RDP 2013-03: Implications for the Australian Economy of Strong Growth in Asia 3. The Resource Boom and Structural Change

Developments in the Australian economy since the onset of the current resource boom have been broadly consistent with the stylised framework outlined in Section 2. In response to the large increase in commodity prices and the terms of trade, investment and employment in the resource sector grew strongly, placing upward pressure on prices and wages. This was offset to some extent by an appreciation of the exchange rate, which led to a loss of competitiveness and downward pressure on prices in other industries exposed to foreign trade.

This section examines these developments in more detail, by considering the three (overlapping) phases of the resource boom:

• Phase I: increase in the terms of trade.
• Phase II: surge in resource investment.
• Phase III: mining production and export phase.

We look at how key variables in the resource, ‘other tradable’ and non-tradable sectors have evolved over recent years, while also noting some important ways in which developments in the Australian economy are not well represented by the theoretical framework presented in Section 2.

Our measure of the ‘resource sector’ is based on the measure of the ‘resource economy’ outlined in Rayner and Bishop (2013).^[8] In this measure, the resource sector is defined broadly to include both resource extraction and resource-related activity:

• Resource extraction. This includes mineral and gas extraction, and also resource-specific manufacturing (such as the production of metals and refined petroleum). This is very close to the ABS' definition of the mining industry, the only difference being that it also includes resource-specific manufacturing.
• Resource-related activity. This includes the provision of intermediate inputs that are used in the current extraction of resources as well as investment that supports future extraction of resources. In other words, it captures activities that are directly connected to resource extraction, such as constructing mines and associated infrastructure, and transporting inputs to, and taking extracted resources away from, mines. It also captures some activities less obviously connected to resource extraction, such as engineering and other professional services (legal and accounting work, for example).

The remainder of the economy – the non-resource sector – can usefully be divided into two parts:^[9]

• ‘Other tradable’ sector: comprises industries (or parts of industries) that are significantly exposed to international trade, but not directly related to the resource sector. This includes agriculture, manufacturing, transport, wholesale trade and accommodation & food services. For each of these industries, exports or competing imports are significant as a share of gross output.
• Non-tradable sector: comprises industries that typically do not have a significant exposure to international trade, and for which production is not directly linked to the resource sector.

The literature proposes several methods for creating this division (of the non-resource economy) in practice; see Dwyer and Groeger (1994) for a review. Our allocation is consistent with a ‘threshold’ approach, in which industries are allocated to the ‘other tradable’ sector if either their exports or competing imports are greater than a certain share of their gross output (generally 10 per cent; see Table A1).^[10]

The key prices and quantities of interest are shown in Table 1, which compares the inflation-targeting period up to 2002/03 with the period of the rapidly rising terms of trade thereafter.

3.1 Phase I: Rise in the Terms of Trade

Australia experienced a large increase in its terms of trade, rising by 82 per cent since 2003/04 to reach their highest level on record in September 2011; they have subsequently declined by around 17 per cent.

This run-up in the terms of trade has provided a significant boost to the real purchasing power of domestic production, given that a larger volume of imports can be purchased with a given volume of exports. For example, Stevens (2010) noted that, prior to the terms of trade boom, a ship load of iron ore was worth about 2,200 flat screen television sets. In 2010, it was worth about 22,000 flat screen television sets. This is akin to saying that Australians have received a transfer of income from the rest of the world. The increase in purchasing power flowing from a rise in the terms of trade can be estimated by comparing real GDP to real gross domestic income (GDI). Since the mid 2000s, growth in real GDI has exceeded that in real GDP by around 10 percentage points. However, Australians did not receive all of this transfer of income from the rest of the world, given that part of the resource sector is foreign-owned.

The distribution of these real income gains across the economy depends, crucially, on how much the exchange rate appreciates in response to the increase in world commodity prices (RBA 2005). An appreciation of the exchange rate means that: the increase in the domestic currency price of commodity exports will be less than the increase in world commodity prices; the income of the ‘other tradable’ sector will fall; and real income gains flow to the broader economy via the associated decline in the price of imports.

