News and events

News from Cass Business School

New research shows that household statistics are inaccurate

Study examined populations in the six Olympic Boroughs

Statistics that inform public policy on health care, childcare, education and housing are fragmented with significant gaps and inaccuracies that limit their usefulness, a new study from Cass Business School, part of City University London has found.

A new approach to population analysis, successfully implemented in the six Olympic boroughs of London during 2011, creates population and household statistics using locally available administrative data, for example GP, council tax, electoral and school registers. The research finds that the actual number of households is in the six Olympic Boroughs is 36,000 more or 6.6% higher than figures reported by the Department for Communities and Local Government (DCLG) at 578,000.

The idea and method of using administrative data to count and classify households has been developed by Les Mayhew, Professor of Statistics at Cass Business School, and Gillian Harper of Mayhew Harper Associates Ltd. Professor Mayhew presented his alternative to the Census to the Science and Technology Committee at The House of Commons in December 2011, before this latest report's full publication this week.

Household characteristics such as occupancy, residents' ages, tenure and benefit status, and gender composition are often good predictors of social outcomes, such as the take up of health, social care, education services, and crime and are highly correlated with the consumption of utilities like electricity or water. Household units are also extensively used in Government to benchmark households below the poverty line and for a range of other purposes including forecasting housing demand.

Current methodology for determining household demographics relies on projections from DCLG, underpinned by population data and household data from the Office for National Statistics (ONS). This information tends to be mainly imputed or modelled using census data that is potentially up to twelve years old. The new method uses locally available information sources and links and analyses them to produce up to date core demographic data by individual and household. 

Commenting on the study Mayhew said: "As well as forming the backbone of evidence for assessing future housing need, household level data is useful for designing services and informing policy in areas such as welfare benefits and childcare. This paper has identified a significant gap in the availability and quality of household statistics at local level which adversely impacts upon the efficacy of local policy making and investment decisions.

"Using locally available administrative data is flexible, because it allows data to be produced at a finer geographic level; cheaper, because it uses readily available data; more up to date, as it does not rely on the ten year census and more dynamic, as it allows for greater range in classifying households."

Read the full report here

Share this article