RDP 2010-08: Sources of Chinese Demand for Resource Commodities 3. Industrial Uses of China's Resource Imports

3.1 Overview

Although the question of what drives resource demand is related to the question of what drives Chinese growth overall, it is helpful to see how the resource commodities that China imports are actually used across industries. This can shed light on the direct sources of demand for resources. With the aid of input-output tables we can also say something about the direct and indirect effects that the manufacturing export sector has on resource consumption, in terms of inter-industry linkages. This section of the paper explores these approaches, bearing in mind that they only provide a snapshot of inter-industry linkages, and cannot properly control for international factors driving the demand for resources.

Iron ore, aluminium ores, base metal ores and coal account for more than half of China's non-oil resource imports. Chinese consumption of imported iron ore and coking coal is driven by steel production in China. China's existing deposits of iron ore have low iron content by international standards,^[7] and the majority of iron ore reserves are located inland in the north and west of China, which makes transportation to the steel mills in the more industrialised coastal areas costly. In addition, in 2008, more than 90 per cent of China' crude steel was produced using blast furnaces and basic oxygen converters (the highest proportion in the world), which tend to rely on coking coal and iron ore, while comparatively little use was made of electric arc furnaces utilising a mix of steel scrap and other iron inputs. Thus the steel industry depends heavily on imported raw materials, with around half of China's total iron ore supply (adjusted for differences in iron content) sourced from overseas.

Data from various sources suggest that Chinese steel production in the 2000s is used in the construction (50–60 per cent), machinery (12–18 per cent), automobile (5–6 per cent) and home appliance (2 per cent) industries, with at least a quarter of domestic consumption being broadly ‘manufacturing’ (Table 2).

While these estimates suggest that the bulk of steel consumption is accounted for by construction, it is unclear how the various sources define end-use sectors and their accuracy is uncertain. Earlier sectoral estimates of steel consumption provided by Wu (1998, 2000) suggest substantially higher shares for manufacturing and lower shares for construction.

Even so, the mix of steel items produced in China underscores the importance of construction as an end use. China's steel industry has traditionally been weighted towards production of ‘long’ products and low-grade ‘flat’ products, both of which have important uses in residential and non-residential construction. But in addition to being used in construction, flat steel products – which account for a rising share of production – are used extensively in manufacturing, especially in appliances such as air conditioners and refrigerators, and in steel casing for vehicles. China's automotive manufacturing sector is now the largest in the world, and (in gross output terms) accounted for about 7 per cent of Chinese GDP in 2009; this share has almost doubled over the past 12 years.

China is dependent on imports for around one-third of its aluminium ore needs. Unlike iron ore, the bulk of aluminium ore demand is driven by the machinery, electronics and transport (particularly automobile) sectors, which together have accounted for around half of total consumption (Hunt 2004). Turning to other base metals, these have a range of industrial uses in both the construction and manufacturing sectors. The chief industrial uses of copper are for electrical/electronic products, engineering, construction and automobiles (Tse 2009). According to the World Bank (IBRD/World Bank 2009), 44 per cent of China's copper demand in 2007 was used in construction and infrastructure (compared to a global average of one-third). Zinc and lead have important uses in manufacturing – especially the automotive industry.

3.2 Input-output Tables

Although the assorted estimates shown above give a rough sense of the relative importance of manufacturing and construction as direct consumers of metal ores, they come from various sources and from different time periods. To avoid some of the problems of comparing data from inconsistent sources, we consider direct and indirect effects of manufacturing and construction activities on the consumption of resources using the official input-output tables for China. Overall, we find that manufacturing is a greater direct consumer of resources than construction. This is not surprising, as manufacturing accounts for about 40 per cent of China's GDP while construction accounts for only 6 per cent.^[8] At the same time, taking indirect effects into account shows that changes in manufacturing and construction may be equally important for resource demand.

To compute the direct effects, we calculate the share of the ‘output’ from resource-related industries used as intermediate inputs in the construction and manufacturing (excluding metal products) sectors. Metal products are excluded from ‘manufacturing’ to enable a better understanding of the relationship between resource imports and manufacturing production.^[9] We define ‘output’ as domestic supply – that is, an industry's gross output plus imports less exports, so as to reflect the value of resource-related output that is available for use in the economy. To gauge the importance of manufacturing exports as a driver of resource demand, we combine these estimates with export data. We use input-output tables published by the National Bureau of Statistics of China (NBS) from 1995 to 2005, with each table containing 17 industries, and update the analysis using the 42 sector table for 2007 (see Appendix B for details).

