TREND ANALYSIS


TOTAL RETAIL

There have been five months of weak trend growth. This follows a slowing in the rate of trend growth from strong growth in August 2007 to no change in the trend estimate for February and March 2008.

INDUSTRY TRENDS

Food retailing (four months), Clothing and soft good retailing and Cafes, restaurants and takeaway food services (both two months) have had moderate trend growth. Household good retailing has had weak trend growth for four out of the last five months. After two months of no change in the trend estimate Department stores had weak trend growth between July 2008 and August 2008. Other retailing has had a decline in the trend estimate for three months after two months of weak growth.


STATE TRENDS

The trend estimate has been in decline in the Australian Capital Territory for five months. There has been no change in the trend estimate in New South Wales for three months after a decline in the trend estimate for five months. Victoria (three months) and Western Australia (two months) have had weak trend growth. All other states have had at least four months of moderate or strong trend growth with:

• Queensland has had two months of strong growth preceded by two months of moderate growth
• South Australia, moderate growth for five out of the last seven months preceded by five months of strong growth
• Tasmania, strong growth for the last five months
• the Northern Territory, moderate growth for six out of the last eight months preceded by six months of strong growth.

REVISIONS TO TREND

The chart above indicates the level of volatility in the seasonally adjusted series. With the new survey design, it is anticipated that the volatility level will increase from July 2008. This increased volatility will contribute to revisions to trend estimates that are larger on average than those produced under the previous sampling methods. The likelihood of revisions needs to be considered when analysing retail trends.

What-if analysis can be used to assess the reliability of the trend estimate at the end points of the series. The following what-if chart presents the possible change in the trend estimates under two different scenarios. The two scenarios for the next month's seasonally adjusted estimates have been derived from the 25th and 75th percentiles of the historical seasonally adjusted movement distribution. The historical seasonally adjusted movements have been adjusted to reflect the expected increase in volatility from the new survey design.

Note that the what-if analysis does not show the unknown impact of revisions to seasonal factor estimates that could arise when the original estimate for the next time period becomes available. For more information see the trend estimates section of the Explanatory Notes.

1 The September seasonally adjusted estimate of retail turnover is 1.39% higher than the July estimate.

2 The September seasonally adjusted estimate of retail turnover is 0.04% higher than the July estimate.