§ Mr. CousinsTo ask the Chancellor of the Exchequer, pursuant to the answer to the hon. Member for the Vale of Clwyd, 16 February 2000,Official Report, columns 610–12W, on household incomes, if he will list the factors underlying the increase in household income in the north between 1997 to 1998 and 1998 to 1999; and if he will also estimate for the North East region the figures given for the years quoted in his answer. [112308]
§ Miss Melanie JohnsonThe information requested falls within the responsibility of the Director of the Office for National Statistics. I have asked him to reply.
Letter from Tim Holt to Mr. Jim Cousins, dated 28 February 2000:
As Director of the Office for National Statistics (ONS), I have been asked to reply to your parliamentary question on the increase in household income in the north.155WThe Answer to the honourable Member for the Vale of Clwyd, Official Report, columns 610–12W mentioned that, like all estimates from sample surveys, the figures were subject to sampling variability, and that care should be taken in interpreting short-term changes in the data. The apparent large increase in income in the North shown by the Family Expenditure Survey between 1997–98 and 1998–99 is almost certainly due mostly to random sampling variability. The estimated increase is £58 a week but this is subject to sampling variability of plus or minus £49 a week, this is the 90% confidence interval. There is only a one-in-ten chance that the true figure lies outside the estimated range. Published regional breakdowns from the FES are now based on three-year averages to reduce their sampling variability.Figures for the Government Office Region of the North East are available only from 1994–95 and are shown below. The comments on sampling variability apply also to these estimates.
Gross weekly household income for the North East Region, 1994–95 to 1998–99 North East Gross weekly households income (£) 1994–95 310 1995–96 340 1996–97 320 1997–98 340 1998–99 400
An alternative source, the Family Resources Survey, would normally be preferred for regional breakdowns because of its larger sample, more than three times that of the FES. It was not used for this question because it does not provide as long a run of data as the FES and because results are not yet available for 1998–99.