RDP 2010-08: Sources of Chinese Demand for Resource Commodities 1. Introduction

China's demand for resources is large. China accounts for around two-thirds of world iron ore demand, around one-third of aluminium ore demand and more than 45 per cent of global demand for coal.^[1] Accordingly, Chinese demand for resources is of considerable importance to resource-exporting countries such as Australia, Brazil, Canada and India. Just as there is debate over the sources of Chinese growth – investment versus exports – there is a similar question as to the relative importance of investment and exports as drivers of resource demand.

Understanding the sources of Chinese demand for resources is important. On the one hand, if the demand for resources is driven by infrastructure and housing investment, then China's resource imports depend on the rate of urbanisation and the government's continued focus on improving public facilities.^[2] On the other hand, if China's manufacturing export industry is an important driver of resource demand, then China's demand for resource commodities may be more vulnerable to external developments.

This paper attempts to shed some light on this issue in two main ways. First, we use input-output tables to trace through the direct and indirect effects that the construction and manufacturing sectors have on resource demand. We demonstrate that, over the past decade or so, manufacturing has been at least as important as construction as a driver of China's consumption of resources. Second, we show that global trade in non-oil resource commodities can be described by the gravity model of trade. Using this model we find that, controlling for domestic expenditure (including investment), total exports are, in general, a sizeable and significant determinant of a country's resource imports, and that this has also been true for China.

The approach taken by this paper differs from previous work on the drivers of China's demand for resources. A few papers consider the question for individual commodities using time-series econometric techniques: for example, demand for energy commodities such as oil and coal (Crompton and Wu 2005; Zhao and Wu 2007), or iron ore (Tcha and Wright 1999). By and large, these papers emphasise industrial production or GDP growth as the main determinant of resource demand; they do not focus on the extent to which exports or investment are important. Focusing on steel demand in particular, McKay, Sheng and Song (2010) report that a country's urbanisation rate, investment and automobile use are statistically significant determinants of its GDP per capita – which in turn displays a positive quadratic relationship with steel use (the ‘Kuznets curve for steel’) – while trade openness is insignificant. In contrast, Kahrl and Roland-Holst (2008) use input-output tables to demonstrate a clear link between China's export production and its demand for oil and coal. Garnaut and Song (2006, 2007) argue that three variables – the investment share of output, the export share of production, and the level of urbanisation – determine China's demand for metals and energy. This view is supported by Zhang and Zheng (2008) who argue that manufacturing production has a significant influence on resource demand. We do not dispute this general assessment, but instead use a gravity model of trade to demonstrate that China's (predominantly manufacturing) exports have been a significant determinant of its non-oil resource imports.

With the exception of Eita and Jordaan (2007), we are unaware of any previous efforts to model trade in resource commodities using this approach.^[3] However, many recent studies have used the gravity framework to model China's role in international merchandise trade. Edmonds, La Croix and Li (2008) and Bussière and Schnatz (2009) find that China's aggregate bilateral trade can be explained by the gravity model. Eichengreen, Rhee and Tong (2007), Greenaway, Mahabir and Milner (2008) and Athukorala (2009) use the gravity model to examine whether China's growth has had an effect on Asian countries' exports to third countries.

Eichengreen et al (2007) find that China's exports tend to displace the exports of less-developed Asian economies (but not those of high-income Asian economies). In contrast, Greenaway et al (2008) find that China's exports ‘crowd out’ the exports of most Asian economies (particularly the more developed economies), while Athukorala (2009) finds little evidence of crowding out.

Work that is more closely related to our paper is Sheng and Song (2008), who model bilateral trade between Australia and China at the industry level using a gravity model. They find that China's trade with Australia is negatively correlated with tariff rates and positively correlated with measures of each country's ‘revealed comparative advantage’. They also report that China's revealed comparative advantage in downstream industries (such as manufacturing) tends to be positively correlated with China-Australia trade in upstream industries (such as resources). This supports the idea that, in recent decades, China's imports of resources have been heavily influenced by its manufacturing export sector.

This paper proceeds as follows. Section 2 discusses the debate on the key contributors to GDP growth in China, and presents some facts about Chinese investment and exports. Section 3 highlights construction and manufacturing as the primary end-use sectors for resources and metal products, and uses input-output tables to analyse these linkages. Section 4 uses a gravity model of resource trade to test the idea that manufacturing exports have influenced China's imports of resources. Section 5 concludes.

Footnotes

The sources for these shares are the US Geological Survey (USGS 2010b) and the United Nations COMTRADE database; USGS (2010a) and CEIC; and BP (2010), respectively. According to the US Energy Information Department, China is also the second-largest consumer of oil after the US, but we do not consider China's demand for oil in this paper. [1]

The view that construction investment drives steel production and hence China's demand for iron ore appears to be common among commodity analysts. See, for example, Komesaroff (2008). [2]

Eita and Jordaan analyse South African metal exports to 33 trading partners using a gravity model. [3]