Wealth and Assets Survey guide

The Wealth and Assets Survey (WAS) was launched in 2006 and is a biennial longitudinal survey conducted by the Office for National Statistics (ONS). This survey measures the well-being of households and individuals in terms of their assets, savings, debt and planning for retirement. Data from this longitudinal survey provides users with the ability to measure changes of wealth in Great Britain (GB) over time.

The WAS is sponsored by a funding consortium, including:

• the ONS, HM Revenue and Customs, Scottish Government and the Department for Work and Pensions up to Round 8

• the ONS, HM Revenue and Customs, and Scottish Government for Rounds 9 onwards

We report transparently on changes that affect quality and comparability, including response rate and achieved sample. 

Accredited official statistics status of the Wealth and Assets Survey core outputs has been suspended from Round 8, covering the period 2020 to 2022 onwards while we undertake further work to improve quality, in line with the Office for Statistics Regulation (OSR) Assessment Report 396 (June 2025) (PDF, 237KB).

For more information about the statistical designation of the survey, go to Quality of the survey.

We use data from this survey in the following publications:

• Household total wealth in Great Britain

• Impact of increased cost of living on adults across Great Britain

• Saving for retirement in Great Britain

• Distribution of individual total wealth by characteristic in Great Britain

• Household debt in Great Britain

We updated this guide on 27 March 2026. Important changes include:

• transformation of the Wealth and Assets Survey (WAS) quality and methodology information (QMI) report into two clearer and more purpose-specific publications – a Household total wealth in Great Britain quality and methods guide and this Wealth and Assets Survey guide

• improved transparency around data quality concerns, for example, response rates and achieved sample

These are material improvements, made in line with Office for Statistics Regulation (OSR) recommendations, to strengthen transparency and ensure users can more easily understand the quality and limitations of the survey.

Data collection method

The Wealth and Assets Survey (WAS) is a biennial longitudinal survey conducted by the Office for National Statistics (ONS). We typically carry out interviews face to face using Computer Assisted Interviewing (CAI).

While in the field, the survey is referred to as the "Household Assets Survey" (HAS). We use this name with participants because the term "wealth" was thought to have negative connotations for individuals who do not identify as someone who has wealth and therefore could potentially discourage participation.

During the coronavirus (COVID-19) pandemic, we carried out interviews by telephone using a shortened questionnaire to maintain participation. These changes to mode and instrument affect comparability for some topics.

The WAS is a longitudinal survey, and attrition naturally occurs between waves. To compensate for attrition between Waves 1 and 2, we introduced a new cohort of addresses at Wave 3 and in each subsequent collection. New cohorts help reduce some of the bias generated by attrition. We select participants from the current population to reflect changes in population characteristics over time, while maintaining the size of the cross-sectional sample.

Sample design

Sampling for the WAS is conducted using a two-stage approach: first selecting postcode sectors, then choosing addresses within those sectors.

Postcode sectors, which are taken from the Postcode Address File (PAF), are used as the units to sample from. These postcode sectors are organised in a list sorted by NUTS level 2 geographies, which are broad sub-regional areas used for statistical purposes. The Primary Sample Units (PSUs) are then selected from this list.

Typically, within each of these postcode sectors, 26 addresses are randomly selected with addresses listed by postcode and street number. Selected addresses are then split into two quotas of 13 addresses to ease the management of fieldwork.

With the distribution of household wealth highly skewed, households likely to be in the top percentile of the Great Britain (GB) wealth distributions are now oversampled by a factor of five, and households in the 2nd to 10th percentile by a factor of three. These addresses are identified using data supplied by HM Revenue and Customs (HMRC). The ONS provides the PSU data to HMRC, who create a wealth indicator using administrative information on financial assets, property income, capital gains, partnerships, Individual Savings Accounts (ISAs), and pensions. Individuals ranked in the top 10% of this distribution are then matched to UK addresses, and the postcode sectors sent from the ONS are merged into the wealth distribution dataset before being delivered back to the ONS for oversampling.

Oversampling helps compensate for lower response rates among high wealth households and improves the precision of wealth estimates at the top of the distribution.

Survey interviews are conducted evenly over the year, to help ensure estimates are not biased by seasonal variations.

"Keep-in-Touch Exercise" interviews are undertaken four months before follow-up mainstage interviews to help maintain the accuracy of contact details and encourage further participation. This reduces loss of participants and improves longitudinal consistency.

