MacLeod et al (2012). Attitudes to Pensions: The 2012 survey. Available at: Attitudes to Pensions: The 2012 survey ↩

20. O’Dea and Sturrock (2018). Subjective expectations of survival and economic behaviour. Available at: https://www.ifs.org.uk/publications/12904 ↩

21. Cridland, J. (2017). Independent Review of the State Pension age - Smoothing the Transition. Available at: https://www.gov.uk/government/publications/state-pension-age-independent-review-final- report ↩

22. Kahneman and Tversky, A. (1979). Prospect theory: Analysis of decision under risk. Econometrica, 47, 263–291. ↩

23. Mather, M et al (2012). Risk preferences and aging: The “Certainty Effect” in Older Adults’ Decision making. Psychology and Ageing, 27, 801-816. ↩

24. Food Standards Agency. (2018). Biannual Public Attitudes Tracker Wave 16, May 2018. Available at: https://www.food.gov.uk/sites/default/files/media/document/biannual-public-attitudes-tracker-wave- 16-final-270718_1.pdf ↩

25. Food Standards Agency. (2018). Biannual Public Attitudes Tracker Wave 16, May 2018. Available at: https://www.food.gov.uk/sites/default/files/media/document/biannual-public-attitudes-tracker-wave- 16-final-270718_1.pdf ↩