back

Notes

Households Below Average Income (HBAI)

Main uses

HBAI is the key dataset for the analysis of income poverty and is treated as such by both researchers and the government.  It importance comes from the fact that the household income data it contains has been extensively reviewed and processed by government to ensure that it is properly comparable between households.  This involves a process (called 'equivalisation') whereby household incomes are adjusted to reflect the fact that, to have the same standard of living, a family of four (for example) requires more money than a family of one but not four times as much.

As well as extensive income data, HBAI also contain data about individuals' pay rates, working hours, tax credits, housing costs, etc, thus allowing complex analyses.  Furthermore, data from HBAI can be linked with the data for the same individuals/households in Family Resources Survey on which it is based.

Finally, despite its name, HBAI is actually a dataset which covers all levels of income, not just low income.  As such, its uses are much wider then income poverty and cover the whole of the income distribution.

Use of the HBAI datasets themselves, rather than the Department of Work and Pension's annual HBAI reports, is recommended because the government sometimes changes and re-publishes the datasets without re-publishing the reports.  In particular, in 2005/06, the Department of Work and Pensions changed the way that they adjusted household income for the size and composition of the household, applying this change retrospectively to all previous datasets.  Also, in 2003/04, they re-weighted all the datasets for revised population estimates from the 2001 Census.  What this effectively means is that any analyses done prior to April 2006 are now out-of-date unless they have been re-worked using the latest versions of the datasets.

Furthermore, when analysing data for below the UK level (e.g. for Scotland, Wales or Northern Ireland), use of the datasets rather than the published reports becomes mandatory as the published reports only contain a very limited amount of information at this level.

The problem is that HBAI is a difficult dataset to use and it is easy to make mistakes.  The lack of documentation adds further difficulty.  But its central importance means that any serious researcher in the field of poverty has to get to grips with it.  The key is to take the time and effort to properly familiarise yourself with the dataset before using it and a useful discipline is to start by trying to duplicate the results in the annual HBAI reports.  Some documentation or advice might also be available from the team at the Department of Work and Pensions (the HBAI team) or alternatively if you contact us (info@poverty.org.uk) then we might be able to help.

The material below is designed to give you a head start in this learning process.  It should be read in conjunction with two other pages from this site which discuss some of the underlying concepts.  These are the pages on choices of low-income threshold and households, families and benefit units.

top

Source

In summary:

From 2006/07 (but not before), each HBAI dataset comes in two flavours:

For most purposes, the version with the rounded numbers will be sufficient, the exceptions being those where the rounding can make a significant difference.  This will only be the case when the purpose involves classifying benefit units by the value in a field and there are many benefit units at the threshold between different classifications.  For example, estimating the number of pensioners with a private income requires the unrounded figures as there are many pensioners with investment income of a few pennies a week.  Note that the approval process required to access these datasets is non-trivial in terms of both whether or not it is granted and how long it takes to gain the approval.

In addition to the dataset itself, the Department of Work and Pensions publishes an annual report giving many statistical analyses from the dataset.

top

General issues

As will become apparent from the discussion below, HBAI is a difficult dataset to use and it is easy to make mistakes.  The lack of documentation adds further difficulty.  But its central importance means that any serious researcher in the field of poverty has to get to grips with it.  The key is to take the time and effort to properly familiarise yourself with the dataset before using it and a useful discipline is to start by trying to duplicate the results in the Department of Work and Pension's annual HBAI reports.

Which software to use

As the annual dataset is around 33,000 records, it can be exported into Excel.

The importance of 'benefit units'

The HBAI dataset comprises a record for each 'benefit unit' and it is therefore important to understand what this term means and how in relates to both 'individuals' and to 'households'.  This is discussed in detail in the page on households, families and benefit units.  In summary, whereas a household is everyone who lives behind the same 'front door', a 'benefit unit' is an adult plus their spouse (if applicable) plus any dependent children they are living with.  So, for example, a young adult living with their parents would count as one 'household' but two 'benefit units'.  More generally:

The importance of 'equivalisation'

Most of the household income variables in HBAI are 'equivalised' income, rather than 'actual' income, and it is therefore important to understand what equalised income is and why it is used.  This is discussed in detail in the page on choices of low-income threshold.  In summary, 'equivalisation' is a process of adjusting household incomes for household composition and size such that those incomes can reasonably be directly compared with each other.  This means increasing the incomes of small households and decreasing the incomes of large households and the extent of these increases and decreases is determined by an internationally agreed set of scales.

Income data before or after deducting housing costs?

Net (disposable) household income in HBAI is available in two forms, namely before or after deducting housing costs.  The Department of Work and Pensions annual HBAI reports includes analyses on both bases and the results are often rather different on the two bases.

