RDP 2012-04: Chinese Urban Residential Construction to 2040 4. Projections

Turning to the future, our methods suggest that the level of urban residential construction is projected to peak in 2017 at a level almost 12 per cent higher than 2011 levels (Figure 9).[16] This marks a moderation of the strong growth we have seen over the past few years. Nonetheless, the outlook is positive. Chinese urban residential construction is expected to stabilise at a high level, and to remain above current levels for almost 20 years. Further out, urban residential construction is projected to taper off slowly.

Figure 9: Urban Residential Construction

An alternative projection is offered in Woetzel et al (2009), where much stronger growth is forecast. We project lower growth rates partly due to our higher starting point, arising from our conjectures in Section 2. Nonetheless, our projected volumes remain above those in Woetzel et al because we project faster growth in floor space per capita. Hu et al (2010) also provide projections; our projections are higher than theirs, which appears to be due in part to their assumption of lower rates of demolition.

We now turn to consider the implications of our projections of urban residential construction for steel demand and the structure of the Chinese economy.

4.1 Steel

Steel intensity is defined as the weight of steel used per square metre of floor space constructed. In this paper, we do not account for steel found in internal furnishing and appliances, such as refrigerators. Average steel intensity in construction depends upon a range of factors. For example, while most modern apartments are constructed with reinforced concrete, steel intensity can vary based on the building's design and legislation requires a higher steel content for buildings in earthquake-prone regions. A key consideration is that steel input varies with the height of the building. Estimates by Walsh (2011) suggest low-rise apartment construction requires, on average, around 40–50 kg of steel per square metre of floor space, whereas around 60–100 kg of steel per square metre is required for a 30-storey building. Basements also increase a building's average steel intensity substantially.

We begin by assuming that the average steel intensity for newly constructed apartments in China was 60 kg per square metre in 2010. This is in line with Woetzel et al (2009, p 373) and Walsh (2011). Given that approximately 1.54 billion square metres of residential floor space was constructed with reinforced concrete in 2010, this implies steel use of about 93 million tonnes – about 14 per cent of China's 2010 crude steel production (Figure 10). Some estimates suggest that total building construction – i.e. residential and non-residential – accounts for about 30 per cent of steel use (KPMG China 2011; Wu 2009). This appears to be consistent with Chinese residential building accounting for about 60 per cent of construction as commercial building should be more steel intensive than residential construction.

Figure 10: Projection of Steel Use from Urban Residential Construction

We also assume that average steel intensity for newly constructed apartment buildings will increase linearly by 1 kg per square metre per year to reflect improved quality and higher buildings.[17] Moreover, the projected increase in incomes is likely to see higher car ownership and demand for underground car parking, which suggests that this is a conservative assumption. If this rate of increase is assumed over the course of the past few years, 19 per cent of steel production for 2004 would have been used in residential construction, which is identical to the share estimated by Hu et al (2010).

Under these assumptions, and our earlier projections for residential construction, steel use in residential construction is likely to continue to grow rapidly for a few more years. Beyond the middle of this decade, growth is projected to moderate, with steel use reaching its peak in 2023, at a level that is 30 per cent higher than consumption in 2011.

4.2 Implications for the Structure of the Economy

Since 2004 residential construction has increased rapidly as a share of GDP (Figure 11).[18] We estimate the current share to be around 9 per cent of GDP, which is high by international standards. Australia's share is 5 per cent and, at the peak of the building boom around 2005, the share in the United States reached over 6 per cent.[19]

Figure 11: Share of Residential Construction in GDP

From 2004 to 2011, China's real GDP more than doubled. Annual floor space completed, which includes both rural and urban buildings, increased by only 41 per cent. Our measure of urban construction increased by 57 per cent. If relative prices remained constant, then these growth rates imply that dwelling investment would have accounted for a declining share of nominal GDP. However, relative price changes, perhaps indicative of quality improvements or changes in the price of materials, are likely to have been important.[20] Steel intensity is arguably a good proxy for quality improvements, in which case overall growth in steel use reflects growth in real construction. The estimated steel use in residential construction grew by over 70 per cent from 2004 to 2011. While this increase is not enough to account for the rising share of GDP if all other relative prices remained constant, it is an improvement upon the raw floor space numbers. On the basis that other factors that may have previously affected the share – for example, increasing materials prices – are not sustained, but that quality improvements continue, the share of residential construction will begin to fall over the projection horizon, but not rapidly (Figure 11). We project residential construction's share to decline to 6 per cent in 2020, and by 2040 it will be down to around 2 per cent of GDP.


Even though the line in Figure 9 is solid up to 2011, the 2011 figure is, in effect, a projection because the data are not yet available to allow us to derive our estimated measure. [16]

This is consistent with the rate of increase in Hu et al (2010). [17]

To calculate this share, we first deduct land investment (CEIC DX Identifier: CRKARPN) from real estate investment (CEIC DX Identifier: CECA), and then multiply this value by the share of residential building (CEIC DX Identifier: CECAA) in real estate investment, to get an approximate series for residential building investment less land. Because fixed asset investment (FAI) data better represent what is calculated in the national accounts, we multiply our approximate series by the ratio of residential building FAI (CEIC DX Identifier: COBDJW) to residential building investment. We then divide the resultant series by nominal GDP to get the share in Figure 11. [18]

Australian figures are for total dwelling investment, which include alterations and additions: Australian Bureau of Statistics Cat No 5204.0, Table 2. Figures for the United States are from the Bureau of Economic Activity. [19]

Another potential explanation is that there has been a bulge of construction that is yet to be completed, which would affect GDP calculations but would not yet show up in completion numbers. However, this seems unlikely to be the full explanation given that it typically takes one to two years for a large building to be built in the United States, and possibly less time in China (US Census Bureau 2012). [20]