SECTION 7: METHOD AND DATA REQUIREMENTS


7.1 As noted earlier, transport activity is the movement of people or goods from one location to another. As a satellite account to the ASNA’s supply-use and input-output tables, a TrSA would primarily provide a systematic and consistent framework and dataset for analysing the role of transport in the economy, both on an industry and product basis. The contribution of transport activity is measured for both own-account and for-hire transport services. Own-account transport consists of the services provided by non-transport industries for their own use.

7.2 While for-hire transport output and value added would be the same as that shown in the annual supply and use tables from the core National Accounts, measuring own-account transport output and value added is less straight forward. There are a few options for measuring own-account transport services and these are discussed in Section 7.7 below.

7.3 A review of international work on TrSAs revealed that only a small number of countries have produced TrSAs (e.g. United States, France, Belgium) or related feasibility studies (e.g. Canada, Italy). The TrSAs can be categorised into either supply-side or demand-side approaches. For example, the United States approach measures transport activities from a supply-side approach. The headline measures include transportation’s contribution to gross value-added and Gross Domestic Product (GDP). That is, this measures the contribution from for-hire and own-account transport to the economy³. On the other hand, France⁴ and Belgium⁵ took the demand-side approach, where TrSAs provide data on total transport expenditures for different types and modes of transport. The Belgian TrSA also provides information on public transport infrastructure expenditure. The Canadian studies analysed the economic importance of transport by measuring the contribution to GDP from for-hire and own-account transport (supply-side approach); however this research was in the form of a research paper only and never officially published⁶.

7.4 The ABS is proposing to measure the TrSA from the supply-side to facilitate the estimation of transport’s impact on the economy, both directly and also in order to allow CGE modelling of downstream impacts and policy analysis. This is also consistent with the SNA recommended approach to satellite accounts, particularly to ensure comparison with the contribution of traditional “industries” to industry gross value-added.

7.5 As the United States approach is the only official TrSA estimate from a supply-side perspective, the ABS proposes to adopt a similar approach, with some refinements (see Section 8.1 below). The US input based approach has some advantage over a wage based approach) although in practice the ABS may use a combination of methods depending on data quality and availability at the time of compilation.

7.6 Some of the key relationships of the United States TrSA to the Input-Output ( I-O) accounts/Supply-Use (S-U) accounts, which the ABS approach would maintain, include:

• The methodology to value own-account transportation activities. That is, using a method similar to that used for own-account construction activities in the ASNA. For example, the intermediate inputs and the value-added inputs associated with or own-account construction, such as capital consumption allowance and labour costs, are moved—or, using I-O terminology, ‘‘redefined’’— to other industries in which the activities are primary. In the TrSAs, these inputs would be similarly redefined, but to a newly defined industry—“Own-account transportation”.
• The overall industry and product classification system and the special definitions and conventions in the I-O accounts would be used in the TrSAs except for own-account transportation, which would form a new industry and a new product category in the TrSAs.
• The value of GDP and total industry gross value-added (would be the same in the TrSA as in the core National Accounts. This is because value-added by own-account transport is redefined or moved from other industries’ value added.

Summary of proposed methodology for an Australian TrSA

7.7 The measurement of own-account transport output would be carried out by first estimating the inputs used by each industry for its own-account transport activities. There are several options for doing this:

1. Indirectly, by a "volume" based measure. This approach estimates the total transport related inputs used by the non-transportation industry (plus a share of non-transport related inputs such as accounting services) and allocates this by industry and mode using number of trucks/buses/trains/planes etc. This approach also uses ratios from the for-hire industry to allocate non-transport related inputs and value added inputs by industry and mode. The United States TrSA uses this approach.
2. Indirectly, by using wages and salaries by industry by transport mode from the Census/Labour Force Survey, focussing on occupations that vary by mode. From this information the ratio of total operating expenses to occupational wages could be calculated by transport mode. The wages are regarded as an indicator of operating expenses in the sense that if these wages and their ratio to operating expenses are known for the for-hire industry, then the operating expenses can be calculated for the non-transportation industry. That is, operating expenses for every mode in the production of own-account transport services can be calculated by multiplying the for-hire mode specific ratio by the industry specific wage estimate.
3. Direct collection/ratio estimation: this approach is similar to approach (a) in that own-account transport output could be measured as the sum of input costs plus depreciation and return to capital. The difference is that some of the own-account transport costs are directly collected by mode and industry as much as possible. In Australia’s case, this direct collection would be via utilising more detailed statistics on transport industry and activity in the years such data is collected in the Economic Activity Survey (which may be at best every five years). It should be noted, non-transport related inputs used in the provision of these services (e.g. accounting services) and Gross Operating Surplus could not be collected and would have to be derived.