Table 1 shows the achieved sample size for the WAS from Wave 1 to the latest completed round. Wave 6 of the WAS was only in the field for 21 months, because of the move of the WAS to financial years (rounds). The number of households interviewed for Wave 6 was approximately 16,000. This was combined with the final three months of the preceding Wave 5 to produce an 18,000 household Round 6 file.

As response rates have declined and attrition has increased, the proportion of the achieved sample derived from households being re-interviewed has diminished.

The fieldwork for Round 8 of the WAS took place during the coronavirus pandemic where, for much of the time, national lockdown conditions and social distancing restrictions were in place. As such, interviewing took place over the telephone for the whole of Round 8. For further details, go to our Impact of COVID-19 on ONS social survey data collection methodology.

Response rates

Response rates are reported internally on a monthly basis and are based on the number of fully and partially co-operating households as a proportion of the numbers of eligible households in the sample. For Wave 1 to Round 9, the overall response rates and the breakdown for new and old cohorts are included in Table 2.

Changes to collection methods in response to the coronavirus pandemic prompted larger issued sample sizes to compensate for lower expected contact and response rates to help maintain achieved sample sizes. For further information, go to our Impact of COVID-19 on ONS social survey data collection methodology.

Table 2 demonstrates a long-term decline in response rates with Round 8 (2020 to 2022) overall response at 41% (down from 66% in Wave 4).

Lower response rates tend to mean that wealth estimates for more granular sub-groups of the population will be less precise (they will have a wider margin for error), and in some cases they cannot be published at all owing to the risk of drawing incorrect conclusions from them. We produce quality indicators, which show confidence intervals around our subgroup total analysis to show the greater level of uncertainty with these estimates.

Coverage

Time periods

Data are available for:

• biennial "waves" (two-year July to June periods) from July 2006 to June 2008 until July 2014 to June 2016

• biennial "rounds" (two-year financial year periodicity) from April 2016 to March 2018 onwards

The shift from waves to rounds aligns the survey period with the UK financial year, improving comparability with other household finance surveys.

The quality indicators datasets have been published for and cover all waves and rounds from July 2006 to March 2022. 

Geographic coverage 

The data covers Great Britain with some exclusions given fieldwork feasibility constraints. All WAS data prior to Round 9 cover Great Britain excluding addresses north of the Caledonian Canal, the Scottish Islands and the Isles of Scilly. From Round 9 onwards, the WAS will cover Great Britain excluding the Isles of Scilly.

Demographic coverage

The survey collects key classificatory details (such as age, sex, and employment status), which are used to support weighting, imputation, and analysis across major demographic and socio-economic groups.

Exclusions

The WAS excludes individuals in communal establishments and certain geographically remote locations, detailed in the geographic coverage section of this guide. Some asset types that are hard to measure in household surveys (for example, specific offshore assets and some trust structures) are likely to be under captured. Pensions wealth estimates in the WAS cover private pension wealth only and therefore exclude State Pension entitlement. Although the WAS collects information on an individual's business assets, these items are not currently included in the total wealth measure. This is because the data collected for these components is not yet sufficiently developed for consistent valuation and would require further methodological work before inclusion in headline estimates. Additionally, ownership of private business assets raises distinct conceptual and valuation challenges that differ from other household assets. Users should consider this when interpreting the very top of the wealth distribution.

Processing survey responses

Editing

An extensive range of validation checks and computer edits are applied to both the household and individual questionnaires during the computer-assisted interview (CAI) and to the consolidated monthly data. These checks include range checks, logical consistency checks, cross-module consistency checks, and validation of skip patterns (these checks are outlined within the How we quality assure the survey section). To aid consistent editing, validation and imputation of collected data, responses to questions with associated period code questions are annualised and "code all that apply" questions are converted to binary format.

Once each stage of editing is complete, checks are run to ensure the correct survey routing has been maintained. This is crucial to ensure the data feeding into imputation is consistent with the intended questionnaire structure, free from routing errors.

Imputation

Missing data is approached in two stages: first, a deductive imputation method, followed by a statistical method.

Deductive imputation is applied where a missing or inconsistent value could be deduced with certainty.

Statistical imputation uses a nearest-neighbour method where information from a donor record (that had no errors or missing values) is used to replace the missing values for a recipient record. In this approach, a donor is selected from a pool of potential donors with similar characteristics based on conditional probabilities for key demographic and wealth variables.