The exchange rate has played a particularly important role in smoothing the effects of terms of trade shocks in Australia during periods of rising commodity prices.^[11] Since the terms of trade started to rise in 2003/04, the nominal exchange rate has appreciated by around 25 per cent in trade-weighted terms. Thus, as described in Blundell-Wignall and Gregory (1990), most of the real exchange rate appreciation necessary to maintain internal balance following the terms of trade shock has been achieved by an appreciation of the nominal exchange rate. This has not been the case in much earlier terms of trade booms in Australia, during which the exchange rate was either fixed or heavily managed, when much of the adjustment to the real exchange rate was achieved via higher inflation.

However, the nature of the increase in commodity prices and the terms of trade, and the exchange rate appreciation, demonstrates a clear distinction between the theory presented in Section 2 and the reality of the current boom. In particular, the large rise in the terms of trade occurred over a number of years, rather than a one-off jump, and the extent of the rise was largely unexpected. This reflects the fact that the pace of development in emerging Asia consistently exceeded expectations (Figure 4). As a result, analysts (including official forecasters) under-predicted the extent of the increase in commodity prices, and therefore the terms of trade, and so the associated exchange rate appreciation also occurred over a number of years.

3.2 Phase II: Surge in Resource Investment

3.2.1 Investment

The resource sector globally has responded to the large rise in commodity prices by expanding its productive capacity. This has occurred gradually over a number of years, reflecting the unforeseen magnitude of the increase in commodity prices and the time required to plan for, approve and then complete large investment projects. In Australia, the growth in investment in the extraction of iron ore, coal and liquefied natural gas (LNG) has been exceptionally strong over recent years, accounting for most of the strong growth of resource sector investment (Figure 5). Indeed, the net capital stock of the resource sector is estimated to have risen by more than 150 per cent in real terms since 2003/04, and is expected to continue to grow rapidly over the next few years, particularly in the LNG sector. At the time of the Bank's February 2013 Statement on Monetary Policy (SMP), the central forecasts for growth embodied the expectation that investment in the resource sector would peak at a little over 8 per cent of GDP in 2012/13, compared to its average of 2 per cent of GDP over the past half century.

The effect of this surge in investment on GDP is lessened by the fact that a significant share of the investment is imported. In aggregate, it appears that around half of the value of these resource investment projects is imported, although this varies somewhat depending on the nature of the project. LNG projects, for example, typically involve higher shares of imported capital inputs. (The estimates of resource sector GVA presented in Section 3.2.2 of this paper have adjusted the value of mining investment for its estimated import content.)

In contrast, investment growth in the ‘other tradable’ and non-tradable sectors has slowed over recent years (Table 1 and Figure 5). Growth in ‘other tradable’ productive capacity has been particularly soft, which may reflect, in part, the high level of the exchange rate, whereas the slowing in growth in non-tradable capacity may reflect other forces acting on demand and confidence in this sector (see below).

3.2.2 Impact of the surge in resource investment on the output of other sectors

As the surge in resource investment gathered pace, the output of the broader resource sector, as measured by its gross value added (GVA), increased strongly (Figures 6 and 7). This is particularly notable for the resource-related construction and business services industries, which have supplied a large quantity of inputs required for resource extraction and investment.

To sustain the growth in the resource sector, factors of production were drawn from the ‘other tradable’ and non-tradable sectors, and GVA growth in those sectors slowed. In the case of the ‘other tradable’ sector, total GVA even declined in some years, with particular weakness in parts of manufacturing that do not supply many inputs to the resource sector, such as textiles, clothing & footwear and wood & paper manufacturing. Growth in GVA of the non-tradable sector has also slowed, but by less than the ‘other tradable’ sector. This suggests that the ‘income effects’ generated by the higher terms of trade (which tend to boost demand for non-tradable goods and services) outweighed the ‘substitution effects’ created by the fall in the price of tradable goods and services relative to the price of non-tradables (which tend to diminish demand for non-tradables relative to tradables).

It should also be noted that there have been factors other than the large increase in commodity prices and the high exchange rate that have had an effect on economic activity in Australia. For example, the global financial crisis caused significant disruption to financial markets and economic activity, albeit to a much lesser extent in Australia than in the north Atlantic economies. There was also an increase in the rate of household saving from the early 2000s, a slowing in credit growth and a transition to more stable levels of indebtedness and housing prices (relative to incomes). In addition to its direct impact on financial sector output, this has contributed to weakness in non-resource construction and turnover in the established housing market, as well as slower growth of retail spending than the strong growth experienced in the years prior to the global financial crisis (Lowe 2012b; Stevens 2012). Furthermore, there was a relatively broad-based slowing in Australia's productivity growth from the early 2000s; some but certainly not all of this can be explained by developments in the resource sector (D'Arcy and Gustafsson 2012).