7.8 The proposed Australian TrSA approach would adopt option 7.7.c and involve the following steps (see Appendix 1 for more detail):

Step 1: Selection of modal industries for deriving own-account transportation
Step 2: Selection of transportation related input (TRIs) expenditure items
Step 3: Estimating other inputs for own-account transport by non-transport industries
Step 4: Calculation of own-account transport output

Data requirements for a TrSA

7.9 As noted earlier, the ABS has already committed funds to researching the establishment of a TrSA (culminating in this paper), as well as the considerable costs associated with defining and collecting a range of additional transport data in the 2010-11 Economic Activity Survey . This data, combined with the other data sources mentioned below, would provide critical data for developing a TrSA. These data sources, and potential key data sets, are discussed further below.

Transport Related Inputs (Expenditure) data

7.10 The ABS proposes to measure own-account transportation output based on the sum of input costs plus depreciation and a return to capital. To measure own-account transport costs requires the collection of transport related operating expenses (inputs), consistent with the ASNA’s supply-use product classification (SUPC) level where possible.

7.11 The 2010-11 Economic Activity Survey will capture a range of additional financial and some non-financial data from both the transport industry and also in terms of transport activity undertaken in all other industries. The process for arriving at the inputs, or expense items to be collected in this survey is outlined below.

7.12 Transport related operating expenses (inputs) were defined for each mode with consideration to the following attributes:

• Uniqueness to transport activities and a specific mode of transport. If a particular product is not unique to transport (e.g. fuel use), then its proportional use for transport purposes needs to be identified.
• Typically, a non-durable product (used within one year) such as fuel, tyres, repairs, insurance etc.
• A relatively high proportion of total intermediate inputs (if possible).
• Ideally at the broader SUPC level of detail, given that it would be preferable to compile the TrSA using supply-use data given its timeliness compared to input-output estimates.
• Around 4-5 transport related inputs specific to each transport mode.

7.13 Many of the inputs chosen by mode on the Economic Activity Survey should be used 100 per cent for transportation purposes, or have been worded in such a way to identify only the transportation use of the input. For example fuel use is defined as “… fuel and gas used to power registered transport and motor vehicles”.

7.14 Due to provider load issues, the same level of detailed transport expense items (inputs) could not be collected from non-transport industries as transport industries, as the items might not be relevant to all businesses/industries. Consequently, broader level transport expenses by mode will be collected from units in all other industries in the Economic Activity Survey in 2010-11, with the more detailed mode inputs either derived from the “for-hire” splits, or derived based on a series of “case studies” or interviews with businesses reporting significant values for some of the broad expenses for a particular mode (e.g. water transport running expenses). See Table 3 in Appendix 1 for details of the transport related inputs the ABS is collecting.

Production of transport services (Income) data

7.15 Additional data on income from transportation services is also being collected in the 2010-11 Economic Activity Survey. As with the expense items identified above, the level of detail collected on income from transport services is more significant for those units belonging to the transport industry itself (Division I of ANZSIC), where sixteen types of transport services are being collected, e.g. rail passenger and freight transport services, postal services and support services to water transport. For units in all other industries, broader groupings of income from transport services are being collected, i.e. income from road freight transportation, income from other modes of freight transportation and income from other transport services.

7.16 Own-account transport activities will not enter the market and will not generate any income, so these activities are not able to be separately identified on the income side. As noted above, these activities will be based on the sum of input costs etc.

Additional data requirements

Transport Physical or Volume Data

7.17 There have been clear messages from the consultations to date that physical or volume data are of interest to policy makers. Satellite accounts present the option to use physical data not only to confront the monetary estimates, but also provide supplementary tables to the monetary accounts.

7.18 Research into the TrSA revealed no comprehensive data is presently available for the number of transport and motor vehicles by industry. Such data would be very useful to determine unit costs per vehicle (in association with the monetary estimates). Whilst the ABS will endeavour to collect Australian Business Numbers (ABNs) in the future to enable volume data by industry, this may be some time off. In the interim, the ABS will collect transport volume data from a sample of businesses in the Economic Activity Survey in 2010-11. Given the data represents a sample of all establishments, and it is the first time this data has been collected in over 25 years, the quality of this volume data is yet to be determined.