For longitudinal households, if a value is observed in one wave or round but missing in another, we impute the missing value by selecting a donor case that has similar characteristics. When imputing, we seek to maintain relationships between waves. Therefore, the imputed value takes into account the respondent's observed value in the previous wave and will select a donor that responded similarly to the recipient in the previous wave. Where there is no response in the previous wave, we rely on other variables that have strong relationships between the variable we wish to impute.  

The imputation is conducted under edit constraints to ensure that outliers and implausible relationships are not introduced during the process.

Post-imputation diagnostics compare pre- and post-distributions to preserve plausible relationships. For further details on imputation, go to How we quality assure the survey. 

Deriving analysis variables

The imputed survey variables are then used to create a series of derived variables that are suitable for analysis. We produce derived variables for each component of wealth, followed by variables that capture both household and individual total wealth. We create additional derived variables for personal and household income and problem debt estimates. You can find specifications for these variables on the UK Data Service (UKDS). These derived variables are quality assured, as outlined in the Quality of the survey section of this guide.

Weighting

The WAS follows a three-stage weighting procedure. First, a design weight is constructed, equal to the reciprocal of the address selection probability. Secondly, a non-response weight is created to reduce potential non-response bias. The non-response model currently includes region, a socio-economic indicator and the HM Revenue and Customs (HMRC)-provided wealth index used to identify the wealthiest households. This applies to a new cohort. In older panels, an attrition adjustment was applied, and joiners were incorporated, before calibration.

The final stage of the weighting procedure calibrates the adjusted weights from the first stages to known population totals present at the time of the fieldwork period. These estimates are now derived from Census 2021 for England and Wales and the 2022 Census for Scotland. Different sets of weights are created so that analysis can be performed both longitudinally (person-level) and cross-sectionally (household-level and person-level).

Calibration is performed at both household and person levels depending on the analytical file, and weighting mitigates (though it does not eliminate) bias from differential response among demographic and wealth groups.

From Round 8, tenure is included in calibration to address response imbalances during the coronavirus pandemic, when renters were under-represented and households who own outright were overrepresented. Calibration totals were also updated using Census 2021 for England and Wales and the 2022 Census for Scotland. These adjustments help improve representativeness, but weighting cannot fully correct sample imbalances, where response patterns differ substantially across regions.

Statistical designation

The Wealth and Assets Survey's (WAS's) core outputs are "Official Statistics". Accredited official statistics status of these core outputs, including Household total wealth in Great Britain, has been suspended from Round 8, covering the period 2020 to 2022 onwards while we undertake further work to improve quality, in line with the Office for Statistics Regulation (OSR) Assessment Report 396 (June 2025) (PDF, 237KB).

How we quality assure the survey

Quality Assurance is applied through collection, initial processing of extracted data, statistical imputation and the preparation of analytical data for publication.

Post data collection checks

For the post-collection editing process, a series of checks are applied to survey data collected for each household and individual interviews on the survey.

Cases and responses are forensically reviewed by editors within spreadsheets and completed questionnaires. Editors record their rationale for making or not making a change based upon the supporting evidence available post-collection. Suggested changes are independently reviewed by an editor manager, who either signs them off or makes further amendments before they are applied to the data and prepared for statistical use.

Issues identified during processing are reviewed to identify changes that may need to be applied to the questionnaire.

The types of checks we conduct at this stage are listed here.

Personal identifier checks

Personal identifiers are reviewed to ensure that they are consistent with past rounds and accurate. These are then required for sample preparation and linkage, and cross-round comparison of responses.

Missing and open-ended response checks

Some unanswered questions are reviewed to determine whether they can be corrected based on evidence within the questionnaire or through external sources.

 This includes:

• unanswered Council Tax questions that can be edited using other sources

• unanswered period code questions and where other questions may provide evidence of the answer

• recorded relationships, for example, marital status

The editors will also review all responses to open questions and back-code to existing answer categories if applicable.

Review of interviewer comments

Interviewers can leave detailed comments about the data and the household while conducting the interview. These can explain context and detail that may not have been captured within the questionnaire. These are checked and edits to the survey data are made to ensure an accurate picture of the household is captured.  

Routing checks

The routing for the questionnaire ensures that respondents receive the appropriate questions based on their previous answers in the interview. Editing during and after the interview can affect the coherence of this routing and so data are checked to ensure the routing rules of the questionnaire have been maintained through each stage of the editing process.  

