5.1
In the inquiry's terms of reference, the committee was tasked with inquiring into the accuracy of statistical data collected by Dairy Australia and the Australian Bureau of Statistics (ABS). The statistical efforts of these two organisations are discussed below, followed by a discussion of key views and concerns raised by inquiry participants. The chapter also discusses industry statistics provided by the Queensland Dairy Accounting Scheme.
Dairy Australia
5.2
Dairy Australia publishes a number of statistical updates throughout the year, including:
a monthly production and sales summary report;
a monthly production inputs update providing information on farm inputs, including water, hay, grain, fertiliser and the weather;
a monthly exports report on various Australian dairy export products such as butter, milk, cheese and whey products; and
a physical and financial analysis of 250 farms under the Dairy Farm Monitor Project to enable farmers to compare their farm performance.
5.3
The main concerns raised by inquiry participants, including Dairy Australia itself, centred on the monthly production data and the information published under the Dairy Farm Monitor Project. These are discussed in further detail below.
Production statistics
5.4
Dairy Australia provides monthly milk and manufactured dairy products production estimates based on data voluntarily shared by processors.
5.5
In its submission to the inquiry, Dairy Australia noted four key issues with its data collection: late delivery of data; quality issues with data; unsuccessful recruitment of new farming entities; and regional misallocation of milk production due to inadequate reporting.
5.6
As a result of these issues, Dairy Australia stated that:
… data collection has become increasingly difficult to maintain total coverage of milk production as the percentage of processors contributing data to the program reduces and new processors enter the industry who are not yet providing data to the DA [Dairy Australia] collection. Unprecedented turnover in supply contracts between different processing companies and some new entrants in the processing market is also making the collection more difficult.
5.7
Recognising these difficulties, Dairy Australia has undertaken a number of measures to enhance its milk production data accuracy. For example, in its submission to the inquiry it states that it has improved its relationship with processors to increase their participation, enhanced its reconciliation procedures, commenced cross-referencing of its data, and added a disclaimer noting the limitations of the data and possible retrospective adjustments.
Dairy Farm Monitor Project
5.8
The Dairy Farm Monitor Project (the Monitor Project) provides a yearly analysis of 250 farms across Australia, with Dairy Australia stating that it 'informs decision making and prioritisation by key stakeholders across the industry including Dairy Australia itself, government bodies and other stakeholders in the industry'.
5.9
Each jurisdiction prepares its own annual report under the Monitor Project with Dairy Australia publishing them on its website. This enables dairy farmers to compare their farm performances against other industry participants, and to identify areas for improvement.
5.10
Although the various publications under the Monitor Project state clearly that they do not represent population averages, as participant farms were not chosen based on random sampling of the population, this was a key criticism of the datasets it provides. These criticisms expressed by inquiry participants are discussed in greater detail later in this chapter.
Australian Bureau of Statistics
5.11
The ABS is Australia's national statistical organisation, and provides official statistics on a wide range of economic, social, population and environmental matters of importance to Australia.
5.12
Primarily through its Agricultural Statistics Program, the ABS produces a range of data on the dairy industry. Specifically these datasets are:
Number of animals on holding and unit counts – Agricultural Commodities, Australia (catalogue number: 7121.0)
Value of milk produced – Value of Agricultural Commodities Produced, Australia (catalogue number: 7503.0)
Dairy and related products price index – Consumer Price Index (catalogue number: 6401.0)
Input to manufacturing dairy cattle farming price index – Produce Price Indexes (catalogue number 6427.0).
5.13
In addition to the above datasets, monthly data are available on international imports and exports of dairy products. Further, estimates regarding employment within the dairy industry are also available from the Census of Population and Housing.
5.14
The ABS states that 'extensive effort' is invested in form design, sample selection, collection, and processing procedures to ensure high quality statistics, and that it utilises a specific framework, the Data Quality Framework, to assess data quality based on its 'fitness for purpose'. In addition, each area works with methodological and technical experts to reduce the four principal influences on data quality: respondent error; processing error; partial or non-response; and undercount.
