4.5  Technology diffusion and adjustment

The literature reviewed so far points to a number of factors that influence the diffusion and application of new technologies in firms and industries over time. Acemoglu (2002) points to the endogenous nature of the supply of skills and new technologies. The firm dynamics literature identifies factors influencing the dispersion of productivity amongst firms, particularly plant level segregation of workers by high-low skill-technology matches.

The OECD has been undertaking a major project on the factors influencing economic growth, with earlier output including The New Economy: Beyond The Hype, (OECD 2001), and The Sources of Economic Growth in OECD Countries (OECD 2003b). More recent evidence on ICT adoption is provided in The Economic Impact of ICT (OECD 2004). These reports note that productivity improvements arising from ICT can occur through three broad channels. One channel is increased investment and productivity in the sectors that are developing or producing ICT technologies. Another is through increasing investment in ICT in other industries (for purposes such as accounting or inventory management), causing capital deepening and boosting labour productivity. A third channel is through spillovers into other areas of the economy (for example, through the internet or other networks, which enhance communications and allow such transactions to be managed differently).[12] The OECD (OECD 2004) identifies three broad groups of factors that are influencing the rate of ICT diffusion at the firm level. First, the regulatory environment affects the ability of firms to take advantage of opportunities. The OECD reports propose a broad policy framework that promotes product market competition, flexibility in the labour market, and firm entry and exit. Second, firms differ in their ability to absorb ICT. The firm’s absorptive capacity will stem in part from the firm’s skill base and level of innovative activity. The reorganisation of work practices is again identified as an important source of gains. Spillover effects around learning from the increased use of ICT also support adoption by remaining firms. Third, large differences persist in the costs of investment in ICT across OECD countries.

Basu et al (2003) explore the reasons why TFP might have accelerated in the US, but not in the UK. They argue that differences in investments in unmeasured intangible organisational capital in industries utilising ICT provides the best explanation of the divergent MFP performance after 1995. Benefiting from ICT appears to require substantial co-investments in complementary learning and organisational change, leading to lags of about 5 years between ICT capital investments and productivity growth in the ICT-using industries. This lagged effect appears to be stronger for the US than for the UK, where MFP growth does not appear to be correlated with lagged ICT growth. They suggest that another factor that may be influencing the UK ICT investments may be the need for skilled labour in the accumulation of complementary capital. They note that the UK has lower levels of skills than the US as measured by the percentage of the work force with degree or higher qualifications. Other factors, such as differences between the countries in terms of rigidities in product and labour markets, or business cycle effects, or in productivity measurement, do not appear to have a significant effect on US-UK differences in productivity growth.

Chapter 6 of the OECD’s report (OECD 2004) on the economic impacts of ICT analyses the Australian experience of productivity effects over the 1990s, using a firm-based longitudinal dataset. The aggregate effects of firm-level effects indicate that ICT usage contributed a relatively small 0.2 percentage points to Australia’s annual MFP growth rate of 1.8%. Their analysis of firm level data found ICT influences in all the industries studied, although ICT capital deepening was particularly strong in service industries especially finance and insurance, and was also above average in manufacturing, electricity and gas and water. It also found a significant positive relationship between increased ICT use and increasing sector wages/skill levels, the use of advanced business practices (eg, business planning, inter-firm comparisons), complementary organisational variables (measured by an intensity of firm restructuring variable), and product innovation. Causation was not determined. ICT uptake was earlier and stronger in large firms. They found that the firm level labour productivity effects tended to taper off over time. Thus, the adoption of ICT appears to provide a level increase rather than a permanent increase in the growth rate.

Related Australian papers (see Barnes and Kennard (2002) and Parham (2004)) suggest that the accumulation skills had only a limited role in Australia’s productivity surge during the 1990s, with growth in skills contributing only about 20% of the annual growth in MFP. Parham attaches greater weight to the increased openness of the Australian economy, increased R&D, and a rapid uptake and use of information and communication technologies (ICT). Nevertheless, these authors suggest that the accumulation of physical and human capital in the past has laid a long term foundation for the productivity growth now occurring. This suggests that the present level of skills may not be a significant constraint on growth. Thus, in relation to capital deepening and MFP growth, the contribution of skills needs to be looked at alongside other inputs, such as the innovation system and capital market effects.

The New Zealand evidence is less developed. Black, Guy and McLellan (2003) summarise labour productivity data for New Zealand industries for 1988-93 and 1993-2002, and make comparisons with Australian industries. These data reflect different rates of productivity growth amongst industries. MFP growth has, however, been uneven across nine broad industry sectors through the 1990s – being better in primary industries, retail and wholesale trade, and personal and community services, but poorer in utilities, manufacturing, mining, and transport and communications. In general, they found that New Zealand has poorer aggregate MFP and labour productivity than Australia, and also particularly in manufacturing. Industry-level investigations allow closer scrutiny of the factors influencing productivity. For instance, Engelbrecht and Xayavong (2004) find a strong correlation between ICT use and labour productivity growth across 29 New Zealand industries. They derive an ICT intensity index from input-output data, and classify industries into more ICT intensive and less ICT intensive categories. The evidence suggests that productivity growth is higher in the ICT intensive industries, though weak overall. Parham and Roberts (2004) explore Australia and New Zealand productivity differences in more detail. They suggest that once periods of rapid labour utilisation, and hard to measure industries, are excluded, that New Zealand’s labour productivity for the period 1996-2002 trails behind Australia by 2%. They contend that capital accumulation appears to be lower in New Zealand, particularly in the uptake of ICT, while the growth in skills in the working age population lags behind Australia.