In common with most other commentators, most of our analysis is presented on an 'after deducting housing costs' basis, which is disposable income after housing costs have been deducted.  The reasons for this are twofold: first, housing costs can vary considerably for people in otherwise identical circumstances (e.g. pensioners who have paid off their mortgage versus pensioners who are renting) without the people having any realistic ability to change these costs; and, second, unlike a 'before deducting housing costs' basis, the 'after deducting housing costs' calculations are not affected by such matters as whether housing benefit – which provides for the housing costs of many of the poorest - is considered to be income or not.

Data levels and naming conventions

Although each record is for a benefit unit, the actual data comprises a mixture of benefit unit, household and individual data.  More specifically:

As discussed above, most of the household income variables are for equivalised, rather than actual, income and these variables typically have 'eq' as part of their name.  But HBAI also records the actual net income from which this equivalisation calculation was performed, and these variables typically have 'nt' as part of their name.  It also record gross (i.e. before tax) income and these variables typically have 'gr' as part of their name.

Finally, for historical reasons, many of the variables are recorded twice, with subtle differences in definition.  The general naming convention here is that variables with the prefix 'h' use out-of-date definitions (and therefore never need to be used) whilst those with the prefix 'e' use up-to-date definitions (and are therefore the ones to use).  One important variable, however, does not use these conventions: it is the family type variable, where the field NEWFAMBU is the up-to-date definition and the field FAMBU is the out-of-date definition.

The table below summarises the naming conventions.

Type of naming convention Value Meaning
Suffix hd the variable relates to the head of the benefit unit
sp the variable relates to the spouse of the benefit unit
bu the variable is a benefit unit level variable
hh the variable is a household level variable (and thus has the same value for all benefit units in the household)
Prefix e the variable uses an up-to-date definition
h the variable uses an out-of-date definition (effectively means that it never need be used and that the equivalent variable beginning with 'e' should be used instead)
Other gr the variable relates to gross income
nt the variable relates to net income
eq the variable relates to equivalised income

What weights to use

In most datasets, the results are calculated by multiplying the field of interest by a weight field.  For example, if there were 3 children in the benefit unit and the benefit unit has a weight of 1,000 then that benefit unit in the survey represents 3,000 children in the population.  HBAI does things differently in that it does this multiplication for you.  Thus, in addition to the benefit unit weight (GS_NEWBU), it also includes variables for the weighted numbers of children, working-age adults, pensioners and population as a whole.  The table below lists these variables.

Variable name Variable description
GS_NEWBU Benefit unit weight
GS_NEWCH Weighted number of children represented by the benefit unit
GS_NEWPN Weighted number of pensioners represented by the benefit unit
GS_NEWPP Weighted number of children represented by the benefit unit
GS_NEWWA Weighted number of working-age adults represented by the benefit unit

UK-wide analysis or Great Britain-wide analysis?

Whilst the HBAI datasets from 2002/03 onwards are for the whole of the UK, those for earlier years are for Great Britain only.  For time trends, this raises the issue of whether these should be for UK as a whole or for Great Britain only.  This website has adopted the following approach:

What variables to use and what do they mean

Clearly, the variables to be used will vary according to the purpose at hand.  There are, however, a small number of variables which will be used much more than any of the others.  The table below lists these variables.

Variable name Variable description
ECOBU The work status of the benefit unit
GVTREGN Government Office Region
NEWFAMBU The family type of the benefit unit
S_OE_AHC The equivalised net income, after deducting housing costs, of the household of which the benefit unit is a part, where OECD scales have been used for the equivalisation
S_OE_BHC The equivalised net income, before deducting housing costs, of the household of which the benefit unit is a part, where OECD scales have been used for the equivalisation

In most datasets, each variable has both a name and a description but in HBAI most of the variables only have names and no description.  The dataset is, however, accompanied by a spreadsheet which provides descriptions for some (but by no means all) of the variables.   The table below lists these for the 2005/06 HBAI dataset.  The list is sorted by whether or not the variable is used in any of the calculations underlying the graphs in this website and, within this, by alphabetical order of name.

Variable name Official variable description Used in this website? Extent of use
ADULTB Number of adults in benefit unit yes minor
AGEHD Age of head yes minor

 