3.2.3 Employment

Following the onset of the terms of trade boom, aggregate employment grew at an above-trend pace. The composition of employment growth also changed significantly (Figures 8 and 9). The share of total employment accounted for by the resource sector (including both resource extraction and resource-related activity) doubled since the mid 2000s, to be around 9¾ per cent in 2011/12.^[12] Around two-fifths of this growth reflected the expansion in resource investment, which increased demand for labour in resource-related construction and other industries that provide inputs to these investment projects (such as some types of machinery manufacturing and engineering services). The share of workers employed in resource extraction accounted for only about one-quarter of the overall increase in the resource sector's share of employment since the mid 2000s, while the remainder has been due to an increase in employment in industries that service the operations of mines (such as transport of output from mine sites to ports, business services and power generation).

Once the peak in resource investment has passed and the extraction of resources increases, the share of labour employed in the more labour-intensive resource-related sector is likely to decline and the share employed in the less labour-intensive resource extraction sector is likely to rise. The expected net effect though is a decline in labour demand in the resource sector.

Employment outcomes in the ‘other tradable’ and non-tradable sectors are consistent with labour moving to the resource sector in response to the higher relative wages on offer (see below), but could also reflect other factors. In contrast to the strong employment growth in the resource sector, employment growth has slowed in the non-tradable sector (particularly in retail and the parts of construction not exposed to the resource sector). The share of labour employed in the non-tradable sector has plateaued since around the mid 2000s, and is in line with the relatively stable share of non-tradable goods and services in production. However, as noted above, other developments largely unrelated to the high level of the terms of trade and the exchange rate are likely to have played a role in the stabilisation of the share of employment and output in the non-tradable sector. Most obvious is the weakness in the housing market; as a share of nominal GDP, dwelling investment peaked in 2003 at about 6½ per cent and has trended down to be around 4¾ per cent in 2012.

Employment growth has been even weaker in the ‘other tradable’ sector, and its share of employment has fallen since the mid 2000s (particularly in manufacturing). It is worth noting, however, that the recent decline in the ‘other tradable’ sector's share of total employment is a continuation of a longer-run structural shift since the 1960s. As real incomes have risen, consumer demand has shifted increasingly toward services, and this, combined with deregulation and privatisation of a range of services industries and the reduction in the level of trade protection provided to goods-producing industries, has seen the tradable sector's share of employment decline, and the non-tradable (service) sector's share increase.^[13]

3.2.4 Wages

The pace of aggregate wage growth picked up between 2003 and 2008 (Figure 10). This reflected considerable pressures on capacity in the economy prior to the global financial crisis, with the unemployment rate declining to its lowest level in more than three decades. When the slowdown associated with the global financial crisis occurred, these pressures on capacity eased and there was a significant moderation in wage growth. Aggregate wage growth subsequently picked up from these earlier low levels as activity recovered.

Wages have risen more rapidly in the mining industry than in the rest of the economy since the beginning of the terms of trade boom, with much of this adjustment occurring between 2004 and 2008 (Figure 11). As a result, the relative wage in mining increased by about 9 per cent over the eight years to 2012 (Figure 12). This was by far the largest increase of any single industry, after having trended lower over the decade leading up to the boom.^[14] It also appears that relative wages increased in sectors complementary to resource extraction, principally resource-related construction and mining services.^[15]

There was very little movement in the relative wage in the non-tradable sector overall and a decline in the ‘other tradable’ sector. This has been a key mechanism facilitating the reallocation of labour between sectors, whereby sectors benefiting from output price increases can afford to pay the higher wage rate and so draw labour away from other sectors.