7.19 Ideally the ABS would release data on the number of transport vehicles and distance travelled classified by type of transport vehicle (e.g. trucks, buses, cars, trains etc.). Further research would be undertaken to determine what other sources may also be available for this type of data. For example, whilst not recording vehicle numbers, the reporting of fuel use by transport companies under National Greenhouse and Energy Reporting (NGER) has the potential to be a good source of data for fuel use.

Transportation Employment Data

7.20 Consultation with transport users indicated that transport employment aggregates and employee characteristics would be high priorities in a TrSA. Interest lies is in the level of employment involved in own-account transport activities compared with for-hire transport activities, hours worked as well as employment status ( full-time versus part-time), which would support labour productivity measurement if data was available over a long time series.

7.21 Compilation of employment relating to own-account transportation activity would present a range of conceptual and practical challenges. The ABS tested the capacity of businesses to report employment in own-account transportation services in the 2009-10 Economic Activity Survey, however businesses reported difficulties in identifying such employees due to a lack of data and classification issues. Consequently any attempt to measure employment in non-transport industries (for own-account transportation occupations) will need to be measured indirectly.

7.22 Should funding for a TrSA be secured, further research into employment methodology will be required. Options for further research might include estimating employment based on "for-hire" transport industry equivalent occupations from the Labour Force Survey. This household survey data is available by industry at various levels of detail, though the quality would need to be assessed. It should be noted that employer based surveys are generally regarded as producing better quality estimates of industry employment, compared to household interview based or self-enumerated industry estimates from household surveys. (Though there can be seasonal issues if the employer data is collected at a point in time only).

7.23 Census data on employment and occupation is another option. Whilst “industry of occupation” is self-enumerated in the Census (some matching to the ABS’s Business Register is undertaken based on industry of the employer), it has the advantage over the Labour Force Survey of not being sample based, and can therefore produce more accurate detailed estimates. The Census’ irregularity (5 yearly) means estimates would have to be extrapolated between Censuses by using movements in Labour Force data if TrSAs are to be constructed on an ongoing basis. However, the same issue is true of the Economic Activity Survey, as detailed transport estimates will not be collected for approximately another 5 years (and have the problems with classification mentioned above).

7.24 Alternatively, using a similar approach to measuring tourism related employment in the Australian Tourism Satellite Account, transport-related employment in the TrSA could be derived by applying the value added ratio (ratio of own-account transportation value-added to total value added for each industry) to numbers of employees in each industry.

7.25 Other options involve using a combination of wages data/labour force ratios. As mentioned above, this issue requires further investigation as to the best approach and what level of detail may be possible.

Issues requiring further investigation

7.26 Apart from measuring broader transport employment, there are a few other key issues which will require further investigation in order to compile a TrSA in Australia. Some of these issues are outlined in brief below. These will be addressed as part of any research and analysis work should the TrSA be funded in the future.

Industry Level GDP

7.27 The contribution of the wider transport industry to Gross Domestic Product (GDP) is a priority for some stakeholders as it enables total transport activity to be directly compared with the most widely recognised national accounting aggregate. This approach is also similar to other satellite accounts, e.g. the Tourism Satellite Account, where contribution to GDP is one of the key aggregates of interest to users.

7.28 GDP is calculated by adding net taxes on products to industry gross value added, and therefore calculating “industry level GDP” requires the identification of net taxes at the industry level. While there are difficulties in deriving net taxes at the industry rather than product level, the inclusion of a measure of “industry-level GDP” in satellite accounts recognises that there is a broad range of satellite account users who vary in their understanding of the key national accounting aggregates and are looking for something familiar, in particular for the advocacy uses of these accounts. The following are options for deriving transport net taxes on products as an intermediate step in deriving wider transport GDP:

• Primary Product: Net Taxes on Product would be allocated to the industry to which the product is primary (i.e. the industry producing a significant proportion of its output), irrespective of any secondary production.
• Output Proportions: Net Taxes on Product would be allocated according to industry in proportion to the output of the products upon which the net tax is incident.
• Value Added Proportions: Net Taxes on Product would be allocated according to industry in proportion to value added of the industry. As such, the industry shares of GDP are identical to the industry shares of value added.
• Value added generated from transport activity could be presented as a proportion of 'the contribution of GDP from production' (i.e. GDP less taxes on production). This approach does not require the problematic derivation of net taxes on products by industry. Specifically, the U.S.A. TrSA shows that the value-added by own-account and for-hire transport services was 4.4 per cent of GDP in 1997. This approach is followed in NZ (TSA), Canada (TrSA) and United States (TrSA).

