SECTION 5: DATA GAPS AND REMEDIES
In section 4 it was explained how statistics that are readily available can be used to construct Australian cultural and creative activity satellite accounts that meet many of stakeholders’ needs. The next questions to consider are:
- What additional data is required to meet the remaining needs?
- How could improvements be made in areas of the satellite accounts where there are reservations about data quality?25
The table below identifies the key data gaps and quality issues identified in section 4, and then suggests how these issues could be dealt with. In general, the solutions range from running a new survey, modifying an existing survey (e.g. by boosting the size of its sample or changing what data is collected), or making greater use of unit record data that is available within current collections but which would require a significant amount of processing.
|Data gap or quality issue||Potential remedies|
|1. Cultural and creative activity in the ANZSIC groups below cannot be identified directly from input-output tables.|
1320 Leather Tanning, Fur Dressing and Leather Product Manufacturing
2029 Other Ceramic Product Manufacturing
2599 Other Manufacturing n.e.c.
6639 Other Goods and Equipment Rental and Hiring n.e.c.
6962 Management Advice and Related Consulting Services
7211 Employment Placement and Recruitment Services
7212 Labour Supply Services
7299 Other Administrative Services n.e.c.
9499 Other Repair and Maintenance n.e.c.
9551 Business and Professional Association Services
9552 Labour Association Services
9559 Other Interest Group Services n.e.c.
3020 Non-Residential Building Construction
3109 Other Heavy and Civil Engineering Construction
4520 Pubs, Taverns and Bars
4530 Clubs (Hospitality)
7510 Central Government Administration
7520 State Government Administration
7530 Local Government Administration
8101 Technical and Vocational Education and Training
8102 Higher Education
|1.1. For Groups A-E, exclude the ANZSICs from the industry supply chains component of the satellite account. Some of the activity in these ANZSIC classes would then be automatically captured in the employment of people in cultural and creative occupations (component 2 of the satellite account). This option can be exercised for any ANZSIC individually.|
1.2. For Group A only, a list of Australian Business Numbers (ABN) is required for significant entities that belong to the ANZSIC classes and predominantly undertake the cultural and creative activities within them. This list could be compiled through desktop research and consultation with government or industry associations. Their share of total activity in the ANZSIC class would then be estimated using entity level data from existing ABS surveys or ATO business tax data.
1.3. For Group B only, Building Approvals and Building Activity Survey records might be used to identify significant construction work on cultural facilities with some help from internet research. The value of work done on these significant jobs should be completely enumerated in the Building Activity Survey. The value of work done on significant jobs could be used to assign a share of total construction activity in input-output tables. Potentially the value of work done on less significant jobs could be separately modelled from Building Approvals.
1.4. For Group C only, an estimate of patron spending during live entertainment events could potentially be collected via a new survey. However, this would be particularly challenging to accurately measure since patrons will have difficulty recalling their expenditure ex-post, and venues will not necessarily be able to provide sales disaggregated this way. Several non-ABS studies have been undertaken on this topic26 and the resulting data was not of sufficient quality for a satellite account.
1.5. For Group D only, administration activity on cultural or creative policies and programs could be estimated by surveying specialist government units (e.g. offices for the arts). Employment or wages in these units would be divided by the ANZSIC class totals to derive a share that can be applied to the input-output table aggregates.
1.6. For Group E only, the share of activity in education and training in cultural and creative fields could be estimated using student course data from the National Centre for Vocational Education Research and Department of Industry, Innovation, Science, Research and Tertiary Education.
|2. Earnings of multiple jobholders, by occupation and industry, have not been collected since SEARS 2007. Updated data would eventually be needed for the occupations component of the satellite accounts.||2.1. Multiple jobholders are identified in the Labour Force Survey each month. Their working hours and earnings in secondary jobs are likely to be collected each August starting in 2014, through the Characteristics of Employment supplementary survey. Questions would also need to be added to capture the industry and occupation of secondary jobs.
2.2. Personal Income Tax (PIT) and Pay As You Go (PAYG) Withholding data from the ATO may be able to fill this need in the longer term, given taxpayers identify separate ABNs for each employer. PIT and PAYG data would need to be analysed in detail in order to determine their suitability for this purpose.|
|3. The GSS’ state and territory estimates of volunteer hours for ‘arts/heritage’ organisations have high relative standard errors (typically 25-50%). ||3.1. Boost the GSS sample size. Only a small percentage of people reached by this survey are volunteers and therefore a substantial increase (e.g. 100%) in the GSS sample would be needed to make a noticeable improvement in the quality of volunteering data. This is likely to make it prohibitively expensive.|
|4. Detail on income, expenses, assets and liabilities are infrequently collected for many parts of the cultural and creative industry supply chains. When surveys are run, they often do not provide estimates for ANZSIC classes or as much detail as sought by stakeholders.||4.1. Boost the samples of EAS surveys on industries that include cultural and creative activities. This would enable income and expense estimates to be produced for selected groups of ANZSIC classes as well as at higher ANZSIC levels. These sample boosts would be quite expensive. The collection of assets and liabilities would require a further allocation of resources, as would the collection of a wider range of income and expense details than on the current standard EAS questionnaire.|
4.2. Business Income Tax (BIT) data from the ATO may be able to fill this need in the longer term. BIT data contains detail on the income and expenses of a business by type, as well as assets and liabilities. BIT data is not currently used for this purpose and would need to be analysed in detail in order to determine its suitability.
|5. Published R&D expenditure does not provide estimates for many of the ANZSIC classes which comprise the cultural and creative industry supply chains. ||5.1. Use R&D expenditure data to compile estimates for the aggregations of ANZSIC classes which make up the cultural and creative industry supply chains, both for Australia and a split by state and territory. A significant amount of compilation work would be required to produce and check the confidentiality of data tailored in this way.|
The ABS Quality Framework
(cat. no. 1520.0) considers data quality using the criteria: relevance, accuracy, timeliness, accessibility, interpretability and coherence.
For example, Ernst & Young (2011), The Economic Contribution of the Venue-Based Live Music Industry in Australia
, report for the Australasian Performing Right Association, <http://issuu.com/apraamcos/docs/livemusic?mode=window