Spillovers from ICT and from ICT Research and Development
Print this pageThe two econometric studies by Diewert and Lawrence also included an examination of possible spillovers (externalities) related to ICT. Positive economic externalities provide the main rationale for government support for innovation.
As noted previously, Diewert and Lawrence show that for sectors with sufficiently robust data, ICT investment is ultimately worth more to producers than the market price, after adjusting for quality. In the follow-up study, they estimated that at the national level ICT investment is worth around 40 per cent more to users than the prices paid. They explain the undervaluation of ICT inputs by a combination of factors including market disequilibrium due to rapidly falling ICT prices, innovation related externalities and intangible investment in human capital and organisational change associated with ICT investment.
Externalities
In the economic literature, externalities reflect differences between the social value and the market price. There are both positive and negative externalities. The best known example for negative externalities is pollution, whereby the consumer or user who causes pollution does not pay for the environmental cost imposed on other members of society. Similar logic applies to congestion related negative externalities.
Knowledge related externalities represent the most important class of positive externalities. These externalities arise because the person or organisation creating new knowledge (innovation) usually cannot appropriate all the commercial benefits arising from it. Due to the leakage of knowledge to others, a large portion of the benefits flow to imitators or consumers through lower prices.
R&D externalities denote the difference between social and private returns to R&D. Social returns represent the contribution of R&D to national productivity growth. Private returns are the commercial benefits received by firms or organisations carrying out R&D. Knowledge related externalities can also occur in non-R&D related economic activities, such as the take-up of new innovations, ICT implementation work or equipment investment.
The common prescription to rectify market distortions caused by externalities is to impose taxes or subsidies that will equalise net marginal social value to the post tax/subsidy price, in other words, equalise marginal social and private benefits and costs.
The DCITA study, The economic impact of ICT R&D: a literature review and some Australian estimates, supports the findings of Diewert and Lawrence. Regression estimates based on Australian national accounting data for 12 market sectors between 1987–88 and 2002–03 indicated that the social rate of return from Australian business R&D is in the range between ten and 30 per cent. This is compared with average real private rate of return to R&D performing firms of around ten per cent. The estimated social rate of return to ICT R&D tends to be considerably higher than for non-ICT R&D.
Results from the labour productivity regressions indicate that the social rate of return to ICT investment is quite high, in the range of eight to 35 per cent per annum. This compares favourably with estimated average real private rate of return in the range of seven to ten per cent. The main reasons suggested for the difference between social and private returns to ICT are knowledge related externalities, rent spillovers due to falling ICT prices and intangible investment in human capital and organisational change that are complementary to ICT investment.
The high social rate of return to ICT would suggest strong reasons for encouraging ICT investment, particularly in ICT R&D. Moreover, as indicated in the box ‘externalities’, on grounds of positive externalities there may be a case to provide government subsidies for some ICT work that is not classified as R&D. This includes assistance for technology diffusion through subsidised ICT advisory services for small business.
Intangible investment in human capital
Part of the reason for the difference between social and private returns to ICT, inferred from national accounts data, appears to be connected with intangible investment in human capital and organisational change that are complementary to ICT investment.
In some econometric studies social returns exceed private returns to investment in machinery and equipment broadly defined, which also includes ICT equipment. The usual explanation is that the difference is related to intangible investment in human capital not recorded in the national accounts.
The productivity impact of public R&D is difficult to estimate and was not covered in DCITA’s regression analysis. However, according to the Digital factories report, industry–university cooperative and collaborative research centres and institutes in the manufacturing arena play a significant role in supporting ICT use in Australian manufacturing. The research outputs of many centres are ICT software, hardware, tools and products designed for adoption in manufacturing processes.
ABS data on public R&D expenditure dissected by fields of study reveals that in 2002–03 ICT R&D accounted for only 7.3 per cent of total expenditure on public R&D in universities and public research institutions. By contrast, the share of ICT R&D in total business R&D at that time was 36 per cent. If research closely related to ICT in other fields of study (particularly in production engineering) is added, then the share of ICT related R&D probably amounts to over half of business R&D.
Closely allied disciplines to ICT accounted for only 0.34 per cent of total expenditure on public R&D in 2002–03. These allied disciplines include nanotechnology, robotics, computer aided automation and control engineering. It is possible that in the statistics some work in these areas was included under other engineering fields of study, nonetheless, the fact that only a miniscule share of public R&D expenditure was devoted directly to these economically promising areas of research may be a cause for concern.
Notwithstanding its large share in business R&D, ICT R&D represents only a small fraction of the total value of ICT implementation work in the business sector. Less than 16 per cent of software development expenditure in the business sector is classified as business R&D. The share of software business R&D in total expenditure on software and information systems (including computer services) is around seven per cent. Most software and information system implementation work is not classified as R&D because it involves applying known software tools to solve routine problems, rather than to developing new software tools.
As indicated earlier, econometric results from DCITA studies (as well as some overseas research) suggest that there may be significant knowledge related externalities and rent spillovers associated with ICT implementation work.
Overseas studies on ICT spillovers
A number of overseas studies support the findings from DCITA’s research about significant spillovers associated with ICT, in a similar manner to spillovers from R&D.
Firm level econometric productivity studies (particularly by Brynjolfsson and Hitt in the United States) reveal that the returns to firms from ICT inputs tend to be much higher than from non-ICT inputs. This is particularly marked for ICT investment compared with other investment, but applies to a lesser extent also to ICT labour.
A number of macro-economic productivity growth studies revealed considerably higher social returns (in terms of contribution to GDP growth) than the market price for ICT purchases and investment. These studies are reviewed in DCITA’s report on ICT R&D.
- Document ID: 68028 |
- Last modified: 5 February 2008, 9:03am