AGESP Age of spouse yes minor
BENUNIT Position of benefit unit in household yes minor
CTBAND Council Tax band yes minor
CTCBU Amount of Child Tax Credit for the Benefit Unit yes minor
CTLIAB Amount of council tax the household is liable for yes minor
CTREBAM1 Amount of council tax rebate the household receives yes minor
DEPCHLDH Number of dependant children in the household yes minor
DISABFLG Number of disabled adults in the benefit unit yes minor
ECOBU Economic status of the benefit unit  yes major
EHBENBU FRS extended - housing benefit for benefit unit yes minor
EHCOST FRS extended - total housing costs. yes minor
EQOAHCHH Household level OECD equivalence scale for BHC income yes minor
EQOBHCHH Household level OECD equivalence scale for AHC income yes minor
ERENTBU FRS extended - rent component of housing costs for the benefit unit yes minor
ESGINVBU FRS extended - SPI'd gross investment income for the benefit unit yes minor
ESGJOBBU FRS extended - SPI'd gross employment income for the benefit unit yes minor
ESGOCCBU FRS extended - SPI'd gross occupational pension income for the benefit unit yes minor
ESGRSEBU FRS extended - SPI'd gross self-employment income for the benefit unit yes minor
ESMISCBU FRS extended - SPI'd total miscellaneous income for the benefit unit yes minor
ESPRIBBU FRS extended - SPI'd private benefit income for the benefit unit yes minor
ETHGRPHH Ethnic group of household yes minor
GS_NEWBU FRS extended - SPI'd benefit unit grossing factor yes major
GS_NEWCH FRS extended - SPI'd dependant children grossing factor yes major
GS_NEWPN FRS extended - SPI'd pensioner grossing factor yes major
GS_NEWPP FRS extended - SPI'd population grossing factor yes major
GS_NEWWA FRS extended - SPI'd working-age adults grossing factor yes major
GVTREGN Government Office Region yes major
INCHILBU Total childrens income in the benefit unit yes minor
ISBU Total income support for the benefit unit yes minor
KID0_1 Number of children aged 0 and 1 in the benefit unit yes minor
KID11_12 Number of children aged 11 and 12 in the benefit unit yes minor
KID13_15 Number of children aged 13, 14 and 15 in the benefit unit yes minor
KID16_18 Number of children aged 16, 17 and 18 in the benefit unit yes minor
KID2_4 Number of children aged 2,3 and 4 in the benefit unit yes minor
KID5_7 Number of children aged 5, 6 and 7 in the benefit unit yes minor
KID8_10 Number of children aged 8, 9 and 10 in the benefit unit yes minor
LONDON Inner or Outer London flag yes minor
MDOEAHC The median of S_OE_AHC yes minor
MDOEBHC The median of S_OE_BHC yes minor
MNOEAHC The average of S_OE_AHC yes minor
NEWFAMBU New family type for the benefit unit yes major
OAHCDEC AHC deciles from S_OE_AHC income variable, including self-employed (OECD) yes minor
OBHCDEC BHC deciles from S_OE_BHC income variable, including self-employed (OECD) yes minor
S_OE_AHC FRS extended - deflated, equivalised, new SPI AHC income (OECD) yes major
S_OE_BHC FRS extended - deflated, equivalised, new SPI BHC income (OECD) yes major
SERNUM Serial number of the household yes minor
SEXHD Sex of the head 1= Male, 2= female yes minor
SEXSP Sex of the spouse yes minor
TENTYP2 The tenure type of the household  yes minor
TENURE Tenure yes minor
UNDER14 Number of children under 14 years of age in the household (OECD) yes minor
WTCBU Amount of Working Tax Credit for the Benefit Unit yes minor
ADULTH Number of adults in household no
AHCDEF Used to deflate AHC income to the average of the survey year no
BENUNITS Number of Benefit Units within the household no
BHCDEF Used to deflate BHC income to the average of the survey year no
BOYS Number of boys in benefit unit no
CHARGHH Total household service charges  no
CHRGAMT1 Amount paid for Ground Rent  no
CHRGAMT2 Amount paid for feu duty  no
CHRGAMT3 Amount paid for chief rent  no
CHRGAMT4 Amount paid for service charge  no
CHRGAMT5 Amount paid for regular maintenance  no
CHRGAMT6 Amount paid for site rent  no
CHRGAMT7 Amount paid for factoring  no
CHRGAMT8 Amount paid for other regular charges no
CHRGAMT9 Amount paid for combined services no
CONTV1 Whether household has colour TV set no
CONTV2 Whether household has black/white TV set no
COVOTHS Insurance premium: what covered no
CWATHH household water charges no
DEPCHLDB Number of dependant children in the benefit unit no
EAHCHH FRS extended - unequivalised, undeflated, un-spi'd after housing costs income for the household no
EBENINBU FRS extended - total benefit income for the benefit unit no
EBENINHH FRS extended - total benefit income for the household no
EGRERNBU FRS extended - gross earnings for the benefit unit no
EGRERNHH FRS extended - gross earnings for the household no
EGRINCBU FRS extended - gross income for the benefit unit no
EGRINCHH FRS extended - gross income for the household no
EMISCIBU FRS extended - total miscellaneous income for the benefit unit no
EMISCIHH FRS extended - total miscellaneous income for the household no
ENTERNBU FRS extended - net earnings for the benefit unit no
ENTERNHH FRS extended - net earnings for the household no