5.15
The ABS also utilises collaborative relationships with various organisations to deliver its outputs. It works closely with the Department of Agriculture, Water and the Environment (the Department); the Australian Bureau of Agricultural and Resource Economics and Sciences (ABARES), rural research and development corporations, and various industry organisations.
5.16
In its submission to the inquiry, the ABS noted that it has also recently established a Regional and Agricultural Statistics Advisory Group comprising senior figures across government and industry. The group's role is to advise the ABS on the future direction of statistics related to the agricultural industry, land and resource use, community-place based socio-economic conditions, and environmental economic accounting.
Queensland Dairy Accounting Scheme
5.17
The Queensland Department of Agriculture and Fisheries (DAF) provides a ‘comparative analysis service available to Queensland dairy farmers to monitor their production and financial performance over time’ through the Queensland Dairy Accounting Scheme (QDAS). Established in 1976, QDAS data includes ‘milk production, stock numbers, income, cash expenses, non-cash inputs such as family labour as well as tonnages of feed used, conserved and stored’.
5.18
In its submission to the inquiry, DAF noted that:
Farmer participation in QDAS is voluntary and free, with individual farm data remaining confidential and only group averages being published. Sixty farms contributed data to QDAS in 2018–19 which represents 19% of Queensland’s dairy farms and 26.5% of Queensland’s milk production. As well as this large sample size, another strength of QDAS is the continuing cooperation between farmers and QDAS, with many farms consistently contributing data for more than 20 years.
5.19
QDAS provides data to Dairy Australia for its web-based farm business performance tool DairyBase and a QDAS report is published by DAF each year.
Views on statistical data collected
5.20
There were mixed views across inquiry participants on the accuracy, transparency, and timeliness of industry statistics. The discussion below covers the key issues raised in submissions and at the inquiry's public hearings.
5.21
A number of inquiry participants were confident in utilising the information prepared by the ABS and Dairy Australia. For example, in its submission to the inquiry, the Department stated that it relies extensively on data collected and reported by both the ABS and Dairy Australia. Further, the Department indicated it has a high degree of confidence in the data these organisations collect and release, and consider their datasets to be the most accurate and comprehensive available. In conclusion, it said:
Without the data provided by ABS and DA, ABARES ability to understand, monitor and forecast the performance of the dairy sector would be severely inhibited.
5.22
In his evidence to the committee, Mr Peter Gooday, a first assistant secretary from ABARES within the Department, highlighted the role of Dairy Australia in providing a population listing, and emphasised his confidence in the estimations based on sampling from it. Specifically he said:
We [ABARES] run a survey of dairy farmers each year; there are a bit more than 300 in our sample. That sample is drawn from a population list provided to us by Dairy Australia. It's a stratified random sample—so we break the sample up into different size categories of dairy farmers and then select randomly from those. The only qualification is that you're on the Dairy Australia list and that you come up randomly when we select. We do like to keep farms in the survey for a couple of years, because there are some pretty big startup costs to being involved in the survey. As I said, we've got about 300 sample points; they're spread across dairy areas. We have a sampling strategy that gives us enough farms in different size categories so that we're confident in coming up with an estimate of the averages for that dairy region.
5.23
The Queensland Dairyfarmers' Organisation (QDO) suggested that the survey methods employed by ABARES allowed for an accurate reflection of dairy farms in Queensland. The QDO stated that:
When organisations such as ABARES undertake farm surveys, they ensure that the population is stratified to break up farms into different size categories. They decide how many farms to choose in different size categories. They select the farms they need to survey with reserves for each selected farm which are similar to the selected farms.
In addition, each farm surveyed has a different weight to reflect how many farms they represent. As a result, the estimates made from surveys have validity as indicated by the residual standard errors (RSE’s) for each data point. Also, it means that the average farm in the survey matches the average farm in real life.