Figure 3 provides our estimate of the extent to which domestic ‘mining and quarrying’ (hereafter ‘mining’) inputs are directly used in manufacturing and construction, over the period 1995–2007. Mining includes coal, iron ore, other metal ores and other mining, while manufacturing (less metal products) is aggregated from individual industries.

Clearly, a larger share of the domestic supply of mining products is used directly in manufacturing than is used in construction, although much more manufacturing output is consumed domestically than is exported. The share of mining directly embodied in manufacturing exports is much lower, since a low share of mining output is used in manufacturing sectors with high export ratios (such as textiles, and machinery and transport). The construction sector and the manufacturing export sector each directly consumed around 5 per cent of the supply of mining products in 2007. However, the majority of mining output was used as intermediate inputs to ‘metal products’ and ‘coal and petroleum’ (not shown in Figure 3). For example, iron ore – a key mining product – is not used directly in construction, but is used to manufacture metal products such as steel, which are in turn used intensively by the construction sector (consistent with Table 2).

Indeed, a substantial proportion of metal products are themselves used as direct inputs to construction and manufacturing (Figure 4). The share of the domestic supply of metal products used in total manufacturing was almost three times higher than construction in 2007, though the construction sector consumed a higher proportion of metal products than manufacturing exports. The high share for total manufacturing is influenced by the use of non-ferrous metal products: the construction and manufacturing industries directly consumed roughly equal shares of the domestic steel supply in 2007. Our estimate of the share of metal products supply directly embodied in manufacturing exports rose strongly over the period 1997–2005, before easing in 2007 to around 10 per cent of metal products output. The share of the domestic supply of coal and petroleum used in manufacturing exports and construction is lower than the share of metal products inputs used in these sectors.

The above calculations only consider the direct shares of resource-related products used in other sectors. For example, it might be the case that a significant proportion of metal products are used as inputs to produce manufacturing equipment that is ultimately used by the construction sector. To get a rough idea of the role of indirect linkages in driving resource demand, we can compute the ‘Leontief inverse’ of the NBS input-output tables to gauge the effect of a one unit rise in construction and manufacturing final demand on different resource-related inputs.^[10]

Figure 5 presents the results of the calculation for metal products. It suggests that accounting for indirect linkages between industries makes a big difference (compare the left-hand panel of Figure 4). Over 1995–2007, a one unit increase in the final demand of the construction sector typically resulted in a larger increase in metal products output than a one unit increase in manufacturing final demand.

However, the effect of a rise in manufacturing demand on metal products output increased strongly over this period, and by 2005 was equal to the effect of an increase in construction demand. The effects of higher demand for manufacturing and construction had both increased further by 2007. These results suggest that while a unit of manufacturing requires a higher direct input of metal products than a unit of construction, construction requires a substantial input of manufactured items that themselves use metal product inputs intensively. Moreover, the total (direct plus indirect) effect of higher final demand in construction and manufacturing has risen since the mid 1990s, suggesting a rise in the intensity with which metal products were used over this period.

To summarise, our analysis of input-output tables indicates that the share of resource/metal product inputs directly used in manufacturing exports has risen, while the share of these inputs used in construction has trended downwards slightly over the 1995–2007 period. Accounting for indirect effects via inter-industry linkages, metal products are used more intensively in construction than their direct input-output shares would suggest. This exercise also indicates that the impact of a one unit increase in final demand, in either construction or manufacturing, on metal products production has been increasing over time. However, when we consider the much larger scale of the manufacturing sector, it seems clear that manufacturing has been at least as important, if not more important, than construction as a source of Chinese resource demand. In the next section we estimate a model of trade in resources in order to test the importance of manufacturing exports as a driver of China's resource imports.

Footnotes

According to the US Geological Survey (Jorgenson 2010), China reports that its iron ore has an iron content of about 33 per cent, compared with a little over 60 per cent in Australia, Brazil and India. [7]

This low share for construction could partly reflect measurement error, particularly with regard to the overlap of classifications. A feature of the Chinese national accounts is that data are often collected for each enterprise as a whole. For example, the OECD (Lequiller and Blades 2006) reports that large state-owned enterprises may extract mineral ores, produce metal products, construct their own buildings and generate their own electricity. Since the total output and value-added of each enterprise can only be assigned to one of its activities, the sectoral classification is chosen on the basis of which activity accounts for the greatest part of the enterprise's gross output. Given the ubiquity of construction in industry, it is likely that classification overlap is an issue for construction. [8]

Metal products manufacturing includes both finished and unfinished metal products (including raw steel), and is thus closely related to mining production and resource imports. [9]

For further details, please refer to Appendix B. We do not compute a separate effect for a rise in manufacturing export demand, since the NBS input-output tables do not distinguish between a change in final demand for manufactures that are exported and a change in final demand for manufactures that are consumed domestically. [10]