ENTINCBU FRS extended - net income for the benefit unit no
ENTINCHH FRS extended - net income for the household no
EOTHDEBU FRS extended - other deductions for the benefit unit no
EOTHDEHH FRS extended - other deductions for the household no
EPRIBNBU FRS extended - private benefit income for the benefit unit no
EPRIBNHH FRS extended - private benefit income for the household no
EQMAHCHH Household level McClements equivalence scale for AHC income no
EQMAHCKD McClements equivalence factor for all children in household - AHC no
EQMBHCHH Household level McClements equivalence scale for BHC income no
EQMBHCKD McClements equivalence factor for all children in household - BHC no
EQMCOUP Status of adults in the benefit unit (McClements) no
ERENTHH FRS extended - rent component of housing costs for the household no
ERICNPBU Benefit unit SPI flag - non pensioner no
ERICNPHH Household SPI flag - non pensioner no
ERICPBU Benefit unit SPI flag - pensioner no
ERICPHH Household SPI flag - pensioner no
ES_HCOST FRS extended - SPI'd total housing costs no
ESAHCHH FRS extended - SPI'd after housing costs income no
ESBENIBU FRS extended - SPI'd total benefit income for the benefit unit no
ESBENIHH FRS extended - SPI'd total benefit income for the household no
ESGINCBU FRS extended - SPI'd gross income for the benefit unit. no
ESGINCHH FRS extended - SPI'd  gross income for the household no
ESGINVHH FRS extended - SPI'd gross investment income for the household no
ESGJOBHH FRS extended - SPI'd gross employment income for the household no
ESGOCCHH FRS extended - SPI'd gross occupational pension income for the household no
ESGRSEHH FRS extended - SPI'd gross self-employment income for the household no
ESMISCHH FRS extended - SPI'd miscellaneous income for the household no
ESNINCBU FRS extended - SPI'd net income for the benefit unit no
ESNINCHH FRS extended - SPI'd  household net income no
ESOTHDBU FRS extended - SPI'd  other deductions for the benefit unit no
ESOTHDHH FRS extended - SPI'd  other deductions for the household no
ESPRIBHH FRS extended - SPI'd private benefit income for the household no
FAMBU The family type of the benefit unit (pre-0102 - replaced with NEWFAMBU)  no
FAMTHBAI Family type-HBAI definition no
FSMBU BU-Value of Free school meals no
FSMLKBU BU-Value of Free school milk no
FWMLKBU BU-Value of Free welfare milk no
G_NEWAD FRS extended - adult grossing factor no
G_NEWBU FRS extended - benefit unit grossing factor no
G_NEWCH FRS extended - dependant children grossing factor no
G_NEWHH FRS extended - household grossing factor no
G_NEWPN FRS extended - pensioner grossing factor no
G_NEWPP FRS extended - population grossing factor no
G_NEWWA FRS extended - working-age adult grossing factor no
GINSEBU Total gross self-employment income for benefit unit no
GNEWHHP Household level population weight no
GS_NEWAD FRS extended - SPI'd adult grossing factor no
GS_NEWHH FRS extended - SPI'd household grossing factor no
HBENBU Benefit unit level housing benefit income no
HBENHH Household housing benefit income no
HBENINBU FES equivalent - total benefit income for the benefit unit no
HBENINHH FES equivalent - total benefit income for the household no
HBMORT FES equivalent - mortgage interest component of housing costs no
HBXMORT FRS extended - mortgage interest component of housing costs no
HGRERNBU FES equivalent - gross earnings for the benefit unit no
HGRERNHH FES equivalent - gross earnings for the houshold no
HGRINCBU FES equivalent - gross income for the benefit unit no
HGRINCHH FES equivalent - gross income for the household no
HGRINVBU FES equivalent - gross investment income for the benefit unit no
HGRINVHH FES equivalent - gross investment income for the household no
HGROCCBU FES equivalent - gross occupational pension income for the benefit unit no
HGROCCHH FES equivalent - gross occupational pension income for the household no
HHCOST FES equivalent - total housing costs no
HHSTAT Household status (conventional/shared) no
HMISCIBU FES equivalent - total miscellaneous income for the benefit unit no
HMISCIHH FES equivalent - total miscellaneous income for the household no
HNTERNBU FES equivalent - net earnings for the benefit unit no
HNTERNHH FES equivalent - net earnings for the household no
HNTINCBU FES equivalent - net income for the benefit unit no
HNTINCHH FES equivalent - net income for the household no
HNTINVBU FES equivalent - net investment income for the benefit unit no
HNTINVHH FES equivalent - net investment income for the household no
HNTOCCBU FES equivalent - net occupational pension income for the benefit unit no
HNTOCCHH FES equivalent - net occupational pension income for the household no
HPRIBNBU FES equivalent - private benefit income for the benefit unit no
HPRIBNHH FES equivalent - private benefit income for the household no
HRENTHH FES equivalent - gross rent for the household no
INCHILHH Total childrens income in the household no
INSEBU Total net self-employment earnings for the benefit unit no
INTDATE Interview date (dd/mm/yy) no
INTPRPAY Amount of last installment on the loan no
KIDECOBU Children's economic status no