While these adjustments in relative wages were substantial, the adjustment to the absolute level of nominal wages overall was smaller than in previous terms of trade booms, such as the early 1970s energy boom (Figure 13). This outcome in the current resource boom was helped by the combination of well-anchored inflation expectations and a more flexible labour market, particularly in comparison to earlier terms of trade booms. During these earlier booms, inflation was more variable and Australia's centralised wage-setting system had the effect of spreading wage increases across the economy to occupational categories for which the value of marginal product had not increased. Not surprisingly then, the result was a rise in inflation and unemployment (Gruen 2006; Battellino 2010; Banks 2011). While the adjustment of relative wages during the current boom has been substantial, the need for relative wages to adjust may have been lessened by a number of factors that have increased the supply of labour to the resource sector, such as: the adjustment in participation rates across different regions; the utilisation of skilled labour sourced from offshore by the resource sector; interstate migration; and employment practices such as fly-in fly-out and drive-in drive-out arrangements (see D'Arcy et al (2012) for details).

3.2.5 Consumer prices

Consumer price inflation has averaged around 2¾ per cent since the mid 2000s (Table 1). This is within the inflation target of 2–3 per cent over the cycle, but marginally higher than the average of 2½ per cent over the preceding decade. Even so, this can be considered as a relatively good outcome given the magnitude of the shock to the terms of trade, and also the much higher inflation outcomes associated with previous resource booms in Australia (Figure 14).

While inflation has been well contained, there were large shifts in relative consumer prices. Non-tradables inflation throughout the period of the terms of trade boom was stronger relative to its pre-boom average, as higher domestic cost pressures fed through to prices. Other factors less directly related to the resource investment boom also contributed to higher non-tradables inflation. For example, utilities price inflation picked up significantly from 2007, reflecting the move towards cost-based pricing, the replacement and expansion of infrastructure, and rising input costs (Plumb and Davis 2010). Productivity growth has also slowed since the early 2000s, although it is estimated to have picked up more recently.

At the same time, the higher exchange rate contributed to a noticeable decline in tradables inflation. Hence, the ratio of non-tradable to tradable consumer prices rose much more rapidly after 2003 compared to the trend of the previous two decades (Figure 15).^[16] This earlier underlying trend reflects the Balassa-Samuelson effect, whereby productivity tends to rise more rapidly in the tradable sector than the non-tradable sector. So, even though wages will tend to equalise across sectors over the longer run, unit labour costs rise more rapidly in the non-tradable sector.^[17] The fact that the ratio of non-tradable prices to tradable prices has risen faster than this earlier trend is consistent with the theory outlined in Section 2.

3.2.6 Interest rates

Interest rates rose over the first part of the terms of trade boom, from a bit below to a bit above their average over the post-1992/93 period. In part, this reflected the response of monetary policy to inflationary pressures building on the back of the boom in the terms of trade, as well as the increasing return to capital in Australia at that time. Thereafter, interest rates declined sharply in response to the global financial crisis. There has been considerably less adjustment of interest rates in the current episode, however, relative to earlier terms of trade booms. One reason is that the flexible exchange rate now provides a considerable buffer against external shocks, so that less adjustment of interest rates is required to manage domestic monetary conditions (Debelle and Plumb 2006).

3.3 Phase III: Mining Production and Exports

3.3.1 Mining production and exports

The response of mining production and exports to the increase in commodity prices followed with some delay, reflecting the time needed to plan, gain approval for, and reallocate scarce productive inputs to enable construction of new infrastructure. For some resource commodities, there has already been a significant pick-up in output and exports. Since the onset of the terms of trade boom, the volume of iron ore extracted and exported has risen at an annual rate of 11¼ per cent (Figure 16). LNG extraction has also risen strongly. Coal production has expanded, but at a broadly similar pace to its pre terms of trade boom average, in part reflecting a sluggish recovery in coal production from the floods in early 2011.

Given the significant expansions in capacity in the resource sector over recent years, and the lag between investment and the corresponding output, the production phase of the resource boom is expected to gather momentum over the next couple of years. In its projections published in March 2013, the Bureau of Resources and Energy Economics anticipated strong growth in iron ore and coal exports over the next half decade, of around 9¾ per cent per year (Figure 16). Growth in exports of LNG is expected to be even stronger, and this could see Australia emerge as the second largest global supplier of LNG in coming years (Jacobs 2011). With the terms of trade forecast to decline gradually over time, it is likely that the growth in the value of resource exports will be less than the growth in the volumes.

