When It’s Not Business As Usual Michael Smedes National Accounts Branch Australian Bureau of Statistics
Overview Australian Context Dramatic Price Change Misbehaving Models Government Stimulus Gaps & Weaknesses Exposed Time Series Challenges Conclusion
Australian GFC Context
Australian GFC Context
Australian GFC Context
Dramatic Price Change Exchange rate and commodity price volatility Coal: +200% 2008, – 60% 2009 Iron Ore : +70% 2008, – 33% 2009 Australian to US $: 98c July 08, 61c Oct 08, 84c Sept 09 Price Indexes Point in time pricing Timing & flow through Aligning prices and values Representativeness
Misbehaving Models Finance Industry Cash rate moves Lack of timely data Use of imputes and trended estimates Cash rate moves 7.25% March 08, 3.25% Feb 09 Nature of the service provided Price and volume components Investment banks fees structures
Government Stimulus Economic stimulus packages Around $60 billion in various initiatives Cash payments, tax breaks, subsidies, infrastructure Determine conceptual treatment Ensure data sources will capture Informing users on treatment Informing users on impacts?
Gaps & Weaknesses Exposed Example - Inventories Difficult to measure Dramatic changes in volumes and prices Hasn't followed historical patterns But very important! Discrepancy can arise right when users need quality data the most Quarterly Supply Use
Time Series Analysis Impact of unusual events Large extreme irregular Change to the seasonal pattern Change in the level of the series Concurrent Seasonal Adjustment vs Fixed Factors Key Issues Judgement Consistency and coherency Keeping users informed
Large Extreme Irregular
Conclusion Estimation or imputation are risky when economy is turning Gaps and weaknesses are exposed Where we do have quality data sources events were picked up Uncertain times can tell us a lot about the quality of our data Judgement & an understanding of the data are critical Coherency and consistency are tested Importance of user communication