§ Mr. CousinsTo ask the Secretary of State for the Environment if he will estimate the further margin of error relating to the tables in annex F of "Paying for Local528W Government" resulting from taking account of the regular 30 per cent. non-response rate to the family expenditure survey, stating for which bands of household income the data would be most affected.
§ Mr. Howard[holding answer 3 February 1988]: Direct evidence of the income of non-respondents is not available. Comparison with the 1981 census does provide an indication of the types of household whose response rate is below average (see "Statistical News No. 72," February 1986, HMSO). This evidence suggests that households with three or more cars and the self-employed are under-respresented. The response rate for the elderly aged over 70 is also significantly below average.
§ Mr. CousinsTo ask the Secretary of State for the Environment if he will publish the repriced tables in his Department's note on the relationship between local taxes and income and the distributional impact of the community charge alongside the results from the 1986 family expenditure survey, together with an explanation of how the repriced figures were derived and what margin of error applies to them.
§ Mr. Howard[holding answer 3 February 1988]: I am afraid the hon. Member's question is unclear. The family expenditure survey does not contain tables comparable with those in the note placed in the Library on 13 January. The table, however, shows the proportion of households in each band of net income as used in that note alongside the most nearly comparable table from the 1986 FES.
529W
1986 FES Table 1 DOE note of 13 January Disposable income Percentage of sample Weekly net income Percentage of sample £ £ Under 45 6.8 45–50 2.2 Under 50 5.4 50–55 1.9 55–60 1.9 60–65 2.5 65–70 2.5 50–75 12.0 70–80 4.0 80–90 4.1 90–100 3.6 75–100 10.3 100–120 7.0 120–140 6.5 100–150 15.7 140–160 6.7 160–180 5.8 180–200 5.5 150–200 15.9 200–225 7.3 225–250 6.0 200–250 13.6 250–275 5.0 275–300 4.2 250–300 10.2 300–325 3.3 325–350 2.7 300–350 6.5 350–375 2.1 375–400 1.4 350–400 4.1 400–450 2.4 450–500 1.4 400–500 4.0 500+ 3.0 500+ 2.3 Percentage of households in each income band
The aggregated FES sample used to exemplify the impact of the community charge has been repriced to 1986–87 prices using a number of appropriate indices. For example earned income in each quarter was updated by reference to the index of seasonably adjusted average earnings, water rates were updated by the water charges component of the RPI and local authority rates were repriced according to the actual level applying to each household in the sample. The margin of error which applies will vary according to the index used.
§ Mr. CousinsTo ask the Secretary of State for the Environment what is the sampling ratio achieved by the combination of family expenditure survey responses for 1981 to 1984 used in annex F of Cmnd. 9714 "Paying for Local Government" and subsequently(a) for Great Britain and (b) for England; and what margin of error should be allowed for in a sampling ratio of this kind.
§ Mr. Howard[holding answer 3 February 1988]: The tables in annex F of the Green Paper "Paying for Local Government" (Cmnd. 9714) have been superseded by those contained in the note I placed in the Library on 13 January showing the relationship between local taxes and income and the distributional impact of the community charge. The sampling ratio achieved by the family expenditure survey is around 6,000 out of 18} million in England. Using four years data gives a combined sampling ratio of around one in 750. It is not possible, however, to estimate the margin of error simply from the sampling ratio. This will depend on
- (i) the sample size
- (ii) the design of the sample
- (iii) the nature of the variable being measured