The strong growth in production and exports of these commodities over recent years has been offset, to a large extent, by weaker performance in other resource commodities. Most notably, oil production peaked in 2000, and since then has fallen by around 45 per cent as a result of the exhaustion of several of Australia's major oil basins. Exports of processed metals also remains below its levels in the early 2000s, and growth in non-ferrous metal ore exports (which include bauxite, copper ore, gold ore and nickel ore) has been sluggish relative to growth in iron ore, LNG and coal exports.

Reflecting these offsetting developments, the volume of Australia's total resource exports has risen at an annual rate of 3½–4 per cent over the course of the terms of trade boom (Table 1 and Figure 17). This is a notable slowing from its 1993–2003 average of 5½ per cent, notwithstanding a more than doubling of the capital stock and employment in the resource extraction sector. However, as noted above, the volume of Australia's resource exports is expected to increase at a faster pace in coming years as a result of the increase in investment.

3.3.2 Non-resource exports

The high level of the exchange rate and the impact of the global financial crisis on external demand have weighed on exports of non-resource goods and services. Exports of manufactured products from Australia remain well below their 2008 peak, even though the volume of global trade has surpassed its 2008 level (Figure 18). In particular, some of Australia's more traditional manufacturing export sectors, such as construction materials and road vehicles, are well down on their earlier levels. That said, there has been growth in some categories of manufactured exports, including specialised industrial machinery and professional and scientific instruments.

Exports of services have also declined significantly since 2008, although this also reflects the tightening of conditions for obtaining student visas, and more recently there has been some recovery in exports of tourism. In contrast to the softness in manufactured and services exports over recent years, rural exports have grown strongly, helped by improved rainfall and relatively high prices for many rural commodities.

The strong growth of economic activity in Asia is clearly evident in the shares of Australian exports by destination. In 2011/12, Asian economies (excluding Japan) accounted for half of Australia's rural exports, compared to around one-third in 2000, and over 30 per cent of manufactured exports (Figure 19).^[18] The share of Australia's services exports to non-Japan Asia has also increased, underpinned by strong growth in exports of education services. Also, short-term arrivals from China have been increasing rapidly – by around 13 per cent per annum over the past five years to be the second most important source of short-term arrivals (after New Zealand).

3.4 Beyond the Resource Investment Boom

The phase of strongly rising commodity prices has now passed. The economy is still in the midst of the boom in resource investment, although prospects are that it will peak sometime this year. So it is important to consider how the broader economy might evolve as the resource sector transitions from the investment phase to the production phase of the boom.

The stylised framework of Section 2 can help us think about the evolution of the economy as this transition occurs. An efficient, forward-looking foreign exchange market should already have factored in the expected decline in commodity prices associated with the gradual increase in global supply. To the extent that commodity prices and the terms of trade decline by more than anticipated, this would be expected to lead to a further depreciation of the exchange rate. Even if this were not the case, with the mining production phase less labour intensive than resource investment, this should free up some labour to move back into the ‘other tradable’ and non-tradable sectors, and wage pressures should moderate. This could lead to a reduction in domestically sourced cost pressures, although this will be offset to some extent by greater imported cost pressures arising from any exchange rate depreciation should that occur. This would also provide stimulus to the ‘other tradable’ sector by improving its international competitiveness.

In terms of how this translates to recent and expected developments in the Australian economy, we have already witnessed a decline in the terms of trade, which has decreased by around 17 per cent from the peak in September 2011. Current forecasts (as of the February 2013 SMP) suggest that the terms of trade will remain high in comparison to pre-boom levels. While it is impossible to be precise about the magnitude or the timing of near-term movements in commodity prices, this assumption seems reasonable on the grounds that industrialisation and urbanisation in China still has some way to run. For example, some estimates suggest that Chinese demand for steel used in residential construction will not peak until several years from now (Berkelmans and Wang 2012) – and the continued development of countries such as India, Indonesia and Vietnam will add significantly to global demand for commodities.

However, while the terms of trade have declined from their peak in late 2011, the Australian dollar remains at a high level by historical standards. One interpretation of this is that the decline in the terms of trade from their peak was anticipated and therefore already ‘priced in’ to the level of the exchange rate. But this ignores the unanticipated deterioration in the global economic outlook that underpinned the decline in commodity prices at that time. The high level of the Australian dollar might also reflect stronger economic growth in Australia relative to many other advanced economies over recent years, as well as increased demand for Australian dollar-denominated assets as other central banks attempt to stimulate their economies via large-scale expansion of their balance sheets (Lowe 2012a; Debelle 2013). How the economy adjusts in the years ahead will depend, in part, on how the exchange rate responds to economic developments; in particular, to the extent that the exchange rate does not depreciate in line with any unexpected declines in the terms of trade, this will affect the adjustment in other sectors of the economy, notably the ‘other tradable’ sector.

