Project Methodology

Constructing a dataset to build Australia’s evidence-base

Information in this report supersedes previous data and provides a more complete snapshot of Australia’s registered charities. Whereas previous reports[1] focused only on those charities which provided their Annual Information Statements (AIS) to the ACNC in time for analysis, in 2015 exclusions have been minimised and the best available data has been used. This is consistent with the ACNC’s goal of promoting transparency.  Whereas the 2014 Reports were based on data from approximately 70% of registered charities, the 2015 dataset includes information for around 95% of the 54,564 charities which were registered with the ACNC at the end of the 2015 financial year[2].

Note that because the data in this report is more complete and based on a different methodology to previous profiles of Australia’s charities, the data presented here is not directly comparable to that contained in previous reports. This dataset provides a new, more comprehensive baseline from which to analyse Australian charities now and into the future.

About the dataset

Data from the following sources was matched using Australian Business Numbers (ABNs):

  • The ACNC Register
  • The Australian Business Register
  • Charities’ Annual Information Statements.

Overall, there were 50,908 records containing information about the activities of 51,679 charities.[3] Each data source is detailed below.

The ACNC Register

The ACNC Register contains information provided by charities for the purposes of registering with the ACNC. This includes their legal name, ABN, date of establishment and their charitable purpose. The register is available publicly on the ACNC website.[4] The ACNC register is a ‘living’ dataset. It is regularly updated by the ACNC as new charities are registered or report a change of details, or as charities have their registration voluntarily or involuntarily revoked. Some data in this report, including charities’ date of establishment, and charitable purpose, is derived from the register, as this information was not captured in the AIS.

The Australian Business Register (ABR) dataset

The ABR contains details provided by Australian businesses and organisations when they register for an Australian Business Number (ABN). It includes information about whether or not charities are endorsed for tax concessions. Relevant ABR data were obtained by the ACNC and provided to the research team for matching to the ACNC Register and 2015 AIS data. In this report, ABR information is used to profile charities’ legal structure, and their DGR status. It is important to note that for some charities, ABR data may be incorrect or out of date compared to data held on the ACNC Register.

The Annual Information Statement (AIS) dataset

Most information in this report comes from information reported by charities to the ACNC as part of their AIS. The 2015 AIS relates to charities characteristics and activities in the financial year ending in 2015. The standard ACNC reporting period is the financial year from 1 July to 30 June. However, charities that use a different reporting period can apply for a ‘substituted accounting period’ (ACNC, 2015e). The time period reported on in the AIS is based on the reporting period used by the charity.

The AIS captures information including charity size, activities, beneficiaries, employee and volunteer numbers, and locations of operations. Since 2014, AIS information has included financial information. Not all registered charities were required to provide an AIS in 2015 (such as newly established charities), and not all providing an AIS were required to provide financial information (such as basic religious charities).

The dataset incorporates information from the 2015 AIS for 44,670 registered charities which had provided the required information to the ACNC by 19 September 2016. Among those who provided their 2015 AIS, there were 885 charities which reported their AIS as part of 114 groups. Each group is treated as a single charity for the purposes of the analysis, as it is not possible to disaggregate the data and assign values to each charity in the group.

Although the 2015 AIS provides the best source of information about a charity in 2015, some charities were not required to provide a 2015 AIS, while others were required to report but had not done so in time for the analysis. In these cases, proxy information for the charity was used.

Where proxy information was required, for the 6,238 charities for which 2015 AIS data was not available,[5] information was estimated based on the best available data about the charity which was held by the ACNC. In most cases information was drawn from the charity’s previous AIS (either the 2014 or 2013 AIS). Where information could not be drawn from a previous AIS, estimation processes were used, based on information in the ACNC Register. The approach to estimation is described in Appendix B :Further methodological details.

The composition of the dataset is presented in Figure 1.1.

Figure 1.1          Summary of data sources

Source Number of Charities Percent Cumulative Percent
2015 AIS 44,670 87.8 87.8
Group charities, 2015 AIS 114 groups* 0.2 88.0
2014 AIS 3,672 7.2 95.2
2013 AIS 500 1.0 96.2
Data estimated, where possible based on information in ACNC Register 1,952 3.8 100.0
Total 50,908 100

Notes: *885 charities reported as 114 groups

Some questions in the AIS were not answered by all charities, so the total number of responses to each question may vary. Non-response reduces the accuracy with which the findings represent the whole population of registered charities. Notwithstanding, the near complete nature of the data, and the approach to identifying and correcting errors means the dataset provides the most accurate and comprehensive information ever available about Australia’s charities.

