The methodology in coming up with a wheat production estimate varies between agencies says the Australian Bureau of Statistics.

ABS says variation in crop estimates is normal

Crop forecasts and estimates will always vary between different forecasters according to the Australian Bureau of Statistics.

IN THE MIDST of the controversy surrounding a near five million tonne discrepancy in the size of the national 2016-17 wheat crop between two government agencies the Australian Bureau of Statistics (ABS) has said it is normal for estimates to vary.

The ABS has come up with a preliminary figure for the 16-17 wheat crop of 30.35 million tonnes, compared to national grain forecaster ABARES, which issued a figure of 35m tonnes.

This divergence has raised eyebrows in the grain industry, with farmer groups saying opaque stocks reporting meant it was difficult for the market to gauge accurately how much grain is around and set a fair price.

Kirrilie Horswill, of the ABS, said data for the ABS estimates came through a survey of Australian farm businesses.

She said the estimates would always vary, not only from ABARES but from the numbers put out by private forecasters due to the way calculations were done.

“The ABS estimates are based on a survey of producers / farm businesses meeting certain criteria and on the goodwill and cooperation of our respondents,” Ms Horswill said.

She stressed there were differences in conducting crop forecasting, which is ABARES’ primary focus, compared to coming up with a post-harvest estimate.

“There are quite different methodologies behind crop forecasting models which are looking ahead, rather than backward and based on expected outcomes.”

For wheat, she said participating farm businesses were asked for estimates of their production and area grown and total base estimates were made off these responses.

Ms Horswill acknowledged there were limitations to the methodology.

“Not everyone responds in a timely way (or at all) and we adjust for this as well as weighting up for farm businesses that we haven’t included in sample,” she said.

She said this meant there were usually changes between the preliminary and final estimates.

“We have more information available to produce our final estimates than preliminary estimates due to more responses being in.”

Along with the estimates, Ms Horswill said other information, such as receival data, exports, crop reports, rainfall and temperature reports were consulted, but she said the final figures were predominately driven by the survey responses.

Would you like to comment on this article?
Thanks for providing feedback.