Consistency checks

Responses on the household survey are checked for consistency where answers to questions have a relationship with one another (for instance someone's gross pay is expected to be higher than their net pay). Checks are also performed across the household to ensure items that can be held jointly have not been double counted.

Outlier checks

From Round 9, outlier thresholds are determined by examining the distribution of each variable, considering the nature of that variable and the spread of values (pre-set thresholds were used based on the variable in question prior to Round 9). A proportion of the highest and, where relevant, the lowest values are flagged as potential outliers. Not all variables have lower-end outliers; for example, low or zero values are valid for many financial account variables.

We check outliers for credibility by reviewing related information, including responses from previous waves, to determine whether there is evidence to support or amend the outlying value. This evidence includes the inspection of wealth through income, assets and debts, and verification from linked variables (such as comparing mortgage balances with monthly payments and remaining terms). Substantial longitudinal changes can be genuine where life events have occurred; for example, alterations to working status or household structure (such as a split in partnership, inheritance, a house move or a change in employment).

We amend outliers only when there is clear evidence that the recorded value is not plausible. Where the review indicates that the value may reflect a genuine change, the value is retained.

Longitudinal consistency checks

We use longitudinal checks to review a respondent's data across waves and rounds when values appear unusual, helping us determine whether changes reflect genuine life events or potential data issues. We also review aggregate estimates over time to identify inconsistencies that may be explained by questionnaire changes or methodological updates.

Where a methods change is the main difference (rather than a respondent level change), we document this in the Changes and their effects on comparability over time section of this guide.

This process helps maintain consistency and reliability in longitudinal estimates. Cross-wave and cross-round comparisons also support the identification of attrition-related anomalies and household composition changes.

Imputation checks

Missing or inconsistent values are addressed through a two-stage imputation process: deductive imputation where values can be logically inferred, followed by nearest-neighbour donor-based imputation for remaining gaps. Longitudinal history is used to inform donor selection for repeat households.

Post-imputation diagnostics compare pre- and post-imputation distributions and preserve observed relationships (for example, income-asset-debt ratios).

In Round 8, a total of 665 variables required some form of imputation (see Table 3 for further information on imputation rates). For most variables (62%), fewer than 10% of respondents required imputed values. For the remainder, broadly equal proportions (12% to 13%) required imputation of 10% to 20%, 20% to 40%, and over 40%. More details can be found in the WAS Round 8 user guide available on the UK Data Service (UKDS).

Analytical derived variable checks

Once survey data has been processed, it is used to calculate analytical derived variables. Checks are carried out on these variables to ensure they have been calculated correctly. Summary statistics are calculated and compared with the equivalent statistics from the previous round. While outliers are checked at earlier stages of the processing, some only become apparent once analytical derived variables are calculated. At this point, any significant ones are identified and investigated.

Strengths and limitations

Strengths

• The WAS provides a unique, single source of detailed information on household and individual wealth and assets in Great Britain, covering property, financial, physical and private pension wealth.

• The inclusion of private pension wealth, captured in depth at person level, is a major strength of the WAS, providing a more comprehensive picture of total household wealth than many comparable surveys, particularly internationally.

• The WAS collects high-level income data for wealth analysis purposes, and it is harmonised with the Living Costs and Food Survey, providing a unique picture of income, spending and wealth. 

• The WAS uses a transparent oversampling approach to improve coverage of high-wealth households, reducing bias relative to surveys that rely solely on general population sampling.

• The longitudinal design allows analysis of wealth accumulation, change and mobility over time, an uncommon feature among international wealth surveys.

• Revisiting respondents in subsequent waves improves internal consistency and enables cross‑round validation.

• Editing and imputation applied before dissemination improves accuracy, comparability and usability of the microdata.

• Data access through the UKDS and the Secure Research Service supports transparency, reproducibility and research use.

• Recent updates to documentation, including the introduction of the Wealth and Assets Survey guide and the Household total wealth in Great Britain quality and methods guide, enhance clarity around methods, quality and limitations, addressing areas for improvement highlighted in the Office for Statistics Regulation (OSR) Assessment Report 396 (June 2025) (PDF, 237KB).

Limitations 

• Self‑reported valuations of property and financial assets can introduce inaccuracies; WAS property values tend to be higher than market‑based sources, and users should take this into account when comparing with the UK House Price Index (HPI) or lender HPIs.

