BESAC(04) - Household Analysis Review Group (HARG)
HARG last met on the 24th February 2004 to discuss methodologies for the 2002 household estimates and 2002-based household projections prior to their release on 23rd April 2004.
Household estimates
Estimates of total number of households for June 2001 and 2002 , based on 2001 Census figures were produced and published in April at the same time as the 2002 household projections. An alternative methodology for the household estimates - using council tax base information, and giving an estimate up to June 2003 - was discussed in HARG. Following the HARG discussion councils' views were sought on the alternative. We decided that, because of the reservations of some councils about the proposed methodology it would be best to take time to review this before making such a radical change. We will pursue this in the HARG work programme for 2004-05.
The 2001 census based household estimates show a discontinuity with the 1991-2001. Scottish total from Census is about 20th lower than previous estimate. HARG has discussed possible methodologies for re-setting estimates back to 1991, but has not yet agreed on a favoured approach. For the next HARG meeting we hope to present a methodology for compiling estimates consistent with the revised mid year population estimates.
Household projections
2002 based projections were published in April. HARG discussed outputs from two different approaches - respectively using head of household (HoH) and household reference person (HRP) bases for compiling household formation rates by age. Both sets of projections were based on changes in household formation patterns between the 1991 and 2001 Censuses; taken from specially commissioned analyses from the Censuses. In particular, the 1991 data were adjusted as far as possible for the change in treatment of students between the two Censuses; and data on household formation patterns on a household reference person basis was provided for 1991 from 10% sample data. [In 91 relationships within households were coded for only 10% of households].
Following discussion in HARG it was agreed that the Head of Household basis was the more reliable of the two approaches. Overall the projections are significantly lower than previous projections. This is due both to the lower population projections and lower growth in the proportion of the population forming households between 1991 and 2001 than between 1981 and 1991.
DD:ASD
June 2004