Guidance

Pensioners’ Incomes data on Stat-Xplore: user guide

Updated 26 March 2026

The Pensioners’ Incomes (PI) Stat-Xplore Database provides information regarding the amounts and sources of the incomes of pensioners in the United Kingdom. Variables are available at a family (benefit unit) level.

Please add “Source: Pensioners’ Incomes Stat-Xplore” to any analysis shared or published.

1. What is Stat-Xplore?

Stat-Xplore is a free tabulation tool available at gov.uk. Users can access DWP data via databases to create their own analysis. PI data is also available via:

	GOV.UK Publication	UK Data Service
Access	Unlimited	Members only
Content	Main report, Tables, Background Information and Methodology report	Rounded variables, Ages of over 80s set to 80 (unless using safe room)

Read more about PI annual reports and accompanying tables, research and technical papers. Versions of the data set are available from the UK Data Service.

2. Benefits of using the PI Stat-Xplore Database

The benefits of using the PI Stat-Xplore Database are:

• that it’s free and accessible to all, with user guidance and virtual tour
• the new analysis of PI data, with a user-friendly interface and quick export of tables and graphs to Excel
• that the data is unrounded, so users can produce more accurate analysis (final estimates must be rounded as described below)
• the open data API functionality that allows users to dynamically create their own tables and data visualisations

3. What are the constraints?

The constraints of using the PI Stat-Xplore Database are:

• that confidence intervals around estimates cannot be produced in Stat-Xplore
• that the analysis based on three-year averages is not possible, so the map feature is not available
• that careful selection of row and column categories is needed, for more information see section 10
• that figures between the years from FYE 1995 to FYE 2021 cannot be compared to the years from FYE 2022 due to the integration of administrative data into the Family Resources Survey

4. PI estimates rounding rules and disclosure

Once you have produced PI estimates using unrounded outputs, the:

• percentages must be rounded to the nearest one per cent
• population numbers must be rounded to the nearest 100,000
• weekly amounts must be rounded to the nearest £1
• annual amounts must be rounded to the nearest £100

This reflects that PI estimates are based on the Family Resources Survey (FRS) and are not actual records of individuals in the UK. Some breakdowns are provided as bands or grouped to further protect against disclosure.

5. PI Stat-Xplore Database: breakdowns available

Stat-Xplore allows users to create their own analysis across all PI years and the following breakdowns.

Descriptions of these breakdowns and any data issues can be found by selecting the ‘i’ icon next to it in the database or via the front-page list.

Variations of these breakdowns are also possible using the ‘Add derivation’ feature. More information on this is in section 9.

Time Characteristic:

The time characteristics are broken down into financial year, 1994 to 1995 to the latest year.

These years are split between two datasets, one with years up to 2020 to 2021 and the other from 2021 to 2022. This is due to the integration of administrative data into the Family Resources Survey from 2021 to 2022 onwards.

Measures of income:

The mean, median and range are provided for:

• gross income:
  • benefit income, which can be broken into State Pension income, income-related benefits income and disability benefit income
  • occupational pension income
  • personal pension income
  • investment income
  • earnings income
  • other income
• net income before housing costs (BHC)
• net income after housing costs (AHC)

Characteristics:

The characteristics are broken down into age of head of pensioner unit, sex of head of pensioner unit and family type.

In receipt flags:

The receipt flags are broken down into:

• benefit income:

  • state pension income

  • income-related benefits income

  • disability benefit income

• occupational pension income

• personal pension income

• investment income

• earnings income

Quintile of the pensioner singles income distribution:

The quintile of the pensioner singles income distribution is broken down into before housing costs (BHC) and after housing costs (AHC).

Quintile of the pensioner couples income distribution:

The quintile of the pensioner couples income distribution is broken down into before housing costs (BHC) and after housing costs (AHC).

6. PI Stat-Xplore Database: how it works

1. Once you have logged in, please take the tour to learn about how to use the Stat-Xplore database. The tour can be accessed by selecting the three dots in the top right-hand corner of the page. Further useful guidance can be found by selecting the ‘?’ icon, also in the top right-hand corner:

2. Scroll down to the Pensioners’ Incomes database and select one of the two datasets: PI - Data to 2021/22 (unlinked), or PI - Data from 2021/22 (admin-linked). Take the time to read the front page for important information on rounding final figures and known issues.

3. Select ‘Go to Dataset’ or double-click a ready-made table.

7. PI Stat-Xplore Database: ready-made tables

There are two ready-made tables in each dataset that demonstrate the types of tables that can be produced using Stat-Xplore, Pensioner units with earnings income and Pensioner units in each quintile of the single pensioner net income distribution (AHC) by age. These are there to be used and can be edited to better suit the information you wish to see.

In each of these tables, the user can view their chosen age of the head of the pensioner unit by selecting from the wafer list.

To export to Excel, click on Download Table in the top right of the screen. This will create a table for every group in the wafer, in this case being for all age bands.