5.24
Mr Rob Miller, of Narrawilly Dairy, observed that the ABS and Dairy Australia provided trend information about, and for, the dairy industry, reflecting performance and profitability. Mr Miller held no concerns about the performance of the ABS, concluding that it does 'good work'.
5.25
In its submission, the Tasmanian Government stated that it has no concerns with the data collected and reported by Dairy Australia and the ABS, noting that 'the data is more comprehensive and accurate than the data collected and publicly reported for many other industries'.
5.26
International banking corporation Rabobank analyses issues and trends in the global food and agribusiness sectors through its RaboResearch arm and shares its findings with its clients. In its submission Rabobank noted that:
Key to research and analysis of the global food and agriculture sector is the ability to collate supply chain data from all parts of the world, which feeds into a global picture. Rabobank utilises Dairy Australia data to analyse these markets and sectoral trends, and finds the data comparable to that provided in other jurisdictions.
5.27
A number of other contributors to the inquiry also recognised the value provided by the statistical information prepared and released by Dairy Australia and ABS, but also raised a number of concerns. For example, Dairy Connect, an industry body representing the value chain of the Australian dairy industry and other industry stakeholders, did not question the accuracy of the data provided by the ABS and Dairy Australia; however, it stated it would prefer greater transparency in the data's collection to 'better understand the reason for the collection of information and what it will be used for'.
5.28
The national advocacy body representing dairy farmers across the six dairying states, Australian Dairy Farmers (ADF), stated that Dairy Australia has secured and maintained the trust of industry to manage data in a confidential manner, and that is it is uniquely capable of aggregating data at the industry level. Notwithstanding, the president of ADF, Mr Terry Richardson, in his opening statement to the committee commented on data limitations and variability in dairy industry statistics. He said:
… we [ADF] recognise these are limited by sample size and data collection variability. There is variation across the agencies, between states and territories and between [commodities], and our suggestion is for the collection agencies to come together to determine ways to streamline and incentivise contributions.
5.29
In relation to Dairy Australia's Monitor Project, ADF stated that, although these datasets are invaluable as a management tool and inform industry participants on the key drivers of farm profitability and performance, they are not statistically representative of Australian dairy farmers and should never be represented as such.
5.30
A number of inquiry participants expanded on lack of representation as a key deficiency in industry statistics. For example, the peak industry body representing Queensland dairy farmers, QDO, was critical of the data provided through the QDAS. In its submission to the inquiry QDO stated that:
The data collected on farm performance in Queensland in the Queensland Dairy Accounting Scheme (QDAS) does not give a representative sample of farms in Queensland. The average Queensland farm is around 1 million litres whereas the average farm in Dairy Australia related surveys is 1.6 million litres.
Many smaller farms are excluded, and it appears are not encouraged to continue or not approached. The population of the data is not stratified. They do not select an unbiased sample but rather a biased one. No weights are used. As a result, the estimates are not representative of average Queensland dairy farms.
Farms for QDAS are selected by direct approach by an officer employed by Dairy Australia. QDO has received numerous reports of farmers, with below average figures or size, being actively dissuaded from participating in QDAS.
5.31
Although conceding he hasn't had a lot of experience with QDAS, in evidence provided to the committee Mr Graham Forbes from Dairy Connect stated that:
I think it [QDAS] is a very good system and I think the information is collected very accurately. Probably my concern about the QDAS pricing is that I think it focuses on some very good farmers and farmers that are prepared to put their information in, and it may not give a total view of the whole industry as well as we'd like to see.
5.32
Responding to these criticisms that QDAS is voluntary and does not include smaller farms, Mr Graeme Busby, a former Queensland Department of Agriculture and Fisheries official, indicated in his evidence to the committee that 20 per cent of the farms in the QDAS produce less than 750 000 litres of milk.
5.33
QDAS in their submission to the inquiry said that while the ‘milk production per farm in QDAS is higher than the average production of a Queensland dairy farm’, the data are analysed using the stratification method which proportionally weights the source data according to the percentage of Queensland farms which occur in five size groups. Dairy Australia provides the percentage of farms in the five size groups.