The scope of the AIS dataset, and an indication of when reported or proxy data were used, is also provided in Figure 1.2.

Figure 1.2          Use of reported and proxy data

proxy-data

 

Financial information in the AIS

Since 2014, AIS information has included financial information for most (but not all) charities. Basic religious charities[6] are not required to answer the financial information questions on the AIS. Further, the ACNC obtains financial information for non-government schools from the Commonwealth Department of Education, rather than the AIS. Medium and large charities were required to submit their annual financial report as well as the AIS. A summary of circumstances for which financial data was estimated is in Figure 1.3.

Figure 1.3          Summary of reported and estimated financial data

reported-data

Use of the 2015 dataset

As explained above, information from the ACNC register, three years of AIS, and ABR data were combined to construct the most complete dataset possible: around 95% of registered charities. In most cases, data in the report uses the full dataset (50,908 cases). However, for some variables, proxy information could not be found and estimation was considered unreliable. Where this was the case, the figures are based on the fullest reported data possible, that is, data drawn from the 2015 with proxy information drawn from a 2014 or 2013 AIS, where available.

The 2014-2015 panel dataset

Due to methodological differences, data in this report should not be compared with data reported in the Australian Charities Report 2014. Instead, to assess change and show which charities were growing, a matched panel dataset was constructed, consisting of charities which reported Annual Information Statements in 2014 and 2015. Findings are in Section 13. Construction of the panel dataset is described in Appendix B: Further methodological details.

Data quality

The use of the best possible data for the charity where the 2015 AIS data was unavailable ensures accuracy. Data was only used from a previous AIS where 2015 data was unavailable. Throughout the process, the research team collaborated closely with the ACNC to refine handling of errors and to improve data accuracy. The data were cleaned prior to analysis using the rules and filters summarized in Appendix B: Further methodological details. This approach to identifying and correcting errors in the data improved standards of data accuracy. However, data which is self-reported by charities may contain some errors, despite the best efforts of the ACNC and research team to identify and address these.

[1] Previous reports using 2014 data include Cortis et al., 2015a; Cortis et al., 2015b; Cortis et al., 2016; Sayers and Mukherjee, 2016. Reports based on 2013 data include Knight and Gilchrist, 2014; Knight and Gilchrist, 2015a; Knight and Gilchrist, 2015b; McGregor-Lowndes and Crittall, 2015; VCOSS, 2015.

[2] Of registered charities, those that report to the Office of the Registrar of Indigenous Corporations (ORIC) were excluded (812). This is because they are regulated primarily by ORIC, and not required by the ACNC to submit an AIS ACNC. (2015a) 2014 Annual Information Statement guide. Available at: http://www.acnc.gov.au/ACNC/Manage/Reporting/2014Guide/ACNC/Report/2014AISGuide.aspx. Charities that failed to provide both their 2014 or 2015 Annual Information Statements, so were liable for revocation, were excluded (2080 charities). Further, 885 reported as part of 114 groups. For the purposes of this report, group reporting charities are treated as a single charity. As such, the total number of records in the dataset is 50,908, representing 51,679 individual charities.

[3] There were 885 charities which reported as part of 114 groups. Each group is treated as a single charity.

[4] Charities are able to withhold commercially sensitive details, if public release could cause harm to the charity or a person, or endanger public safety ACNC. (2016) Information on the ACNC Register: withheld information. Available at: http://acnc.gov.au/ACNC/FindCharity/About_Register/Withheld_info/ACNC/Reg/With_info.aspx.

[5] Charities did not report 2015 either because they were not required to, or in a smaller number of cases, because they had not done so in time for the analysis.

[6] Basic religious charities are those that are registered with the subtype of a purpose of advancing religion, that could not be registered under another sub type. Charities are not basic religious charities if they are incorporated, endorsed to receive deductible gifts, or if they fall into other categories. Please refer to the full definition in section 205-35 of the ACNC Act, which is replicated in Appendix D: Charities and the Charities Act 2013.