|Page tools: Print Page Print All|
IMPROVED ESTIMATES OF THE ANNUAL NATIONAL ACCOUNTS: RESULTS OF THE 2017 HISTORICAL REVISIONS
REVISIONS TO KEY NATIONAL ACCOUNTS AGGREGATES
Australian gross domestic product (GDP) was revised for most years of the revisions period (1960-61 to 2015-16). Graph 1 provides the overall impact of the revisions to annual percentage change in GDP in volume terms. The revised long-run growth rates of GDP were similar to those previously published. For 1960-61 to 2015-16, the average annual growth rate remained unchanged at 3.5%.
Graph 1. IMPACT ON GDP GROWTH, Volume measures
Revision triangles show how published estimates for a particular period were revised over time. The data and revision triangles for GDP are shown in Table 1. The growth rate of GDP remained unchanged in the latest four years: 2012-13 to 2015-16. In most respects, the picture of the economy shown by the revised estimates is similar to the picture shown previously. Additional revision triangles for selected components of GDP are provided in Appendix A.
REVISIONS TO INCOME FROM GDP
The revisions by Income category are presented in Table 2.
Compensation of employees: The revisions to the level of compensation of employees (COE), which begin from 2003-04, reflected the inclusion of updated source data from Labour Costs, Australia (MLC) (cat. no. 6348.0), Australian Industry (AIS) (cat. no. 8155.0) and the Australian Prudential Regulatory Authority (APRA). For 2006-07 to 2015-16, revisions to COE were of mixed sign; the largest downward revision was for 2013-14 ($0.4b), and the largest upward revision was for 2015-16 ($6.1b).
Taxes less subsidies on production and imports: Revisions to this category were mainly associated with the removal of fines, interest charges and provisions on late payment from the GST.
Gross operating surplus and gross mixed income: The revisions to gross operating surplus and gross mixed income (GOSMI) mainly reflected the improved picture of business profits based on the reconciliation of updated source data through the production, income and expenditure measures of GDP.
The revision triangle in Table 3 shows that the largest upwards revision to growth in COE was 0.7% in 2014-15, mainly due to revisions to wages and salaries from AIS. The upward revision of 0.4% in 2010-11 resulted from all data source updates.
The larger revisions to COE recorded in the 2009 to 2011 editions of the Australian System of National Accounts (ASNA) reflected a range of improvements to compilation methods including the implementation of the System of National Accounts 2008 (2008 SNA), the introduction of off-june reporting adjustments and the move to the Australian and New Zealand Standard Industrial Classification, 2006 (ANZSIC06) (cat. no. 1292.0).
REVISIONS TO EXPENDITURE ON GDP
The expenditure measure of GDP consists of three components – final consumption expenditure by households (HFCE) and government (GFCE); gross fixed capital formation (GFCF) by private and public sectors; and international trade (exports and imports). All three components of the expenditure measure were revised through the time series, with the largest revisions occurring in HFCE, private GFCF and international trade. The revisions by expenditure category are presented in Table 4.
HOUSEHOLD FINAL CONSUMPTION EXPENDITURE
Household final consumption expenditure (HFCE) was revised through the time series to incorporate updated data sources that are usually unavailable on an annual basis. The revisions to HFCE growth in volumes terms range from -0.3 percentage points to 1.0 percentage points.
Graph 2. IMPACT ON HFCE, Volume measures
The main data sources include retail sales data from Retail and Wholesale Industries, Australia, 2012-13 (RISWIS) (cat. no. 8622.0) and household expenditure data from the Household Expenditure Survey, Australia (HES) (cat. no. 6530.0). The list of primary data sources used to benchmark annual estimates of HFCE categorised by the Classification of Individual Consumption by Purpose (COICOP) are shown in Table 5.
Graph 3 shows revisions to components of HFCE by COICOP. Although HFCE was revised through the entire time series, the revisions prior to 2006-07 were minor. The revisions range from a downward revision of $2.2b in 2003-04 to an upward revision of $18.0b in 2012-13.