MAHCDEC AHC deciles from S_MC_AHC income variable, including self-employed (McClements) no
MAHCDECX AHC deciles from S_MC_AHC income variable, excluding self-employed (McClements) no
MBHCDEC BHC deciles from S_MC_BHC income variable, including self-employed (McClements) no
MBHCDECX BHC deciles from S_MC_BHC income variable, excluding self-employed (McClements) no
MC_AHC FRS extended - deflated, equivalised AHC income (McClements) no
MC_BHC FRS extended - deflated, equivalised BHC income (McClements) no
MC_GRO FRS extended - deflated, equivalised gross income (McClements) no
MDMCAHC The median of S_MC_AHC no
MDMCAHCX The median of S_MC_AHC, excluding the self-employed no
MDMCBHC The median of S_MC_BHC no
MDMCBHCX The median of S_MC_BHC, excluding the self-employed no
MDOEAHCX The median of S_OE_AHC, excluding the self-employed no
MDOEBHCX The median of S_OE_BHC, excluding the self-employed no
MNMCAHC The average of S_MC_AHC no
MNMCAHCX The average of S_MC_AHC, excluding the self-employed no
MNMCBHC The average of S_MC_BHC no
MNMCBHCX The average of S_MC_BHC, excluding the self-employed no
MNOEAHCX The average of S_OE_AHC, excluding the self-employed no
MNOEBHC The average of S_OE_BHC no
MNOEBHCX The average of S_OE_BHC, excluding the self-employed no
MORINPAY Amount of mortgage interest paid no
MORINPD Pcode: amount of mortgage interest paid no
MORINUS Whether mortgage interest usual no
MORTLEFT Amount outstanding on mortgage no
MORTSEQ Mortgage Sequence Number no
MORTTYPE Type of mortgage/loan no
MORUS Amount of usual mortgage interest paid no
MPPINT Total payments per mortgage no
MQINAHC AHC quintiles from S_MC_AHC income variable, including self-employed no
MQINAHCX AHC quintiles from S_MC_AHC income variable, excluding self-employed no
MQINBHC BHC quintiles from S_MC_BHC income variable, including self-employed no
MQINBHCX BHC quintiles from S_MC_BHC income variable, excluding self-employed no
MSFLAGHH Household flag to identify missing spouse benefit units no
NINRV Net Rateable Value Of Property (NI) no
NIRATE Amount of District Council Rate (NI) no
NORATE Why  no  rate bill (NI) no
OAHCDECX AHC deciles from S_OE_AHC income variable, excluding self-employed (OECD) no
OBHCDECX BHC deciles from S_OE_BHC income variable, excluding self-employed (OECD) no
OE_AHC FRS extended - deflated, equivalised AHC income (OECD) no
OE_BHC FRS extended - deflated, equivalised BHC income (OECD) no
OE_GRO FRS extended - deflated, equivalised gross income (OECD) no
OQINAHC AHC quintiles from S_OE_AHC income variable, including self-employed no
OQINAHCX AHC quintiles from S_OE_AHC income variable, excluding self-employed no
OQINBHC BHC quintiles from S_OE_BHC income variable, including self-employed no
OQINBHCX BHC quintiles from S_OE_BHC income variable, excluding self-employed no
OVER14 Number of children over 14 years of age in the household (OECD) no
RENTBU FES equivalent - rent component of housing costs for the benefit unit no
RTCHECK Weeklyised amount of Annual rates paid (NI) no
RTREBAMT Amount of Rate Rebate allowed (NI) no
S_MC_AHC FRS extended - deflated, equivalised, new SPI AHC income (McClements) no
S_MC_BHC FRS extended - deflated, equivalised, new SPI BHC income (McClements) no
S_MC_GRO FRS extended - deflated, equivalised, new SPI gross income (McClements) no
S_OE_GRO FRS extended - deflated, equivalised, new SPI gross income (OECD) no
SEHHFLG Number of full time self-employed persons in the household no
SERINC1 Services included in rent: heating no
SERINC2 Services included in rent: lighting no
SERINC3 Services included in rent: hot water no
SERINC4 Services included in rent: cooking fuel no
SERINC5 Services included in rent: TV licence no
SEWAMT Sewerage rates: amount paid last time no
SEWERAGE Sewerage charges for households in Scotland no
SFRPABU Amount of social fund repayments from the benefit unit no
SFRPAHH Amount of social fund repayments from the household no
SHTSEFLG Number short-term self-employed (less than one month) in the household no
SNPHCOST Average of housing costs for rich non-pensioner households no
SPHCOST Average of housing costs for rich pensioner households no
SRENTBU Ben unit accumulation of rent paid by adults in shared households no
STRAMT1 Amount: Insurance part of repayment no
STRAMT2 Amount: Insurance premium no
STRCOV Items covered by insurance policy no
STRMORT Whether mortgage payments inc. insurance no
STROTHS Whether pay structural insurance no
SUBRENT Amount of rent from subletting no
TOTSTHH FES - Total structural insurance no
TOTSTRHH FRS - Total structural insurance no
TVLIC Whether claims concessionary TV licence no
TVLICBU Amount of concessionary TV licence allocated to the benunit no
TVLICHH Amount of concessionary TV licence allocated to the household no
WATAMT Water rates: amount paid last time no
WATSEWHH Total amount of water and sewerage charges for the household no
WATSEWRT Total Water and Sewerage no
WHYNOCT Reason for paying no Council Tax (Great Britain) no
WINPAYBU Total winter fuel payment for benefit unit no
WSEWAMT Combined water/sewer rates: amount paid no