• Not all forms of wealth can be captured, including some informal debts, offshore holdings and certain trust structures, which is particularly relevant when analysing the top of the distribution.

• Defined Benefit (DB) pension wealth is valued using superannuation contributions adjusted for past experience (SCAPE)‑based assumptions introduced in Round 8, which prioritise stability; alternative valuation methods may be more suitable for some analytical purposes.

• The Round 8 DB pension methodology change introduces a discontinuity in the time series; comparisons of pensions and total wealth across Round 7 to 8 require caution.

• Coronavirus (COVID‑19) pandemic‑related changes in Round 8 (telephone mode, shortened questionnaire) reduce comparability for non‑core topics and may have introduced mode‑related measurement differences.

• Significant delays between data collection and publication reduce timeliness, such as the 34‑month lag for Round 8.

• The length and complexity of the questionnaire can contribute to respondent burden and increase risks of partial response or drop‑off.

• Declining response rates and attrition across waves and rounds reduce representativeness and longitudinal comparability, especially when differential non‑response affects younger renters or very high‑wealth households.

• Precision is lower for granular breakdowns, with London estimates showing higher volatility because of smaller sample sizes and greater non‑response.

• Despite oversampling, the highest‑wealth households and those with complex financial arrangements remain under-represented.

Data quality dimensions

The Government Data Quality Framework recommends the use of the Data Management Association (DAMA UK) data quality dimensions to assess and improve data quality. These are:

• completeness
• uniqueness
• consistency
• timeliness
• validity
• accuracy

We have integrated these considerations into the guide.

Latest changes

Improving the presentation of publications for users

On 27 March 2026, we implemented a set of material improvements in line with recommendations from the Office for Statistics Regulation (OSR). These changes included transforming the quality and methodology information (QMI) report for the Wealth and Assets Survey (WAS) into two clearer and more purpose‑specific publications:

• Household total wealth in Great Britain quality and methods guide

• Wealth and Assets Survey guide

This transformation improves transparency around data quality considerations, strengthens signposting, and provides clearer, more consistent definitions. In doing so, we aim to strengthen transparency and ensure users can more easily understand the quality and limitations of the survey.

Past changes

These changes are ordered by date, with the most recent first.

Round 9 updates to processing systems

Data collection for Round 9 is complete, with an achieved sample of 11,708 households. Work is underway to strengthen data processing, quality assurance and operational resilience, as set out in the Economic Statistics Plan and Survey Improvement and Enhancement Plan. This includes the introduction of a Reproducible Analytical Pipeline (RAP) to modernise internal processing, provide clearer specifications for derived variables, and support more consistent editing and quality checks.

These developments aim to improve transparency, reproducibility and efficiency in the production of WAS data. They do not affect the underlying concepts or definitions used in the survey.

Round 9 coverage and sampling updates

All WAS data prior to Round 9 cover Great Britain excluding addresses north of the Caledonian Canal, the Scottish Islands and the Isles of Scilly. From Round 9 onwards, the WAS will cover Great Britain excluding the Isles of Scilly as these addresses have been incorporated within the sample for Round 9 onwards. 

Round 9 questionnaire updates

Alongside routine updates to the questionnaire, Round 9 saw further developments to improve the questionnaire. The pension questions were restructured to provide clarity over Defined Benefit and Defined Contribution pensions and improve the questionnaire experience for respondents, remove any unnecessary or duplicated content and improve data quality, with a particular focus on pensions currently in payment. There was also a restructuring of the financial assets section. Identity, wellbeing, and benefits questions were reinstated after being removed for Round 8 and new questions on income, business and self-employment were added, while several sections, including on housing, mortgages, loans, Individual Savings Accounts (ISAs), and insurance, were streamlined. Some showcards were also reinstated, having previously been removed in Round 8 because of the questionnaire being shortened during the coronavirus (COVID-19) pandemic. 

Round 8 weighting

For Round 8 of the WAS, there have been two updates to the weighting scheme to ensure it is as representative as possible of the Great Britain (GB) population.

Update one

Tenure was included as an additional calibration control to counteract the increased selection bias in the achieved sample during the coronavirus pandemic, which under-represented renters and over-represented households who own outright. More detail on how response by household characteristic has been affected during the coronavirus pandemic can be found in our Impact of COVID-19 on ONS social survey data collection publication.

