Producing State Level Estimates of Hours Worked
The ABS publishes estimates of total hours worked by employed Australians, which complement the employment rate as a measure of economic activity and employment in Australia. They are also used as a labour input for the productivity estimates in the Australian System of National Accounts. Recently, there has been increased interest in using the these data to measure the impact of, and the recovery from, the global financial crisis and its effect on the Australian economy. In meeting user demands for finer level estimates, the ABS has improved methods of accounting for holiday effects in these time series and has published estimates at the state and territory level. These flow estimates of state level hours worked have been first published in the January 2011 Labour Force publication (cat no. 6202.0).
Accounting for holiday effects is involves a number of stages. In the context of actual hours worked data, holiday correction is a process of removing the effects of regular non-random events and is performed prior to adjusting the data for monthly seasonality. A holiday corrected series is one without known holiday impacts that is intended to reflect the hours worked in a working week unaffected by holidays. The overall process involves taking holiday corrected stock series of hours worked (derived from the ABS Monthly Population Survey) to produce monthly flow estimates of working weeks unaffected by holidays by using a linear interpolation methodology. Holiday effects are then reintroduced into these estimates to give monthly flow estimates of actual hours worked.
The effect of state specific holidays are better estimated using state level series, and such estimates were introduced earlier this year. Likewise, the effects of one-off events can be seen more clearly in these series when the impact of the event is localised geographically. For example, the estimates of hours worked for Queensland last January were noticeably affected by the wide-spread flooding occurring at that time.
The methodology for estimating and accounting for moving holidays in these series was presented at the Australian Statistical Conference in December 2010. For more information, please contact Philip Crouch at email@example.com.