RDP 2013-02: Industry Dimensions of the Resource Boom: An Input-Output Analysis 3. Results

This section presents the key results. First, we document the extent to which the broader resource economy is larger than traditional estimates of the mining industry due to the inclusion of resource-related activity. Second, we examine the industry composition of resource-related activity. Third, we show that growth in the resource and non-resource parts of the economy has been very different over the course of the 2000s. And finally, we present a narrower definition of the resource economy that includes only those sub-industries of the resource extraction sector that have been most heavily exposed to external demand conditions and the associated boom in resource export prices and investment.

3.1 Size and Composition of the Resource Economy

We estimate that the resource economy accounted for around 18 per cent of nominal GVA in 2011/12, around double its share of the economy in 2003/04 (Figure 6). This reflects an increase in both resource extraction and resource-related activity.

Our estimate of the size of the resource economy is larger than the measure presented in RBA (2011a). The difference between the two measures (which is around 1¾ per cent of GDP in 2011/12) can be attributed to our inclusion of intermediate sales (as described in Section 2.6) and domestic final demand for resource output, although this is partly offset by subtracting resource imports and imported intermediate inputs used in resource extraction.^[21]

Resource extraction is estimated to have accounted for around two-thirds of the value of the resource economy in 2011/12 (11½ per cent of GVA; Figure 6 and Table 4). This includes the extraction of the resources themselves (the ‘mining’ industry as it is referred to by the ABS) – 9¾ per cent of GVA – and also the processing and refinement of those resources – 1¾ per cent of GVA. The large rise in resource extraction as a share of nominal GVA largely reflects higher export prices for resource commodities over the past decade; as a share of real GVA, resource extraction has been broadly unchanged, with strong growth in iron ore production having been offset by a sharp fall in oil production and falls in the production of other ores (which include bauxite, copper, gold, lead, nickel and zinc), as discussed in Section 2.5.

Table 4: Industry Composition of the Resource Economy

Estimated share of nominal GVA, per cent – 2011/12

Resource economy	18
of which:
Resource extraction	11½
Resource-related activity	6½
of which:
Business services	2¼
Construction	1¼
Manufacturing	1
Transport, postal & warehousing	¾
Wholesale trade	½
Electricity, gas, water & waste services	¼
Other	½
Sources: ABS; authors' calculations

The most important contribution of our methodology is to estimate resource-related activity and to decompose this activity by industry. As the resource boom has gathered pace, resource-related activity has picked up sharply, rising from an estimated 3 per cent of GVA in the early 2000s to around 6½ per cent in 2011/12. The largest contributions to resource-related activity in 2011/12 came from the business services, construction, manufacturing, transport, and wholesale trade industries (Table 4 and Figure 7).^[22] While construction and transport have obvious connections to the resource sector, business services (for example, engineering, legal and accounting services) account for a larger share of resource-related activity. As shown in Figure 7, business services are key inputs to both resource extraction and resource investment. In part, the relatively small share of construction reflects the fact that the construction industry itself draws on a relatively high share of intermediate inputs from other industries, and that a large share of resource-related construction investment is imported. However, consistent with the significant increase in resource investment since the mid 2000s, resource-related construction has increased sharply as a share of nominal GVA.

One key driver of the increase in the nominal share of the resource economy has been higher prices for resource exports (relative to other prices). To abstract from changes in relative prices, we construct a measure of real GVA of the resource economy by deflating nominal GVA by industry-specific implicit price deflators, and then aggregating across industries.^[23] Over recent years, the resource economy has, in real terms, grown much faster than the non-resource economy (Figure 8).

Since 2004/05, the resource economy has averaged 7½ per cent growth per year, while the non-resource economy has averaged 2¼ per cent growth. This difference largely reflects very strong growth of 16 per cent per year in resource-related activity, whereas to date, resource extraction has expanded only a little faster in real terms than the non-resource economy.

3.2 A Narrower Definition of the Resource Economy

Given that a strong rise in external demand, particularly from emerging Asia, has been the key driver of the recent resource boom, it is useful to consider a narrower definition of the resource economy that excludes the parts of the resource economy that have not directly benefited from this increase in external demand and the associated boom in resource export prices and investment.

In our narrower measure we exclude resource-specific manufacturing from our measure of final demand, as well as intermediate and final sales of any resources to the domestic economy. This leaves us with a measure of the resource economy that includes only resource exports (excluding resource-specific manufactured exports) and resource investment, less any intermediate and capital imports. We make the simplifying assumption that all resource investment is undertaken to increase future resource exports, not domestic sales, so all resource investment is included in our narrower measure. Under this narrower definition, the size of the resource economy accounted for 14¼ per cent of GVA in 2011/12, compared to 18 per cent for the broader definition. The average rate of growth for the narrower definition has been faster than the broader measure, reflecting the divergence in growth patterns between external and domestic demand (Figure 9).

Footnotes

Our estimate of the size of the resource economy is within the range of estimates published by the Australian Treasury (15–20 per cent over the forecast horizon) in the May 2012 Budget (Australian Government 2012) and in Gruen (2011). Our estimate is a little smaller than in Shann (2012); Shann estimates that the mining and mining-related sectors accounted for 19½ per cent of GDP in 2010/11, whereas our measure suggests these sectors accounted for 17 per cent of GDP in 2010/11. Shann makes adjustments to include mining services investment and exports in the scope of the ‘mining-related sector’. We do not make this adjustment for investment, as we have already captured most of this activity in our measure of resource investment, and hence would be double-counting. Likewise, we do not make Shann's adjustment for exports of mining services as these exports are not separately identified in ABS data and estimates of their size vary widely. [21]

A finer disaggregation of the resource economy by industry can be found in Appendix C. [22]

Implicit price deflators (IPDs) for GVA are published annually in the Australian System of National Accounts (ABS Cat No 5204.0). The ABS publishes these IPDs at the broadest industry level (i.e. the industry ‘division’). In instances where we have used a more disaggregated industry definition, such as with the manufacturing industry, we deflate all sub-divisions by the aggregate IPD for that industry. To account for compositional change in output over time, we sum the chain volumes using an appropriate chain-volume aggregation method. The ‘implied’ price deflator for the resource economy using this approach is not strongly correlated with the relevant final price deflator from the expenditure side of the national accounts (i.e. the weighted average of the IPDs for resource exports and investment). The reason for this difference is difficult to identify, and could reflect measurement errors in our approach or in the published expenditure IPDs, or both. [23]