8. PI Stat-Xplore Database: user-defined analysis

When you select ‘Go to Dataset’ (as shown in section 6 of this guide), a page containing a variable list and an empty table is displayed:

Common PI analysis

The following table provides steps for producing some common PI analysis:

Analysis	Filter	Wafer	Row	Column	Numbers to Percentages
Mean income by income source by family type by financial year	Not applicable	Not applicable	Choose the ‘mean’ box under ‘measures’ for as many types of income as you’re interested in. Then add family types.	Financial year (tick the boxes for the years you’re interested in)	Not applicable
Percentage of pensioner units with earnings income by age	Not applicable	Age of head of pensioner unit	In receipt of earnings income	Financial year (tick the boxes for the years you’re interested in)	Select Table options then Percentages then Column
Percentage of pensioners in each quintile of the AHC pensioner singles income distribution by sex	Select Family type then single pensioner	Sex of head of pensioner unit	Quintile of the AHC pensioner singles income distribution (choose all except not applicable)	Financial year (tick the boxes for the years you’re interested in)	Select Table options then Percentages then Column
The occupational pension income distribution for those in receipt	Not applicable	Not applicable	In ‘Measures’ under ‘occupational pension income’ select ‘Range’ and create your desired range (for example from 0 to 500, increment 20). Select your new range listed under ‘Ranges’ and choose all boxes except ‘0 or less’.	Financial year (tick the boxes for the years you’re interested in)	Select Table options then Percentages then Column

Adding derivations

You can use the ‘Add Derivation’ feature which allows you to create your own variation of a category. Here we add a simple derivation to compare half the median income (AHC) of pensioner couples to the median income for pensioner singles.

1. The median income (AHC) of pensioner couples and pensioner singles is added as a row. Note that it is necessary to choose the measure(s) first, select ‘Row’, and then do the same for the family types. Next, ‘Add Derivation’ can be used by selecting the three dots next to the label for Family type:

2. Create a name for the derivation (for example half pensioner couple) and add the formula (in this case, V1*0.5). Select ‘Advanced’ and choose where you want the new row to display. Then press the ‘Create’ button.

3. Select ‘Retrieve Data’ and the numbers, including your new derivation, will appear.

9. PI Stat-Xplore Database: top tips

Stat-Xplore vs Published Tables

When deciding whether to use Stat-Xplore or Published Tables, you need to:

• check whether the breakdown you require is currently available in the published tables and use the published tables where possible

• know that not all breakdowns are available, more information on this is in ‘Exclusions’ in the next section

Build a table in the following order:

1. Filter
2. Wafer
3. Column
4. Row

Select ‘Family Type’ or another classification variable as a wafer to produce the same cross-tabulations for each type.

Convert from percentages

To convert from percentages to population numbers, choose ‘Table Options’ then ‘Percentages’ and select ‘None’.

Convert a table into a graph

Once the table has been created, select the Graph view at the top of the page. You may need to change the ‘Graph by’ to ‘Column’

Removing totals

For some tables, the ‘Total’ column does not add useful information. In these cases, select the three dots next to the variable label and untick ‘Total’.

Unless you are averaging across multiple years, this should be done for ‘Financial year’ for all tables. There may also be other tables you create where the ‘Total’ column or row is not relevant.

Stat Xplore Default Options

There are a number of default functions on Stat Xplore platform. Not all are relevant to the PI database, including the RSE function. RSE stands for Relative Standard Error. The PI database is weighted to provide the correct figures. The weighted plugin also adds RSE figures. This has been set to zero so users can ignore.

10. PI Stat-Xplore Database: exclusions, important footnotes and user feedback

Current Database exclusions (available in published PI tables)

The following breakdowns are not included in this version, due to either small sample sizes or complexities involved with displaying them in Stat-Xplore. They are:

• survey sample sizes
• the percentage of pensioner units with more than 50% of gross income from private sources
• income from annual payments such as Winter Fuel Payments and free TV licences
• the position of pensioners in the overall UK income distribution (including non-pensioners)
• both members over State Pension age (SPa) vs one over SPa and one under SPa
• average incomes of single retired benefit units under SPa
• average incomes of pensioner units where at least one member is aged over 65
• married vs cohabiting couples

Three-year average estimates are not available in Stat-Xplore. As single-year PI estimates for the breakdowns are considered too volatile, estimates based on country, region or ethnicity must be calculated using three-year averages. Output at least three financial years and calculate a three-year average as follows: (yr1 estimate + yr2 estimate + yr3 estimate)/3.