Update two

Calibration targets are now derived from population series based on Census 2021 for England and Wales and the 2022 Census for Scotland, respectively. Deflation factors have been applied to the census-based estimates of the whole population to account for the fact that the target population for the WAS are private households. All calibration totals have been adjusted to refer to March 2021, the midpoint of the Round 8 data collection period.

The adjustments to the weighting scheme help improve the overall representativity of Round 8 at the GB level. However, it remains unlikely that the weighting scheme will fully compensate for all imbalances in the sample introduced during the coronavirus pandemic. The inclusion of tenure in the weighting scheme helps address renter under‑representation, which is particularly pronounced in London. However, weighting alone cannot fully correct for these imbalances, and London estimates therefore require additional caution.

Furthermore, the weighting scheme is also unlikely to account for unusual household circumstances resulting from the coronavirus pandemic, such as where people were more likely to relocate or change household composition temporarily.

Round 8 data collection

Fieldwork for Round 8 was conducted during the coronavirus pandemic, where national lockdown and social distancing restrictions were in place (see our Impact of COVID-19 on ONS social survey data collection publication).

To conduct fieldwork during this time:

• survey data were collected over the telephone throughout Round 8, rather than face-to-face interviewing

• the questionnaire was shortened to increase accessibility by removing questions that did not feed into core WAS wealth estimates

• Wave 1 address quotas were adjusted several times to manage fieldwork feasibility during the coronavirus pandemic; planned Wave 1 quotas were originally 13 addresses each, but once operationally possible, they were increased to 26 (June to September 2020), then reduced to 24 (October 2020 to January 2021), and further reduced to 20 (February and March 2021)

• Following a significant reduction in funding part way through Round 8, no Wave 1 quotas were issued for the second year of the round

These adjustments enabled data collection to continue under severe operational constraints but reduced comparability with previous rounds, particularly for non-core topics omitted during the shortened questionnaire period. The suspension of Wave 1 quotas in Year 2 also had a material impact on overall response and the depth of analysis possible for Round 8.

Round 7

Round 7 of the WAS commenced in April 2018 where a level of integration took place with the other household finance surveys, with a common primary sample being drawn for all three surveys, and harmonisation of some income questions across the surveys.

Round 6

In April 2016, the survey period moved to a two-year, financial year-based periodicity (April to March), with this periodicity being referred to as a "round". This move to a two-year, financial-year basis allowed the WAS to be integrated with other household financial surveys as part of the Office for National Statistics (ONS) Data Collection Transformation Programme. This programme aimed to bring several surveys together to form the Household Finance Survey. More details are available in our Moving the Wealth and Assets Survey onto a financial years' basis methodology publication.

Wave 2

In Wave 2, additional longitudinal editing was introduced. This used information gathered at Wave 1 to validate the Wave 2 data, but also looking at the Wave 1 data alongside the data given at Wave 2.

In any sample survey, there will always be missing values for individual questions. However, when constructing estimates of wealth, it is necessary that valid responses have been given to all the component estimates. Therefore, any missing values are imputed. The imputation methodology was improved with Wave 2 to take into account the information gathered at both waves.

Upcoming changes

These changes are ordered by date, with the most immediate changes first.

Upcoming publication of a coherence report

A separate report on the UK Data Service (UKDS) to improve transparency on comparability and coherence of the survey is under preparation, users will be updated when this is published.

Round 11 questionnaire updates

Round 11 introduces a substantial questionnaire update. The redesign draws on feedback from expert users, internal stakeholders and field interviewers, with changes developed and tested to improve clarity, flow and data quality.

Improved signposting

The Round 9 (and subsequent) publications will include direct links to this guide and the relevant Household total wealth in Great Britain quality and methods guide in the Related links section.

We compare the Wealth and Assets Survey (WAS) to other sources where users might expect parallel trends or complementary coverage, and where differences in purpose, coverage, concepts, and periodicity explain nonalignment.

A separate report on the UK Data Service (UKDS) to improve transparency on comparability and coherence of the survey is under preparation, users will be updated when this is published.

These resources provide clearer guidance on how WAS statistics compare with other sources, including differences in purpose, coverage and concepts, and offer practical advice to support appropriate use and interpretation.