Estimating profits on own-account transport activity

7.29 As the value of own-account transport output cannot be measured directly, the United States currently estimate the output of own-account transport by summing the costs of all intermediate inputs and value-added inputs of production including consumption of fixed capital (or depreciation). Although this approach is frequently used to measure the value of own-account types of production, it is recognised that the resulting estimates of output are understated as they do not include profits or return to capital. As a result, such estimates have limited value for productivity analysis and other similar studies.

7.30 Should the ABS proceed with developing a TrSA, identification and assessment of options to estimate return on capital for own-account transport activity will be another area of further research.

7.31 One option could be to utilise Australian National Accounts data on capital stock and depreciation estimates by industry. This data includes depreciation related to transport capital in non-transport industries. This option could entail deriving Net Operating Surplus on transport capital for each non-transport industry, by applying the ratio of for-hire transport's (ANZSIC Division I) Net Operating Surplus/Intermediate inputs to the sum of intermediate inputs required to produce own-account transport output by each non-transport industry to estimate Net Operating Surplus associated with own-account transport activity. Gross Operating Surplus could be derived by adding depreciation to Net Operating Surplus.

Treatment of transport margins

7.32 The System of National Accounts (SNA) 2008 (SNA08) and 1993 (SNA93) versions differ from SNA 1968 in their definition of what constitutes basic prices with regard to the treatment of freight transport services provided by a third party. Under SNA08/SNA93, only transport costs which are separately invoiced are treated as transport margin output and are excluded from the basic price of the product being transported. However, transport charges which are not separately invoiced are included in the basic price of the product being transported. This was a change from the SNA68 definition of basic price which excluded the transport component whether separately invoiced or not.

7.33 Analysts who use input-output tables however, express a strong preference for the SNA68 standard, as it provides more useful statistics for detailed analysis of the economy. ABS input-output tables use the SNA68 definition and this is being considered for implementation in the ASNA’s supply-use tables in the future. The approach used in the TrSA would most likely be determined by the ASNA’s supply-use treatment, since these are used in compiling quarterly and annual GDP and as these tables will be the main compilation source for the TrSA.

Treatment of light commercial vehicles (LCV’s)

7.34 Measuring the "transport output" of light commercial vehicles (LCVs) is another issue that requires further investigation as part of the TrSA development process. The measurement of LCVs is important due to their significance to road-transport activity. The significance of LCVs is not so much in terms of their total tonne-kilometres travelled (which was only 3.9% of tonne-kilometres travelled by freight carrying vehicles in 2010) but rather by their large fleet (82.6% of the freight-carrying vehicle fleet in 2010) and also the total kilometres travelled (72.8% of freight carrying vehicle kilometres in 2010). An issue complicating measuring own-account output of LCV services is the fact that despite contributing a high proportion of the kilometres travelled by freight vehicles, they consumed less fuel overall (5,546 million litres in 2010) compared to other freight carrying vehicles (6,403 million litres).

7.35 The key issue for valuing output from use of LCV vehicles on own-account relates to finding businesses which use corresponding vehicles in the for-hire transport industry. Some of the options include: matching to courier/postal transport inputs; identifying particular businesses with high proportion of LCVs to adjust the road transport related input proportions; or using Survey of Motor Vehicle Use characteristic information for LCVs.

Other issues

7.36 There are also some other issues which will need to be further investigated in developing a TrSA. For example, treatment of Government units and “output” associated with own-account infrastructure maintenance. In terms of the latter, if you consider how this activity is treated in Division I (for-hire transport), infrastructure maintenance activity e.g. production of road maintenance services, is recorded to the Construction Industry. This issue will therefore need further consideration.

---

³ US Transportation Satellite Accounts, http://www.bts.gov/publications/transportation_satellite_accounts/2011/
⁴ French Transportation Satellite Accounts, http://www.statistiques.equipement.gouv.fr/rubrique.php3?id_rubrique=253
⁵ Federal Planning Bureau, Headlines Belgian Economy, Analysis of public expenses and revenues for transport, http://www.plan.be/admin/uploaded/200812231923500.stu_0408.pdf
⁶ Statistics Canada - Cat. no. 11-001-XIE , The Daily, May 19, 2006, Study: Economic importance of transportation.