Important footnotes

These important footnotes are displayed on tables, which users must comply with (while displaying footnotes on percentages tables is not possible, they still apply):

Symbol	Description
I	Figures are for the United Kingdom. The reference period is single financial years.
II	Figures derived are unrounded. Before use of these figures, users must use the following rounding conventions: a) Percentages must be rounded to the nearest 1 per cent. b) Numbers must be rounded to the nearest 0.1 million. c) Amounts must be rounded to the nearest £pound (weekly) and nearest £100 (annual). These rounding conventions have been set to reflect that PI estimates are based on survey data and not actual records of individuals in the UK.
III	When comparing year-on-year changes, users are advised to refer to the suite of tables providing confidence intervals around the key PI estimates in the ‘Estimating and interpreting uncertainty in PI’ section of the background information & methodology.  These confidence intervals present how estimates might have varied if a different FRS sample had been created and to help the user to understand where some differences seen in the estimates do represent a true change (and not a result of variation from sampling different people in the UK over time).
IV	The tables use grossing factors based on 2011 Census data, so caution should be exercised when making comparisons with published reports and tables prior to 2012 to 2013.
V	“..” indicates data not being available in that year.
VI	Estimates based on country, region or ethnicity must be calculated using three-year averages. Output at least three financial years and calculate a three-year average as follows: (yr1 estimate + yr2 estimate + yr3 estimate)/3.
VII	Ethnic background is self-declared; data is available for known declarations and excludes ‘choose not to declare’ and ‘unknown’.
VIII	It is recommended that any estimates for groups of less than 50,000 pensioner units are not used as sample sizes are too small for robust analysis.
IX	From the 2024 to 2025 publication, FRS survey data has been integrated with DWP and HMRC administrative records. This replaces most survey-reported benefit amounts with linked administrative data and also adds benefit records where respondents did not report receipt in the survey. This reduces the under-reporting of state benefits and tax credits that is previously seen in the FRS and therefore PI. Due to this methodological change, there is a break in the series between 2020 to 2021 and 2021 to 2022. We advise users not to make a direct comparison of changes across the break. For further details, please see FRS Transformation.
i	Click to view information about the category and any data issues.

11. A) Worked example: average incomes of pensioners by income source, family type and age

1. Start with an empty table (select ‘Clear Table’ if necessary). Expand ‘Family type, tick the boxes for ‘Pensioner couple’ and ‘Single pensioner’ and then select ‘Add to: Wafer’.

2. Select the financial years you are interested in. For this example, we choose all years, so select the arrow next to Financial year, then on ‘Financial year’ in the dropdown menu, then ‘Add to: Column’.

3. Choose the measures you are interested in. Here, we select the means of each different source of income, as well as the medians of net income BHC and AHC. This is to replicate Table 2.1 of the Pensioners’ Incomes series publication tables. Then select ‘Add to: Row’.

4. Finally, select ‘Retrieve data’ and the table will be shown. You can also view information about the measures or classification variables such as family type by selecting the ‘i’ buttons.

5. You can switch between pensioner couples, singles and all pensioners using the ‘Wafers’ dropdown menu.

6. You can add additional breakdowns to the table, for example by selecting ‘Under 75’ and ‘Over 75’ and selecting ‘Add to: Row’. This produces the data in Table 2.6 of the Pensioners’ Incomes publication tables.

7. To download the data into Excel, select ‘Download Table’ in the top-right corner of the page.

8. Producing a graph: Producing a graph: For clarity of presentation, remove all measures except for ‘Median net income after housing costs’ by selecting all other measures and selecting ‘Remove’. Select ‘Graph view’ at the top of the page and change the ‘Graph by:’ from ‘Row’ to ‘Column’.

9. You can export the graph in the same way as for the table – by selecting the ‘Download Graph’ button in the top-right corner of the page.

11. B) Worked example: Number of single pensioners with a gross income above a specified threshold using ranges

1. For this example, you will need a Stat-Xplore account to access ranges. Start by logging into your account then choose either ‘Pensioners’ Incomes’ dataset and select ‘Go to Dataset’.

2. Ensure you have an empty table (select ‘Clear Table’ if necessary). Expand ‘Family Type’ in the left-hand column, tick ‘Single Pensioner’ and select ‘Add to: Filter’.

3. Select the financial years you are interested in, then ‘Add to: Column’.

4. Choose the measures you are interested in. Here, we are interested in different thresholds of gross income. Expand the ‘Measures’, click the blue ‘Range’ button under ‘Gross income’.

5. Input the range of weekly gross income in the ‘From’ and ‘To’ sections, then add the increment. Give the range a name and click ‘Next’ and then ‘Create’. For example, create Range 1 from 0 to 242 with increment 242.

6. Expand the ‘Ranges’ folder in the left-hand column, then expand the range you just created. Tick the bands required in the table, then select ‘Add to: Row’.

7. Finally, select ‘Retrieve data’ and the table will be shown. This will be a table of the estimated total number of single pensioners, by financial year and weekly gross income amount. Note: You can also view information about the measures or classification variables such as family type by selecting the ‘i’ buttons.

8. To download the data into Excel, select ‘Download Table’ in the top-right corner of the page.

9. To view the data graphically, select ‘Graph View’. Then select the type of graph that best fits the data.

12. Feedback

The PI team are actively seeking feedback from users.

Please email: pensioners-incomes@dwp.gov.uk if you have any comments on how you have found using Stat-Xplore or if there are further tabulations you would like to produce.

ISBN: 978-1-78659-660-4