Working Paper No. 233 April 2008 www.esri.ie ICT Diffusion, Innovation Systems, Globalisation and Regional Economic Dynamics: Theory and Empirical Evidence a b b Charlie Karlsson , Gunther Maier , Michaela Trippl , c d c Iulia Siedschlag , Robert Owen , and Gavin Murphy Subsequently published as: Karlsson, C., Maier, G., Trippl, M., Siedschlag, I., Owen, R., Murphy, G., “ICT and Regional Economic Dynamics: A Literature Review”, European Commission, Joint Research Centre, Institute for Prospective Technological Studies, Luxembourg: Publications Office of the European Union, October 2010. Abstract: The objective of this paper is to review the relevant theoretical and empirical literature to provide a conceptual and methodological background for the analysis of the consequences of ICT use and globalisation on the regional economies in the European Union. We highlight the key aspects of ICT as a general purpose technology, discuss the economic impacts of ICT diffusion from a macro as well as from a micro perspective, and examine the spatial consequences of ICT diffusion. We focus on regional innovation systems and globalisation in order to propose an organizing framework for the analysis of the impact of ICT diffusion on regional development. Key words: Technology diffusion, Innovation systems, Globalisation, Regional development JEL classification: F23, O33, R11 Corresponding Author: CONTENTS 1 INTRODUCTION 1 2 ICT: DEFINITIONS AND MEASURES 4 3 THE ECONOMIC IMPACT OF ICT DIFFUSION 8 3.1 ICT and Structural Change 10 3.2 The Effects of ICT on Aggregate Productivity and Output Growth 12 3.3 The Effects of ICT on Productivity and Output Growth at Firm Level 18 4 REGIONAL DYNAMICS AND ECONOMIC TRANSFORMATIONS DURING THE DEPLOYMENT PHASE OF ICT 21 4.1 ICT and Spatial Transformations 24 4.2 ICT and the Location of Firms 28 4.3 ICT and Regional Economic Growth 30 5 REGIONAL INNOVATION SYSTEMS AND THE LAYERS OF INNOVATION 38 5.1 Regional Innovation Systems: Key Characteristics and Dimensions 38 5.2 RIS and Linkages with National and International Innovation Systems 42 6 GLOBALISATION AND ICT IN THE KNOWLEDGE-BASED ECONOMY 55 6.1 ICT-enabled Outsourcing, Offshoring and Firms Performance 57 6.2 The Impact of Globalisation and ICT on Regional Performance 60 6.3 Policy Stakes of Globalisation, ICT and Regional Development 62 7 SUMMARY AND CONCLUSIONS 64 7.1 Summary 64 7.2 ICT, Innovation Systems and Globalisation: An Integrated View 72 References 75 1 1. Introduction Information and Communication Technologies (ICT) is no doubt one of the key innovations of the last century. ICT represent a new technological paradigm that belongs to the family of General Purpose Technologies (GPTs). A GPT has the potential for pervasive adoption and adaptation in a wide range or even all sectors of the economy in ways that drastically change operations and products as well as the relationships between different sectors. The characteristics of GPTs have been described by Bresnahan & Trajtenberg (1995, p.84): “Most GPTs play the role of ‘enabling technologies’, opening up new opportunities rather than offering complete, final solutions.” GPTs also involve ‘innovational complementarities’, i.e. “the productivity of R&D in a downstream sector increases as a consequence of innovation in the GPT technology”. Thus, GPTs have two major characteristics: generality of application; and, innovational complementarities. However, other characteristics of GPTs are also important (Lipsey, Becar & Carlaw, 1998): (i) much scope for improvement initially, (ii) many varied uses, (iii) applicability across large parts of the economy, and (iv) strong complementarities with other technologies. Some innovations are incremental and some are drastic. ICT are an example of a drastic innovation, which qualifies as a general purpose technology (GPT), since it has shown to have the potential for (i) pervasive use in a wide range of sectors in ways that drastically change their modes of operation as well as the character of their output, (ii) setting the stage for series of incremental innovations, and (iii) producing discontinuities in the observed pattern of resource allocation and the evolution of output. ICT being a GPT has many implications: i) adoption of such a technology entails experimentation, which may lead to innovation by adopting firms, which in turns show up as total factor productivity growth, ii) as well as innovating themselves, firms can learn from the (successful or unsuccessful) innovation efforts of others, so there are spillover effects (Bresnahan & Trajtenberg, 1995), and iii) successful implementation of an ICT project requires reorganisation of the firm around the new technology (Helpman & Trajtenberg, 1998, Yang & Brynjolffson, 2001; Brynjolfsson, Hitt & Yang, 2002). ICT are composed of a wide range of product and service technologies including computer hardware, software and services and a host of telecommunications functions 2 that include wire or wireline, and wireless, satellite products and services. The rapid diffusion of ICT has produced important changes in how and where goods and services are produced, the nature of goods and services produced, and the means by which goods and services are brought to the market and distributed to consumers. This implies that ICT has had an impact on the industrial structure of regions and on the geographical location of different industries not only within the EU but worldwide. ICT has also influenced the relationship between customers and suppliers and the way many markets for intermediate as well as final goods and services are organised. However, there are substantial differences among countries and regions, also among the developed economies, as regards their role in the development of ICT and their ability and propensity to adopt ICT applications in various activities and sectors (Johansson, Karlsson & Stough, 2006). This implies, among other things, that there is a substantial variation in the impact of the use of ICT on efficiency, productivity, and economic growth in different countries and regions. The objective of this paper is to review the relevant theoretical and empirical literature to provide a theoretical and methodological background for the analysis of the consequences of ICT use and globalisation on the regional economies in the European Union (EU). This review focuses on the following core issues and analytical questions: ICT: DEFINITIONS AND MEASURES 1. What do we understand with ICT as a set of technologies and as a sector of the economy? 