IMPACT OF THE FLOODS ON THE LABOUR FORCE SURVEY
Due to flooding in Queensland, operational difficulties were experienced in conducting the Labour Force Survey in January 2011. There was a larger than usual number of households in the Queensland sample which could not be interviewed in the fortnight commencing 9 January 2011. Investigations were conducted to identify possible impacts on the estimates for this period.
The weighting and estimation methods used in the Labour Force Survey are designed to compensate for sample loss by inflating the survey weights for collected responses up to the population benchmarks and applying composite estimation to produce the current month's estimates. While the disruption to survey operations will have slightly reduced the quality of some Queensland estimates, the impact on the estimates is not statistically significant for most series. However, a noticeable impact on aggregate monthly hours worked was observed; see further details below.
Due to the sample loss noted above, there will be increased volatility in the Queensland estimates, particularly in the original and seasonally adjusted estimates. Given increased volatility, the ABS continues to encourage users to focus on trend estimates in monitoring the underlying level of series.
AGGREGATE MONTHLY HOURS WORKED
The method for estimating aggregate monthly hours worked accounts for systematic calendar-related effects (such as seasonality and holidays) and uses data from the reference week to estimate aggregated hours worked for the calendar month. When an unusual event occurs outside the reference week, the effects of this event will not be captured by the survey, and as a result, estimation of aggregated monthly hours worked is problematic. Specifically, the effects that the flooding in Queensland may have had on hours worked in the latter weeks of January may not be well captured by available survey responses and so the aggregate hours worked estimates for January 2011 should be used with caution.
For aggregate hours worked estimates for Queensland, a large extreme correction has been applied to stabilise the trend estimate. A similar treatment has been applied to the Australian total series. The ABS will continue to monitor movements in these series.
A large extreme correction can be applied to stabilise the trend estimates and seasonal factor estimates so as to reduce bias caused by an unusually large or small original estimate. It can also reduce subsequent revisions to seasonally adjusted and trend estimates. Details on the use of large extreme corrections to stabilise estimates of the trend and seasonal pattern during unusual events can be found in the feature article When it's not "Business-as-usual": Implications for ABS Time Series in the August 2009 issue of Australian Economic Indicators (cat. no. 1350.0).
Further information on the methodology used to produce the aggregate monthly hours worked estimates is available on the ABS website in Information Paper: Expansion of Hours Worked Estimates from the Labour Force Survey (cat. no. 6290.0.55.001).
The impact of the floods on the Queensland estimates will vary across regions, as a result, the relative standard errors (RSEs) for estimates of employment and unemployment for some Queensland regions will be higher than normal. The RSEs for the Darling Downs-South West and Ipswich City Statistical Regions are expected to be approximately 50% higher, while the RSEs for the Brisbane City Inner Ring Statistical Region will increase by approximately 25%. The Brisbane City Outer Ring, West Moreton and Mackay-Fitzroy-Central West Statistical Regions will have RSEs approximately 10% higher. All other regions have minimal differences.
For further information about this article contact Labour Force Estimates on Canberra 02 6252 6525, or via email <email@example.com>.