Graph 3. REVISIONS TO HFCE BY COICOP, Current Prices
The upward revisions to HFCE between 2006-07 and 2012-13 were predominantly driven by an increase in expenditure on goods as a result of the implementation of the RISWIS. The main HFCE categories that were revised upward include recreation and culture (driven by audio-visuals and garden supplies), furnishings and other household equipment, clothing and footwear, operation of vehicles and alcoholic beverages. Household consumption of food was revised downwards.
Household consumption of tobacco was revised from 2006-07 onwards due to a review of tobacco supply and use estimates based on data from RISWIS and tax administration datasets from both ATO and Customs.
Investigations into household and business purchases of motor vehicles showed that, in recent years, household purchases of motor vehicles were understated, and conversely, business purchases were overstated. The current HFCE estimate for purchases of motor vehicles utilises car sales data sourced from the VFACTS produced by the Federal Chamber of Automotive Industries, business car purchases from AIS and state government new car registrations. This change has resulted in upward revisions to HFCE from 2006-07 and downward revisions to private GFCF of machinery & equipment.
The implementation of new dwelling counts (owner occupied and rented) and rents paid estimates from the latest Census of Population and Housing: Mesh Block Counts (cat. no. 2074.0) drove the downward revisions to rent and other dwelling services.
GROSS FIXED CAPITAL FORMATION
The implementation of a number of methodological improvements resulted in revisions to gross fixed capital formation (GFCF). The revisions to GFCF also resulted in revisions to the estimates of capital stock and consumption of fixed capital. Additional details on the methodological changes to GFCF are provided in the Information Paper: Changes to the 2016-17 release of the Australian System of National Accounts, 2016-17 (cat. no. 5204.0.55.012).
Capital formation for Machinery & Equipment
The estimates of GFCF machinery & equipment (M&E) were revised downwards across the time series to more closely align with CAPEX and AIS data. In 2015-16, GFCF M&E was revised downward by $3.3b.
Graph 4. MACHINERY & EQUIPMENT GFCF, Current prices
Capital formation for Private Dwellings
Alterations and additions, a component of Private Dwellings GFCF, was revised based on data from the BIS Shrapnel survey on the Home Improvements Market in Australia. This survey was used to update an under-coverage factor for alterations and additions that are out of scope for the Building Activity, Australia (BACS) (cat. no. 8752.0) survey.
Graph 5. PRIVATE DWELLINGS GFCF, Current prices
The revisions to international trade reflect a range of improvements to the model used to estimate trade in travel services as announced in International Trade in Goods and Services, Australia (cat.no. 5368.0). In 2015-16 net exports was revised downward by $0.9b.
Graph 6. REVISIONS TO TRADE IN GOODS AND SERVICES, Current prices
REVISIONS TO COMPONENTS OF GVA
Similar to GDP, the revisions to gross value added were minimal throughout the time series.
Graph 7. GROSS VALUE ADDED, Volume measures
Graph 8 shows the average changes to the industry estimates of GVA between two decades: 1996-97 to 2005-06, and 2006-07 to 2015-16. The revisions to GVA between 1996-97 and 2005-06 were minimal for all industries. From 2006-07 onwards, larger revisions were observed, notably to Financial and Insurance Services, Professional, Scientific and Technical Services, Administrative and Support Services, Ownership of Dwellings and Construction.
Graph 8. AVERAGE GVA REVISIONS ($m), Current price
Finance and Insurance Services industry
The revisions to balance sheet and income and expenditure data reported to the Australian Prudential Regulation Authority (APRA) contributed to revisions to this industry from 2002-03 onwards. In addition, Financial Intermediation Services Indirectly Measured (FISIM) has been updated back to 1994-95 due to improved methodology, and revised source data which includes an improved RBA interest rate indicator. Additionally, COE levels were benchmarked to data provided by APRA and AIS from 2003-04 onwards.