What do the values of the variables mean

The values of some variables take the value of codes which only make sense if one knows what the codes mean.  In most datasets, both the code and its meaning are stored.  In HBAI, however, only the codes themselves are stored and are thus meaningless unless one knows what the codes mean.  The table below list the fields in the 2005/06 HBAI dataset which are of this type together with the meaning of each code value.

Variable name Description Code values and descriptions
SEXHD/SP Sex of head/spouse

1 = male

2 = female

ECOBU Economic status of the benefit unit

1 = One or more self-employed

2 = Single or couple, all in full-time work

3 = Couple, one in full-time work, one part-time

4 = Couple, one in full-time work, one not working

5 = No full-time work, one or more in part-time work

6 = Workless, head or spouse aged 60 or over

7 = Workless, head or spouse unemployed

8 = Workless, other economically inactive

Note that this coding is implemented in a hierarchical fashion.  So, for example, code 2 means that the benefit unit is a single or couple, all in full-time work, but that neither is self-employed.  And, code 7 means that the benefit unit is workless, with head or spouse unemployed, but with neither aged 60 or over.

ETHGRPHH Ethnic group of household

1 = White - British

2 = Any other white background

3 = Mixed - White and Black Caribbean

4 = Mixed - White and Black African

5 = Mixed - White and Asian

6 = Any other mixed background

7 = Asian or Asian British - Indian

8 = Asian or Asian British - Pakistani

9 = Asian or Asian British - Bangladeshi

10 = Any other Asian/Asian British background

11 = Black or Black British - Caribbean

12 = Black or Black British - African

13 = Any other Black/Black British background

14 = Chinese

15 = Any other

KIDECOBU Children's economic status

1 = Lone parent: in full-time work

2 = Lone parent: in part-time work

3 = Lone parent: not working

4 = Couple: one or more full-time self employed

5 = Couple: both in full-time work

6 = Couple: one in full-time work, one in part-time work

7 = Couple: one in full-time work, one not working

8 = Couple: one or more in part-time work only

9 = Couple: both not in work

NEWFAMBU Family type of benefit unit

1 = Pensioner couple (where one or more adults are state pension age or over)

2 = Single male pensioner

3 = Single female pensioner

4 = Working-age couple with children

5 = Working-age single with children

6 = Working-age couple without children

7 = Single working-age male without children

8 = Single working-age female without children

Note that there is another variable which give the family type - FAMBU - but that there is not need ever to use this variable as it uses out-of-date definitions.

GVTREGN Government Office Region

1 = North East

2 = North West

4 = Yorkshire and the Humber

5 = East Midlands

6 = West Midlands

7 = Eastern

8 = London

9 = South East

10 = South West

11 = Wales

12 = Scotland

13 = Northern Ireland

LONDON Inner or outer London flag

1 = inner London

2 = outer London

3 = not London

TENTYP2 Tenure type of household

1 = LA rented unfurnished

2 = Housing Association

3 = Other rented unfurnished

4 = Private rented furnished

5 = Owned with mortgage

6 = Owned outright

7 = Rent free

8 = Other Crown Est/Govt.

9 = Co-ownership schemes

10 = Shared ownership

TENURE Tenure

1 = Owns it outright

2 = Buying with the help of a mortgage

3 = Part own, part rent

4 = Rents

5 = Rent-free

6 = Squatting

CTBAND Council Tax band

1 = Band A

2 = Band B

3 = Band C

4 = Band D

5 = Band E

6 = Band F

7 = Band G

8 = Band H

9 = Household not valued separately

WHYNOCT Reason for paying no Council Tax (Great Britain only)

1 = Bill not yet received/not previously liable

2 = Bill not yet paid/not previously liable

3 = Deliberate non-payment

4 = Household only recently moved in

5 = Household has formal exemption from the tax

6 = Other reason

NORATE Why no rate bill (Northern Ireland only)

1 = Rented accommodation with rates included in rent

2 = Rent/rates free

3 = Receive rebate

4 = Other reason

top

Specific issues

Links to the Family Resources Survey dataset

The HBAI dataset is derived from the FRS dataset and, as there is precisely one record in the FRS benefit unit table for every record in the HBAI dataset, the two datasets can be linked together using a combination of the household serial number and the benefit unit number.  This can be useful for certain types of analysis (e.g. by religion) where the data is in FRS but not in HBAI.

