Department of Communications, Information Technology and the Arts
While it is generally acknowledged that ICT has played some part in Australia's recent good productivity growth performance, research done to date leaves much of Australia's recent productivity growth unexplained. This unexplained growth could be interpreted as resulting from micro-economic reform, suggesting that the Australian experience may be somewhat of an exception to the trend noted above in other countries.
The key findings from research
It is clear from Australian studies that microeconomic reform has played an important role across the economy in general. However, a study by the DCITA indicates that, for many parts of the manufacturing sector, new technology, including ICT, has made a much more significant direct contribution than was hitherto suspected.
This study
This study examines productivity growth in Australian manufacturing over the period from 1984-85 to 2000-01. Our main interest is to identify to what extent recent productivity growth was driven by technological progress, particularly the rapid technological advances in ICT, or by other factors.
During the 16 years covered in this study, labour productivity of manufacturing increased by 43.9 per cent. The study found wide disparities in the productivity growth rates of 19 industries, which cover the entire manufacturing sector. While industries such as electronics, medical and scientific instruments, non-ferrous metals, motor vehicles and pharmaceuticals recorded labour productivity growth in excess of 100 per cent over 16 years, other industries such as wood and paper, non-metallic mineral products, simple metal fabrications, furniture and miscellaneous small articles recorded productivity growth of less than 15 per cent. The general pattern of results suggests that high-technology and capital intensive industries tended to record higher productivity growth than the rest.
The strong correlation between productivity growth and R&D intensity highlights the importance of "technological paths" in determining productivity growth. Some industries, such as electronics and pharmaceuticals, are on rising technological paths, where the rapid pace of innovations opens the way for substantial cost reductions and/or new product developments. Usually, industries on rising technological paths are characterised by high business R&D intensity, because in these fields R&D tends to be more fruitful in developing new products or improving production processes.
This study could not adequately resolve the issue of how much of technological progress in manufacturing was due to the rapid advances in information and communications technologies (ICT) and how much can be attributed to other types of technological innovations. Indications from the literature suggest that a large proportion of technological advances in manufacturing over the last two decades were driven by the ICT revolution. Major innovations along this line include the incorporation of (computerised) numerical controllers into machines, robotics and local area (LAN) communication and control networks in factories. Also, the strong correlation found in this study between domestic ICT inputs and productivity growth suggests that ICT was a central factor driving productivity growth.
A further complementary study is proposed looking at the impact of ICT technologies in manufacturing at the firm level. This survey could shed more light on the relative importance of ICT versus non-ICT innovations in manufacturing.
The key methodological issues involved in this study
The estimation of productivity growth is beset by methodological and measurement difficulties. Further complications arise when analysing the data. A number of other studies have analysed Australian productivity growth at the national or sectoral level using growth accounting techniques. The present study addresses the issue from a different perspective, using straightforward regression and correlation techniques to decompose the drivers of productivity growth. There are a number of other possible approaches, although the scarcity of suitable data imposes severe constraints on what can be applied in practice.
The wide disparities in growth rates can provide an intuitive explanation for the finding from the regression analysis that technological factors were much more important than institutional-economic factors in driving manufacturing productivity growth. One would expect that changes in institutional-economic factors, such as improved labour market flexibility, improved competition in product and capital markets or rising education standards, would tend to affect productivity growth rates relatively evenly. In this case the wide dispersion of sectoral productivity growth rates would indicate that the main drivers of productivity growth were different technological paths and not changes in the economic environment.
Differences in productivity growth rates were investigated quantitatively through regression and correlation analysis. The explanatory variables used in the statistical analysis are allocated into three categories:
Technological variables including:
R&D intensity,
capital intensity,
labour productivity growth in other OECD countries (reflecting international technological paths),
inputs of locally produced electronic equipment and telecommunications services,
inputs of the above plus professional and engineering services.
Institutional variables including:
reduction of tariff protection,
reduction in the number of days lost per employee due to industrial disputes,
the share of university graduates in the industry's workforce,
the share of persons with post-school qualifications (from universities or technical colleges) in the workforce,
changes in the above mentioned education variables.
Changes in the capital-labour ratio (that is, capital deepening).
The regressions
Given that the regressions are based on 19 observations (for 19 industries), the explanatory variables were applied in a parsimonious manner, with no more than five variables in each regression, and no more than two variables from either the technological or institutional categories.
The results from the regressions were decomposed in two ways:
The portion of variations that were accounted for by variables from the three groups. The components of the aggregate productivity growth figure in terms of regression coefficients multiplied by the mean values of the independent variables.
The focus of attention in this study was on labour productivity growth (measured in terms of value-added per hour) because of the availability of suitable data on constant-price value-added and total hours worked per year in each industry. Some relevant data was also available on multi-factor productivity (MFP) growth and were used as dependent variables in some regressions. MFP growth is labour productivity growth minus the effect of capital accumulation on productivity. In other words, MFP represents the portion of productivity growth that cannot be explained by changes in labour and capital.
Most of the variations in productivity growth rates were accounted for by four or five variables. When technological variables were included in the regression, the relationships almost always turned out to be highly significant (R2 ratio above 0.7 and sometimes even above 0.8).
Increase in the capital-labour ratio (capital deepening) account in the regressions for around 20 per cent of the growth in labour productivity. Our interest was in decomposing the variations that were not due to capital deepening, thereby obtaining an estimate on the contribution of other factors to MFP growth. After discounting the impact of change in capital per worker, in regressions with high R2 ratios, technological variables usually account for over 80 per cent of explained variance and institutional variables account for less than 20 per cent.
In regard to the regression-estimated components of aggregate productivity growth, the results are slightly less clear cut. But in regressions where R2 exceeds 0.6 and the regression constant represents less than a third of aggregate growth, technological variables make up the bulk of the aggregate growth figure, while institutional variables account for a minor portion.
These decomposition estimates suggest that between 65 and 85 per cent of multifactor productivity (MFP) growth in manufacturing during the 1984-85 to 2000-01 period was due to technological factors, while institutional-economic changes may explain between 15 and 35 per cent of MFP growth. These estimates accord with the widely held view among economic historians that MFP growth in the long run reflects mainly technological progress.
It should be noted, however, that any decomposition between technological and institutional variables is difficult, because the two groups are inter-related. Indeed, most explanatory variables in the model are interdependent to some extent, but the size of impact is inferred from statistically estimable direct effects, without estimating possible indirect effects. Moreover, the estimates could be subject to inaccuracies due to insufficiently strong cross-sectional proxy variables to represent changes in the competitive environment of individual industries.