If the resource investment profile described above is realised, then overall demand for labour in the resource sector is expected to ease at some point. By itself, this would tend to lead to some moderation in inflationary pressures in the non-tradable sector. Conversely, any depreciation would exert some inflationary pressure on the prices of imported goods and services. Looking even further ahead, as Asian economies continue to develop and incomes rise, history suggests that demand will broaden beyond resources to other goods and services. Given the population of developing Asia, the Asian market is large and is expected to get much larger still; by 2020, more than half of the world's ‘middle class’ could be in Asia (Kharas 2010). From Australia's perspective, this could result in rising Asian demand for industries such as tourism, education, financial and other professional services, food and agribusiness, and specialised manufacturing, particularly related to agriculture and mining. As noted above, there is some evidence that this is already occurring.

Another important feature of emerging Asian economies that will have implications for Australia is the continued development and integration of Asian financial markets. While the economies of Asia are highly integrated into the global trading system for goods and services, this is less true of financial systems in general across Asia (Lowe 2009). Even though saving ratios have been higher than investment ratios in emerging Asia in recent years these ‘excess savings’ have predominantly been channelled to the rest of the world by the public sector as authorities, such as central banks and sovereign wealth funds, buy foreign assets. While investment in Australia from emerging Asian economies has grown in recent years, it still represents only a small share of total foreign investment in Australia. For example, Chinese investment in Australia has increased rapidly since the mid 2000s – driven by strong foreign direct investment, particularly in the Australian resource sector – but still represented only 1 per cent of total foreign investment in Australia at the end of 2011 (investment from the United States in Australia was 27 per cent of the total).

Footnotes

The methodology detailed in Rayner and Bishop builds on that in Kouparitsas (2011), which was implemented in Gruen (2011). [8]

Rayner and Bishop (2013) provide estimates of the gross value added (GVA) and employment in each industry that is directly connected to the resource sector. For this reason, in the ‘other tradable’ and non-tradable sectors used in the current paper, the parts of industries that are directly related to the resource sector are removed. This means that part of manufacturing GVA and employment is classified to the resource sector and part to the ‘other tradable’ sector, for example. The approach is different for sectoral estimates of investment and wages, which are based on the simple industry classification in Table A1. [9]

More precisely, an industry is classified to the ‘other tradable’ sector if more than 10 per cent of its total production was exported, or if competing imports accounted for more than 10 per cent of the industry's total supply, in 2008/09. The ‘other tradable’ sector in this article differs slightly from that in the previous version of this paper (Plumb et al 2012). In the current paper, the wholesale trade industry is also included in the ‘other tradable’ sector, reflecting updated information from the 2008/09 input-output tables. [10]

See Blundell-Wignall and Gregory (1990), Gruen and Wilkinson (1994) and Debelle and Plumb (2006) for a discussion of episodes following the floating of the Australian dollar in 1983. [11]

These estimates for employment assume that the productivity of a worker who works in a particular industry will be the same if they supply their labour to the resource or non-resource sectors of the economy (see Rayner and Bishop (2013)). [12]

This structural trend is particularly evident for manufacturing, whose share of employment fell from one-quarter in the 1960s to less than one-tenth in 2012. [13]

In this section, mining is defined as resource extraction excluding resource-specific manufacturing. [14]

Wages in construction and professional services increased strongly between the mid 2000s and 2012, relative to other industries. While ABS data on wages cannot be disaggregated into resource- and non-resource-related construction and business services, the RBA's liaison program suggests that the wage data by industry are likely to mask stronger growth in resource-related construction and services and weaker outcomes in construction and services not exposed to the resource sector. [15]

The picture is similar if utility prices are excluded from the calculation. [16]

See Balassa (1964) and Samuelson (1964). [17]

In this section, Asian economies include China, Hong Kong SAR, India, Indonesia, Korea, Malaysia, the Philippines, Singapore, Taiwan, Thailand and Vietnam. [18]