Analysis by fixed income thresholds

To calculate the numbers below a fixed threshold, this threshold needs to be calculated after taking account of inflation.  Obtain the monthly retail price indices from the ONS site.  For the years from 2000/01 onwards, calculate the index as the average for the twelve months covered by the dataset using the 'all items excluding housing' (CHAZ) index.  For earlier years, use the following table to decide which monthly index to use.
Year Appropriate monthly index
1994/95 Jul 1996
1995/96 Jun 1997
1996/97 Apr 1998
1997/98 Apr 1999
1998/99 Feb 2000
1999/00 Dec 2000

This is because it is only from 2000/01 onwards that HBAI incomes are based on the average prices for the survey year.  Prior to that, prices for particular months were used as itemised in the table above.

top

Relevant graphs on this website

UK graphs

Indicator Graphs Comments
Numbers in low income first and second

Need to calculate the value of the fixed threshold for each subsequent year.

Location of low income all  
The impact of housing costs all  
Low income by age group first  
second

Use the first year of HBAI (1995/95) rather than the usual 'decade earlier' as the baseline year.

third  
fourth

HBAI now only records the age of the children by age group (0-1, 2-4, 5-7, 8-10, 11-12, 13-15 and 16-18) rather than by individual years (because the OECD equivalisation scales only require this).  To calculate the numbers for a individual year, assume that the numbers for each of the individual years within the age group are the same).

Low income by family type first  
second

The amounts below the 60% low-income threshold are unequivalised rather than equivalised.  This is calculated by multiplying the equivalised amounts by the equivalisation factors.  The equation is therefore (0.6 * MDOEAHC - S_OE_AHC)* EQOAHCHH.

Low income and ethnicity first and second

Use 'White' rather than 'White British' as the latter is not available for the earlier years.  Calculate 'White' as 'White British' plus 'Any other white background'.

third to sixth

Calculate 'ethnic minorities' as 'total' - 'White British'

Low income by gender first to fifth

Exclude children as the graphs are for adults only.

sixth

Exclude children as the graph is for adults only.

For couples, calculate the number of adults and assume that half are male and half are female (totally accurate apart from same sex couples).

Income inequalities first to fourth

Follow the following process:

  • For each BU, multiply the household income by the population weight (i.e. S_OE_AHC * GS_NEWPP)

  • For each income decile, sum the result to get the total amount of income by decile.   This gives the third and fourth graphs.

  • Divide the whole population's total income by the total size of the population to get income per pop.

  • Put the income per pop onto a consistent price basis for the start and end years using the monthly retail price indices from the ONS site.

  • Compare the end year's results with the start year's results to get an overall % increase in real incomes.

  • For the both start and end years, estimate each decile's share of total income as a three-year average.  This is done, rather than take the actual year's share, to improve statistical reliability, as the subsequent calculations are very sensitive to the precise shares used.

  • For each income decile, calculate the percentage increase between start and end years as that decile's share of total income in the end year * the overall % increase in real incomes / the decile's share of total income in the start year.  This gives the first graph.

  • Turn these percentages into absolute amounts.  This gives the second graph.

fifth

Use the deciles field OAHCDEC to allocate people to income deciles.

In receipt of tax credits third

Requires the unrounded version of the dataset in order for the 10.50 threshold to be applied properly.

Exclude both workless BUs and pensioner BUs.

Allocate the BU to a tax credit group as follows:

  • Calculate total tax credit as Working Tax Credit (WTCBU) + Child Tax Credit (CTCBU).

  • If this amount is more than 10.50 (the value of the basic family element), then the BU is deemed as getting tax credits.  If the amount is less than or equal to 10.50, then the BU is deemed as not getting tax credits.

  • Calculate the amount by which the unequivalised household income is below the 60% low-income threshold by multiplying the equivalised amounts by the equivalisation factors.  The equation is therefore (0.6 * MDOEAHC - S_OE_AHC)* EQOAHCHH.

  • Allocate the BU to a tax credit group by comparing the two amounts above.

Note that there are quite a lot of BUs in the dataset who are recorded as receiving tax credits just above (and just below) 10.50 and it is not immediately obvious whether such BUs are in reality receiving a small amount above the family element (as the numbers state) or if it is an issue of rounding (e.g. because a BU who receives just the family element of 10.50 on average sometimes receives a bit more and sometimes a bit less).  After discussions with DWP, it was decided to assume the former, and thus apply the 10.50 rule above.  This is because, if the latter were the case, then one would expect to see an upper limit (say, 11) at which there were very few BUs but this is not in fact the case.  Note that, even if this assumption is wrong, it should not affect the trends as the same assumption has been applied to each year.

Children in low-income households all bar the fourth  
fourth

Exclude pensioner BU family types as analysis by work status is not relevant to them.

Children in receipt of tax credits all

Requires the unrounded version of the dataset in order for the 10.50 threshold to be applied properly.

Exclude both workless BUs and pensioner BUs.