2. How relevant is the existing statistical data for analysing different aspects of the impact and effects of ICT? THE ECONOMIC IMPACT OF ICT DIFFUSION AND REGIONAL DYNAMICS 3. What do we know about the effects of ICT diffusion on structural change, productivity and output growth? 4. What are the effects of region-specific mechanisms, such as technological and knowledge spillovers, interactions and networking between firms, on ICT diffusion and its impact on the economic efficiency across EU regions? 5. How does ICT influence spatial patterns of economic activity and what type of regional transformations can be linked to ICT? 3 REGIONAL INNOVATION SYSTEMS AND ICT DIFFUSION 6. What key features characterise regional innovation systems in relation to ICT production and use? 7. What are the linkages between regional innovation systems and larger institutional frameworks at national and international levels? 8. How do regional innovation systems influence the effective use of ICT at regional level? 9. How does ICT influence the functioning of regional innovation systems? CONSEQUENCES OF GLOBALISATION AND ICT DIFFUSION 10. What are the consequences of the interactions between ICT use and globalisation on location of economic activities? 11. What are the consequences of outsourcing /offshoring driven by globalisation and ICT use on the home and host economies? 12. What is the impact of ICT use and globalisation on regional economic performance? The paper is organised as follows. Section 2 provides an overview of definitions and measures with the objective to provide a general conceptual framework for understanding ICT. Section 3 discusses the impact of ICT diffusion on economic performance, in particular the impact of ICT on structural change, productivity and output growth. Section 4 examines the spatial impact of ICT diffusion including ICT- related spatial transformations, the location of economic activities and the impact of ICT on regional economic growth. Section 5 provides an analysis of the key features of regional innovation systems in relation to ICT production and usage. Section 6 discusses the interactions between ICT and globalisation and their consequences on firms’ organisational structures, location of economic activities, and economic performance. Section 7 summarises the main theoretical arguments and concepts and articulates an integrated framework for the analysis of the interplay between, ICT diffusion, regional innovation systems, globalisation on regional economies. 4 2 ICT: Definitions and Measures At face value, ICT are a collection of technologies and applications, which enable electronic processing, storing, retrieval, and transfer of data to a wide variety of users or clients. According to Cohen, Salomon & Nijkamp (2002), ICT are currently characterised by: • very dynamic technological changes, with rapid penetration and adoption rates; • decreasing costs for new equipment and features; • a rapidly increasing range of applications and penetration in an increasing number of realms of professional and personal life; • an intertwined institutional market place, with the private sector acting in a decreasingly regulated environment (in most countries); • a production and services package dependent on a range of qualities of skilled human resources, and • a convergence of technologies. Unfortunately, there exists no clear or unambiguous agreement on the definition of ICT, or what sectors should be termed ICT sectors, which has been generally accepted (Schwartz, 1990; Malecki, 1991; Graham & Marvin, 1996). Furthermore, as ICT over time penetrate more and more sectors, more and more sectors deserve to be classified as ICT sectors. However, it is possible to provisionally delineate the providers of ICT in terms i) manufacturing of ICT, ii) wholesale and retail trade of ICT, iii) ICT network services, iv) other ICT services. These providers provide ICT to ICT users, i.e. to households, firms, and public sector organisations. To analyse the growth effects of ICT it is necessary to measure the extent of investment in ICT in the manufacturing sector, in the private service sector and in the public sector. Usually, statistical bureaus estimate investments from surveys among businesses specifically designed to capture investments. Based upon data over investments in well-defined asset groups, such as plant & machinery, infrastructure, dwellings, vehicles, and intangibles, such as R&D, with internationally agreed definitions, is it possible to estimate total investments by aggregation. However, this 5 doesn’t work in the case of investments in ICT due to a lack of generally agreed 1 definitions. To develop such generally agreed definitions for investments in ICT is by no means simple for several reasons. First, the investments are of at least four kinds: hardware, software, network infrastructure (communications equipment), and, in principle also 2 training of personnel to handle the equipment but only the three first types are normally capitalised. However, even at this relatively aggregated level comparability problems remain (Ahmed, Schreyer & Wölfl, 2004). When software is sold together with hardware, its value may be recorded as either software or included in the hardware value depending upon the circumstances. ICT components included in other 3 products, such as machinery, will not be directly recorded as an investment in ICT. The implication is that comparisons of investments in ICT in the manufacturing and the service sector may be biased, since substantial expenditures on ICT in the manufacturing sector might be recorded as intermediate consumption, while