FORTHCOMING CHANGES


BACKGROUND

Following the completion of the 5 yearly Census of Population and Housing, the ABS reviews the LFS sample design. The review ensures the survey accurately reflects the geographical distribution of the Australian population, and remains efficient and cost effective.

The outcome of the review based on 2006 Census data is being implemented over the period November 2007 to June 2008. In developing the 2006 sample design, the decision was made to achieve cost savings by taking advantage of the sampling efficiencies related to the introduction of composite estimation. This enabled an 11% reduction in the LFS sample with only minor reductions in data quality relative to the previous design. Full details of the 2006 sample design are presented in Information Paper: Labour Force Survey Sample Design (cat no. 6269.0).


RECENT DEVELOPMENTS

The ABS is facing a tight budget situation in 2008-09, which has led to a range of reductions in the ABS work program. One of the reductions is that from July 2008 the sample size of the LFS will be reduced by 24% when compared with the June 2008 sample being implemented under the 2006 sample design. The ABS is implementing this sample reduction in such a way that the sample can be easily increased again in the future should the ABS funding position change.


IMPLEMENTATION

The ABS will fully implement the sample reduction in July 2008. Implementing the full reduction in a single month will mean less common sample between June and July, and hence the standard error on movements will be slightly larger than if the reduction was phased-in. However, in order to maximise the savings in 2008/09 the reduced sample needs to be fully implemented from July 2008.


SAMPLE SIZE

Table 1 presents the expected number of persons in the LFS sample for June 2008 under the 2006 sample design, and the proposed LFS sample for July 2008 following the sample reduction. It should be noted that the sample size will gradually increase between sample redesigns due to population growth prior to the next sample redesign following the 2011 Census.

Table 1. Estimated Fully Responding Person Records
	June 2008	July 2008
New South Wales	12 900	9 600
Victoria	11 500	8 700
Queensland	9 600	7 300
South Australia	6 400	4 900
Western Australia	6 200	4 600
Tasmania	3 500	2 700
Northern Territory	2 100	1 700
Australian Capital Territory	2 200	1 600
Australia	54 400	41 100

QUALITY OF ESTIMATES

The new sample, while smaller, will still be representative, with selections made in all parts of Australia. There will be increased volatility in the estimates, particularly the original and seasonally adjusted estimates, but this volatility will be random. Given the increased volatility in the original and seasonally adjusted estimates, the ABS would continue to encourage users to focus on trend estimates as the increased volatility seen in the original and seasonally adjusted estimates will be dampened through the 'trending' process.

The most common way to quantify the volatility is to examine the relative standard errors (RSEs). Table 2 below shows the target RSEs following the 2006 sample design and introduction of composite estimation and the target RSEs for the sample from July 2008.

Table 2. LFS Relative Standard Errors
	Employment RSE		Unemployment RSE
	2006 Target RSE(a)	2008 Target RSE	2006 Target RSE(a)	2008 Target RSE
	%	%	%	%
New South Wales	0.8	0.9	5.0	5.7
Victoria	0.8	0.9	5.3	6.1
Queensland	0.9	1.0	5.7	6.6
South Australia	1.1	1.3	6.8	7.7
Western Australia	1.0	1.1	7.1	8.1
Tasmania	1.5	1.7	8.3	9.5
Northern Territory	2.6	3.0	13.4	15.1
Australian Capital Territory	1.3	1.5	13.9	16.0
Australia	0.4	0.4	2.6	2.9
(a) Due to recent improvements, the 2006 RSEs for the NT and WA and also for unemployment in SA differ from those published in the information paper Labour Force Survey Sample Design, Australia. (cat. no. 6269.0).

Overall, the RSEs for estimates of employment and unemployment at the national, state and territory level are expected to be approximately 15% higher than those expected from the 2006 sample design. The impact of the increased RSEs is best demonstrated by consideration of the confidence intervals surrounding the respective estimates. By way of example, say the estimate for employment is 10,000,000. Under the 2006 design, there would be 19 chances in 20 that the real value falls within the range 9,932,600 to 10,067,400. With the reduction in sample from July 2008 that range will increase to 9,922,800 to 10,077,200. Similarly for unemployment, if the estimate is 500,000 then under the 2006 design, there would be 19 chances in 20 that the real value falls within the range 474,000 to 526,000. With the reduction in sample that range will increase to 470,200 to 529,800.


LFS PRODUCTS

Key monthly estimates from the LFS are published in Labour Force, Australia (cat. no. 6202.0). As well, there are a range of other products presenting detailed estimates from the LFS, which are available at the ABS website, www.abs.gov.au.

The sample reduction will also increase the standard error on the detailed estimates in these products by approximately 15%. Some of the estimates in these products, such as detailed industry estimates for small states and territories, have always had high standard errors and the sample reduction will increase this number.

To assist users in understanding the quality of the estimates published, the ABS will be producing a new standard error model for the LFS. This model will be used to populate the standard error tables in this publication and to annotate those estimates in our products with an RSE of 25% or higher. Estimates with an RSE of 25% or higher should be used with caution. This new model will also be incorporated into the spreadsheet Labour Force Survey Standard Errors, Datacube (cat. no. 6298.0.55.001) to allow users to calculate the standard error for any LFS estimate.


ASSOCIATED SURVEYS

In addition to impacting on the LFS, the sample reduction will also affect the supplementary surveys which are conducted on part of the LFS sample and cover a range of different topics. Due to the infrequent nature of topics in the supplementary surveys, they do not benefit from the efficiency gains associated with composite estimation. As a result the sample reduction associated with the 2006 design is resulting in increased standard errors for these estimates. The further sample reduction to be introduced in July 2008 will increase the standard errors further. In combination, the sample reductions are expected to increase the standard error for estimates from supplementary topics by approximately 22% relative to the 2001 design. The level of disaggregation of estimates possible from these topics will be assessed as each topic is prepared for release. However, it is likely that the level of disaggregation will need to be reduced, especially for topics which relate to small sub populations.

The ABS also conducts a Multi Purpose Household Survey on part of the LFS sample each year. The impact of the sample reduction on the MPHS will not be as large as for the LFS or the supplementary topics as the MPHS sample size will be kept approximately the same after July 2008. Under the 2001 sample design, the MPHS sampled one-third of those dwellings in the LFS for the last month, which achieved a sample of 13,500 dwellings per year. Following the sample reduction, the proposal is to increase the proportion of dwellings sampled to 50% of those in the LFS for the last month. This proportion is expected to achieve a MPHS sample size of 13,000 dwellings per year.