Administration and Support Services Industry
This industry’s output, COE and gross operating surplus deviated from the benchmark data source (AIS), particularly in the most recent years. This deviation was corrected, resulting in upwards revisions to industry GVA, beginning in 2003-04 at $0.4b, and rising to $7.5b in 2014-15.
Professional, Scientific and Technical Services industry
GVA was revised upwards for Professional, Scientific and Technical Services (PST), from $0.5b in 2006-07 to $3.5b in 2014-15. Similar to the Administration and Support Services industry, this revision was largely attributable to the benchmarking of COE for this industry to reflect estimates from AIS. In addition, a more detailed survey of this industry was conducted for the 2015-16 year, which provided an improved understanding of cost structures at the finer industry level.
Ownership of Dwellings
Ownership of Dwellings was revised to include data from the 2016 Census of Population and Housing. Census data on the number of owner-occupied and rented dwellings and rents paid for rented dwellings were used to update benchmark estimates of HFCE of actual and imputed rent. Industry estimates of Output and GOS were updated to reflect these changes.
Moderate revisions to construction industry GVA mainly reflected the inclusion of the Private Sector Construction Industry, Australia, 2011-12 (CIS) (cat. no. 8772.0) and changes to private dwelling GFCF. In 2015-16, there was a downward revision to Construction industry GVA of $4.8b. The volume measures for the Construction industry were revised throughout the time series due to the introduction of a new producer price index in the deflator for heavy and civil engineering products.
REVISIONS TO KEY AGGREGATES
Table 6 presents a number of key current price measures. Gross domestic product was revised down by $2.3b in 2006-07 and revised up by $4.7b in 2015-16. Revisions to gross national income were of a similar order of magnitude. The revisions to GDP per capita contains the estimated resident population (ERP) rebased to the 2016 Census. Revisions to national saving and household saving are also shown.
Household net saving is the difference between household net disposable income and household final consumption expenditure. The household net saving ratio was revised downwards significantly from 2008-09 onwards, mostly reflecting the impact of the revisions to household final consumption expenditure.
Graph 9. IMPACT ON HOUSEHOLD NET SAVING RATIO, Current prices
REVISIONS TO NET WORTH
The revisions to national net worth reflected source data updates for land estimates and a correction back to 1998-99 of the real discount rate used in the net present value estimate for sub soil assets. In 2015-16 net worth was revised downward by $0.3b.
Graph 10. IMPACT ON NATIONAL NET WORTH, Current prices
REVISIONS TO MARKET SECTOR PRODUCTIVITY GROWTH CYCLES
Productivity growth cycle peaks are determined by comparing original MFP estimates with their corresponding long-term trend estimates. Graph 11 shows the 2007-08 peak previously identified has weakened due to the full historical revisions, including revisions to GVA, capital stock, and hours worked. This data point no longer passes the criteria for a peak. However, a robust 2011-12 peak has become available, its strength supported by the 2010-11 trough.
To ensure that peaks are more resilient to revisions, a three-filter approach (Henderson 11, H(11); Hodrick and Prescott, HP (1997) and Christiano and Fitzgerald, CF(2003)) was implemented in 2016 to identify cycle peaks. For more information about criteria to determine productivity growth cycle peaks, please see the feature article Experimental Estimates of Industry Value Added Growth Cycles.
Deviations of original MFP from long-term trend estimates were significant for 2011-12 by all three filters (H11), HP(1997) and CF(2003), indicating a robust cycle peak for this year.
Graph 11. DEVIATION FROM LONG TERM HENDERSON TREND MFP FOR MARKET SECTOR
There is a substantial amount of change in this edition of the ASNA. Employing new methodologies and the latest data sources ensures that the best available data are used in national accounts statistics. In most respects, the picture of the economy shown by the revised estimates is similar to previous published estimates. The impact on the movements in GDP in current price and volume terms were minor.
For questions concerning the nature of the revisions, contact the National Information and Referral Service on 1300 135 070 or National Accounts by email <firstname.lastname@example.org>
APPENDIX A - Data and Revision Triangles for selected GDP components
These documents will be presented in a new window.