Allocate the children to a tax credit group as follows:

  • Calculate total tax credit as Working Tax Credit (WTCBU) + Child Tax Credit (CTCBU).
  • If this amount is more than 10.50 (the value of the basic family element), then the BU is deemed as getting tax credits.  If the amount is less than or equal to 10.50, then the BU is deemed as not getting tax credits.
  • Calculate the amount by which the unequivalised household income is below the 60% low-income threshold by multiplying the equivalised amounts by the equivalisation factors.  The equation is therefore (0.6 * MDOEAHC - S_OE_AHC)* EQOAHCHH.
  • Allocate the BU to a tax credit group by comparing the two amounts above.

Note that there are quite a lot of BUs in the dataset who are recorded as receiving tax credits just above (and just below) 10.50 and it is not immediately obvious whether such BUs are in reality receiving a small amount above the family element (as the numbers state) or if it is an issue of rounding (e.g. because a BU who receives just the family element of 10.50 on average sometimes receives a bit more and sometimes a bit less).  After discussions with DWP, it was decided to assume the former, and thus apply the 10.50 rule above.  This is because, if the latter were the case, then one would expect to see an upper limit (say, 11) at which there were very few BUs but this is not in fact the case.  Note that, even if this assumption is wrong, it should not affect the trends as the same assumption has been applied to each year.

Young adults in low-income households all The analysis is for adults aged 16 to 24.
Low income by work status all

The analysis is for Great Britain only as it is primarily numbers rather than rates.

Exclude pensioner BU family types as analysis by work status is not relevant to them.

Low income and disability all

The analysis is for adults aged 25 to retirement rather than all working-age adults.  This is because the high prevalence of low income among younger adults combined with the low prevalence of disability in that age group arguably distorts, and certainly changes, the comparison between disabled and non-disabled people.

Where the BU is a couple and only one of the couple is disabled, the dataset does not identify which is the one who is disabled.  In those cases, where one is in the age group being analysed and the other is not, it is assumed that half of the disabled adults are in the age group being analysed.

Low income by age (working-age) all Note that, the analysis by disability is whether or not there any of the adults in the BU is disabled rather than whether or not the particular individual adult is disabled.
Low income and council tax all

The analysis is for England and Wales only.  In Scotland, water charges are included with Council Tax and everyone pays water charges so it is not possible to distinguish people who are paying no Council Tax.  In Northern Ireland, there is a Rates system rather than Council Tax.

Exclude those BUs where Council Tax is not applicable (CTLIAB=0)

Decide whether the BU is in a household which is paying full, partial or no Council Tax by comparing their Council Tax liability (CTLIAB) with their Council Tax rebate (CTREBAM1):

  • If CTREBAM1 = 0, then paying full.
  • If CTREBAM1 = CTLIAB then paying none.
  • If CTLIAB>CTREBAM1>0 then paying partial.

For the third graph, then exclude those BUs who in the household who are not responsible for paying the Council Tax.  This is because, in the HBAI dataset, Council Tax is actually a household-level variable rather than a BU-level variable but that it is the work status of the BUs that are responsible for paying this Council Tax which is relevant.  Analysis of the Family Resources Survey, where Council Tax is recorded at the BU level, shows that this exclusion can be done with reasonable accuracy by excluding those BUs who are not the first BU in the household (i.e. BENUNIT > 1).  Note that the consequent results are effectively the number of households in each Council Tax category for a given work status of the BU responsible for paying that Council Tax.

Older people in low income all bar the fifth  
fifth The amounts below the 60% low-income threshold are unequivalised rather than equivalised.  This is calculated by multiplying the equivalised amounts by the equivalisation factors.  The equation is therefore (0.6 * MDOEAHC - S_OE_AHC)* EQOAHCHH.
Older people with no private income first

Requires the unrounded version of the dataset as there are many pensioners who have an investment income of a few pennies a week and these will be wrongly classified as 'no private income' if the rounded version of the dataset is used.

The total income from all sources other than State benefits is INCHILBU + ESGJOBBU + ESGRSEBU + ESGINVBU + ESGOCCBU + ESPRIBBU +ESMISCBU - TVLICBU.  BUs are then considered to have not income other than State benefit if and only if this total is zero.  The deduction of TVLICBU is crucial: it is the amount of the free TV licence for those aged 75 and over and, for some reason, HBAI includes it as income in the ESMISCBU variable.  So, unless it is excluded, all pensioner households aged 75 and over would count as having some income other than State benefits.

Polarisation by housing tenure first  
Housing Benefit second

Analyse by BU rather than by individual.

Only include BUs who are paying rent (i.e. where the rent variable ERENTBU > 0).  Inter alia, by excluding those BUs who are not paying rent but are living with other BUs who are paying rent (e.g. young adults living with their parents), this means that the count of relevant BUs is the same as the count of relevant households so the results can reasonably be interpreted as households.

Decide whether the BU is getting Housing Benefit or not using the Housing Benefit variable EHBENBU (if >0 then getting Housing Benefit)

For those getting Housing Benefit, decide whether it is full or partial by comparing ERENTBU with EHBENBU (only full if the two are equal).

Scotland, Wales and Northern Ireland graphs

These are effectively a subset of the UK graphs using government region as a filter.

top