5.34
United Dairyfarmers of Victoria (UDV), the peak representative body for dairy farmers in Victoria, also raised lack of representation in relation to Dairy Australia statistics. In its submission it stated that 'many of Dairy Australia's statistics used to determine industry performance are not representative of the entire industry'. The UDV went on to claim that there are not enough farmers providing data and, hence, the sample size is too small and contains a disproportionate number of high performing farms. In conclusion the UDV states that it 'believes these sample sizes need to be expanded to better reflect the broader industry'.
5.35
UDV also noted that Dairy Australia commonly relies on data to be voluntarily provided by processors and food safety regulators and, hence, its data can only be as accurate as what is voluntarily reported. UDV believes this arrangement is inadequate and, to improve transparency and reliability, recommends that processors be required to report on milk production volumes and the number of dairy farms by region.
5.36
Mr Kevin Ashworth, representing Farmer Power, questioned the validity of the target for farm profitability articulated in the Australian Dairy Plan 2020–2025. Mr Ashworth contended that the Dairy Plan states that the target is 'based on a national analysis of farm business performance data monitored by Dairy Australia' but that 'the participant farms were not chosen to be statistically representative'. He argued that this statement 'goes to the very core of the accuracy of statistical data collected by DA'.
5.37
Similarly, Far North Coast Dairy Industry Group (FNCDIG) indicated that smaller, or struggling, farms are less likely to reveal their information and, hence, data collected through voluntary surveys represents the 'top end of the industry'. FNCDIG concludes that '[i]f data collected was a true reflection of the whole industry we would see a very different picture of where the total industry is situated'.
5.38
In addition to these accuracy issues, Dairy Connect and UDV also criticised the timeliness of the data provided. Dairy Connect stated that Dairy Australia needed to improve the timely dissemination of its data, and UDV claimed that it is often released with a two to three month lag time.
5.39
In responding to criticisms and adverse commentary expressed by a number of inquiry participants, Dairy Australia stated that:
[a]ll statistics collected and published by Dairy Australia … include quality control measures to ensure that the data is as complete and accurate as possible. It is important to note that much of the key data provided by farmers and processors is provided on a voluntary basis. The limitations of this voluntary data are recognised with appropriate disclaimers included in relevant reporting. Further to [Dairy Australia's] original submission, the Committee may wish to consider mandating data reporting requirements if more complete information is required.
The Dairy Farm Monitor project is a comprehensive monitoring program of farms in all eight dairying regions that have volunteered to take part. The farms demonstrate wide variations in performance in each region, and due to the intensity of data collection, are a small subset of farms. The data is used for benchmarking and comparison purposes, not as a statistical representation of performance in each region. We recognise the important contribution of state governments in some regions that resource on-farm activities and ensure there is strong data integrity.
Committee view
5.40
The collection of accurate statistical data is of critical importance to the understanding of trends in milk production and sales, costs of production and industry performance. An accurate measure of the costs of production is also a prerequisite to determine a fair price for milk.
5.41
To improve data accuracy the committee endorses mandating the timely provision of milk production data and reporting by the postcode of origin of the source of the milk.
5.42
The committee notes a high level of satisfaction among inquiry participants with data provided by the ABS, perhaps with the proviso that it and other government agencies collecting statistical data be more forthcoming to industry participants on why they are collecting data and the purposes for which the information will be used.
5.43
The committee recommends that the government expand the representative sample of statistical information collected by Dairy Australia so that it better represents the population of Australian dairy farms.
5.44
The committee recommends that the government maintain a single authoritative measure of the cost of production of milk for the eight regional milk districts.
5.45
The committee recommends that the government consider approaches to improve the timely provision of milk production data to Dairy Australia by processors.
5.46
The committee recommends that the government consider approaches to enhance the information processors provide Dairy Australia regarding the locations in which they source their milk.