RDP 9612: External Influences on Output: An Industry Analysis 2. A Summary View of Sectoral Output

Gordon de Brouwer and John Romalis

December 1996

A number of papers have pursued an explanation of the strong contemporaneous and long-run relationship between Australian and OECD/US output. Debelle and Preston (1995), for example, looked at aggregate consumption and investment links with other countries. They found that US and Japanese activity provide some information about domestic income and hence have an indirect effect on Australian consumption. But they also reported that overseas developments contain little information about domestic investor confidence, profitability or investment.

Gruen and Shuetrim (1994) tried to explain the correlation of Australian and foreign output in terms of the strength of foreign demand for Australian goods and services. They found, however, that export shares reveal little about the importance of foreign demand.[1] De Roos and Russell (1996) explained that this will usually be the case since an increase in foreign demand lifts Australian exports, but the associated pick up in domestic demand induces a supply shift away from foreign markets to the home market. The net effect of an increase in foreign demand on exports may be quite small – indeed, the correlation of exports and national income is quite low.

They, accordingly, model Australian exports as a function of both foreign and domestic demand, and find that the effect of foreign aggregate demand on Australian exports can be high, particularly when the foreign-income elasticity of demand for Australian exports is high, as happens to be the case for Japan and the United States. They also found that the stock market in the United States affects that in Australia, and that this may induce a common cycle, especially in investment.

This paper shifts focus to the production side of the national accounts. Since production is a result of both supply and demand effects, this should not necessarily be interpreted as an examination of the supply linkages between Australia and the rest of the world. There appears to be nothing published on the transmission of foreign shocks to Australian output at the industry level. Prasada Rao, Shepherd and Pilat (1995) used input-output tables to compare the levels of Australian and US manufacturing productivity and real output. They report that the structures of the manufacturing sectors in Australia and the United States are similar, although Australian manufacturing uses more intermediate inputs in production and so has a lower value added. They find that output, value added and prices have tended to be higher in protected manufacturing sectors in Australia. Productivity levels are about half those in the United States, and are highest in agriculture and mining and lowest in heavy engineering. Ergas and Wright (1994) reported that openness, as indicated by trade intensities and FDI, has increased in manufacturing. These studies focussed on manufacturing rather than general industrial output, and they did not explore the interactions of domestic with overseas production.

In this study, output interactions for most industrial classifications are examined. The reference country is the United States, since it is the largest economy in the world and has the highest productivity levels, and so may be viewed as the leading source of productivity shocks. Moreover, recent work indicates that US productivity shocks propagate quickly to other economies, while those from Japan or Europe do not (Elliott and Fatas 1996). The relationship between Australian and US sectoral output is examined at the one and two-digit levels of aggregation from 1977 to 1993. There are 11 one-digit sectors and 12 manufacturing sub-sectors, and these are graphed in Figures 1 and 2. The data are annual, since this is the frequency of the US data, and descriptions and reconciliations of the data are provided in Appendix A. The limited number of observations makes it difficult to draw strong inferences from the data. Table 1 presents summary statistics, including the share of each sector in total private sector output, the contemporaneous correlation of output growth in each Australian industry sector with the rest of Australian output growth, and the correlations of Australian and US sectoral output growth.

Figure 1: Real Sectoral Output
Index: 1977–1993 = 100
Figure 1: Real Sectoral Output
Figure 2: Real Manufacturing Output
Index: 1977–1993 = 100
Figure 2: Real Manufacturing Output
Table 1: Summary Statistics on Australian and US Sectoral Output
  Per cent of total private sector output
(period average) 		Sectoral &
total output correlation		 Correlation of Australian
and
US outputs
−1 (+1) indicates Australia (US) leads
   
Level   First difference
Australia US Australia −1 0 +1
Total GDP(P) 100   100   – −0.26 0.54** 0.65***
Agriculture 6   3   0.16 0.08 0.03 −0.40
Mining 5   3   0.33 0.23 0.29 0.26
Manufacturing 22   27   0.78*** −0.39 0.44* 0.46*
Food, beverages and tobacco   21   11 – 0.06 0.00 0.11
Textiles   2   2 – −0.13 −0.18 0.30
Clothing   4   3 – −0.17 0.55** 0.08
Wood and furniture   5   3 – −0.04 0.28 0.84***
Paper, printing and publishing   8   11 – −0.13 0.35 0.52**
Chemical, petroleum and coal   10   12 – −0.26 0.04 0.51**
Non-metallic mineral products   5   3 – −0.18 0.10 0.35
Basic metal products   11   5 – −0.01 0.51** 0.00
Fabricated metal products   7   7 – −0.44* 0.54** 0.63***
Transport equipment   11   12 – −0.45* 0.17 0.03
Other machinery and misc   15   26 – −0.23 0.51** 0.50**
mfg                
Utilities 4   4   0.55** −0.24 0.00 0.15
Construction 10   6   0.88*** −0.33 0.31 0.43*
Wholesale trade 11   8   0.82*** −0.32 0.19 0.02
Retail trade 13   13   0.62*** 0.19 0.19 0.02
Transport and storage 6   5   0.62*** −0.18 −0.02 0.60**
Rail   10   14 – −0.08 −0.03 0.50**
Water   11   6 – 0.17 0.22 −0.06
Air   18   20 – −0.29 0.12 0.39
Road   62   60 – 0.06 −0.23 0.30
Communications 2   4   0.38 −0.29 −0.13 0.24
Finance 15   25   0.61*** −0.22 0.25 0.53**
Recreation and personal services 6   3   0.64*** 0.25 0.28 0.00

Note: *, ** and *** denote significance at the 10%, 5% and 1% levels, respectively.

There are three striking features in the summary statistics:

  • while US production is about 20 times larger than Australia's, the industrial structures of the Australian and US economies are similar. For example, manufacturing accounts for about one-quarter of private sector output in both countries, and most other shares are broadly the same. The exceptions are that the US finance sector is relatively larger than Australia's, while the production of food and beverages is relatively more important in Australian manufacturing;
  • when there is a significant correlation between sectoral outputs, it is almost always the case that the correlation is contemporaneous or, more typically, that the development in the United States leads that in Australia. Australia follows the United States, rather than the other way round; and
  • the correlations appear to be concentrated in the manufacturing sector, and in areas which involve processing and technology. There is little correlation in services sector output, although construction and finance are the exceptions here. As a casual observation, both of these sectors would be thought to be the more open and traded of the services sectors.

Footnote

Bodman (1996), however, found that Australian GDP and exports are cointegrated and stable, but that the effect of exports on GDP is relatively small. He also reported that exports Granger-cause productivity, and that reverse causality is rejected. [1]