Annual Survey of Hours and Earnings

The Annual Survey of Hours and Earnings (ASHE), which is produced by the Office for National Statistics (ONS), is broadly comparable with the WAS, but be aware of the following differences:

• the WAS covers Great Britain (GB) (England, Scotland and Wales) and is biennial; the ASHE covers the UK (England, Scotland, Wales and Northern Ireland) and is annual (April snapshot)

• the WAS captures pensions at person and household level and uses self-reported data; the ASHE is an employer survey completed by the payroll administrator

• the WAS collects two employer pensions per person, so numbers are always slightly misaligned with the ASHE; the ASHE only captures pension information for the employer completing the survey

English Housing Survey

The English Housing Survey (EHS), which is produced by the Ministry of Housing, Communities and Local Government (MHCLG), is broadly comparable with the WAS, but be aware of the following differences:

• the WAS statistics cover GB (although the WAS overall property ownership rates are available at country level, allowing an England-only comparison); the EHS is England-only and focuses on current dwelling conditions and tenure

• the WAS captures property wealth within total wealth; the EHS captures housing conditions and tenure

• the WAS is biennial (based on financial years since 2016); the EHS is annual (financial year)

Family Resources Survey

The Family Resources Survey (FRS), which is produced by the Department for Work and Pensions (DWP), is broadly comparable with the WAS, but be aware of the following differences:

• the WAS uses the small-user Postcode Address File (PAF) for Great Britain; the FRS uses the small-user Postcode Address File (PAF) for the United Kingdom

• the WAS includes ownership of all property; the FRS includes ownership of current dwelling only

• the WAS focuses on wealth accumulation and assets with broader categories and longer reference periods for income; the FRS focuses on income and benefits, includes stronger validation of income components, and has more granular categories and shorter reference periods for income

Living Costs and Food survey

The Living Costs and Food Survey (LCF), which is produced by the ONS, is broadly comparable with the WAS, but be aware of the following differences:

• the WAS samples private households in GB only (England, Scotland and Wales); the LCF samples the whole of the UK (England, Scotland, Wales and Northern Ireland)

• the WAS focus is on lifetime accumulation and depletion of wealth and assets; the LCF focus is on recent flows of income and expenditure 

• the WAS collects income for wealth analysis purposes, and it is harmonised with the LCF; the LCF collects more details on income than the WAS, and it allows more in-depth income analysis

Vehicle Ownership

The Vehicle Ownership Administrative Data, which is produced by the Department for Transport and the Driver and Vehicle Licencing Agency (DVLA), is not comparable with the WAS, because of the following differences:

• the WAS covers self-reported ownership of vehicles in GB households; the vehicle ownership data covers all registered vehicles across GB

• the WAS statistics are biennial (based on calendar years for 2006 to 2016 and financial years since 2016); the DVLA publishes the number of vehicle registrations on a quarterly basis

Calibration (survey weighting)

Adjusting weights to align with known population demographic totals (from 2021 and 2022 censuses), including tenure from Round 8 to mitigate coronavirus (COVID-19) pandemic era imbalances.

Confidence interval

Range conveying uncertainty around an estimate. Published for headline measures and being extended to granular breakdowns.

Deductive imputation

A method where missing values are logically inferred from other known responses.

Defined Benefit pension

A workplace pension promising a specified level of income in retirement. It is valued using Superannuation Contributions Adjusted for Past Experience (SCAPE)-based discounting from Round 8 onwards, improving stability but creating a break in the series.  

Donor (nearest-neighbour) imputation

A method where missing values are filled using data from respondents with similar characteristics.

Longitudinal household

A household followed across waves and rounds, enabling analysis of change, attrition, and mobility.

Outliers

Values that fall outside expected ranges or that represent unusually large changes over time, requiring validation.

Oversampling

The deliberate selection of more cases from specific groups (for example, high wealth households) to improve estimate precision.

Primary sampling unit

The first-stage sampling unit, based on postcode sectors selected before individual addresses.

Postcode Address File

The sampling frame for the Wealth and Assets Survey (WAS), comprising all known residential addresses in Great Britain.

Round and wave

Two-year periods (rounds are financial year based since 2016; earlier waves were July to June).

Routing

The automated skipping logic in computer assisted interviewing (CAI) that directs respondents to the correct questions based on previous answers. It is essential for editing and quality assurance (QA).

SCAPE discount rate

The superannuation contributions adjusted for past experience (SCAPE) rate is used to determine employer contribution rates in the valuations of the public service pension schemes. Following review and a series of recommendations from the Government Actuary's Department (GAD), SCAPE is currently used for discounting future DB pensions promises to a present value, and prioritises stability over market-based approaches in order to minimise undue volatility in our pensions estimates.