they are capitalised as investments in ICT in the service sector. Furthermore, there are large differences between countries regarding the extent to which expenditures for software are capitalised. Implementing ICT in an organisation entails reorganisation costs. These adjustment costs create a stock that yields future benefits. However, this investment is not measured as such in the national accounts (Oulton & Srinivasan, 4 2005). Second, we have the rapid improvements in capacities and speed and at the same time a drastic drop in costs of in particular the hardware. In many cases, it isn’t enough to measure expenditures on ICT at current prices. Instead a volume measure is needed that controls for changes in the price level of ICT products. Thus, price indices are needed to deflate expenditures at current price to get a ‘constant price’ measure. Due to rapidly decreasing production costs and strong competition, the prices of key ICT products has fallen drastically in recent decades. At the same time, their capacity has 1 The problems of measuring ICT investments and to make comparisons between countries are discussed in Ahmed, Schreyer & Wölfl (2004). 2 The fast development of ICT implies that the labour force must learn, relearn, train, and retrain, i.e. there is a continuous need for households and for entire economies to make specific investments in human capital as long as standardised ICT solutions are not established in each area. 3 Focusing exclusively on ICT investment products does not fully reflect the benefits of ICT diffusion within investment products or in the economy at large (Papaconstantinou, Sakurai & Wyckoff, 1996). 4 The EU KLEMS project provides data on ICT capital assets comparable across a number of OECD countries (see www.euklems.net ). 6 increased rapidly. Obviously, the construction of price indices for ICT products is no 5 easy task and a possible source of unreliable results in empirical studies. There are also other measurement problems involved (Howitt, 1998): (i) many of the knowledge-creating activities are not registered as part of GDP under conventional national accounting, (ii) the contribution of new or improved products to output is typically underestimated, and (iii) the arrival of new technologies makes machines and equipment that were designed for the old one obsolete, and it reduces the value of the skills that workers acquired for the old technology. However, a time series of investments in ICT in total and in the three major types is not enough to measure the effects of investments in ICT. What is needed is an aggregated, discounted measure of all historic investments in ICT equipment and 6 systems making up the current ICT capital stock , or rather of the flow of capital services from the stock of ICT capital. However, normally a strict proportionality between capital services and capital stocks but the ratio can vary between different types of assets. Thus, there is also an aggregation problem. Starting first with the problem of how to estimate the ICT capital stock (or the capital stock of any of the three major types, we assume that there exists a sufficiently long time series of investments in ICT at current prices and a suitable corresponding price i index. If we let the current price investment for ICT asset type i in year τ be I t−τ , i and the relevant price index be pt,0 , the productive stock of the actual type of ICT i asset Kt at the beginning of period t can be computed as: i T i i i i i Kt = ∑ τ=0 (I t−τ pt,0 )hτ Rτ (2.1) i 7 i where T represents the maximum service life of asset i , hτ is an age efficiency function representing the lower efficiency of older vintages of ICT capital goods, and i Rτ describes the probability of survival of capital goods over a cohort’s life span. 5 For an overview of the problems measuring ICT prices, see Ahmed, Schreyer & Wölfl (2004). 6 This becomes extra problematic at the sectoral level since ICT equipment might be rented and ICT services out-sourced. 7 In OECD work the average service life for different types of ICT capital is assumed to be 3 years for software, 7 years for ICT hardware and 15 years for communication equipment (Schreyer, Bignon & 7 Turning now to the aggregation problem, we need a means to effectively incorporate differences in the productive contribution of heterogeneous investments as the composition of investment and capital changes over time. User cost weights provide such a means since, under assumptions of competitive markets and equilibrium conditions, they reflect the marginal productivity of different assets (Jorgensen, 1963; Jorgensen & Griliches, 1967). User costs are imputed prices and reflect how much would be charged in a well-functioning market for a one period-rental of a given i capital good. Ignoring tax effects, the user costs of a capital good i , ut , are composed i of (i) the net rate of return rt applied to the purchase price of a new capital good pt , i (ii) the costs of depreciation, captured by the rate of depreciation dt , and (iii) the rate i i of change of the price of the actual capital good expressed by ψ t ≡ d ln pt : i i i i ut = pt (rt + dt −ψ t ) (2.2) The expression in the parenthesis represents the gross rate of return on an investment in a new capital good in year t . The gross rate of return on investments in infrastructure capital goods tends to be higher than for other types of capital goods. This is a result of the rapid obsolescence of ICT capital goods, which enters the user cost of capital in the form of purchase prices of new capital goods and via the rate of depreciation. Falling purchasing prices makes it less expensive to buy new capital goods but raises the costs of holding old capital goods. Depreciation rates may be computed according to different formula but they all shall reflect the relative loss of the value of a capital good due to ageing. Dupont, 2003). Of course, these assumptions are critical and one can wonder whether the assumed average service lifes might be to long given the rapid technological changes in the field of ICT. 8 3 The Economic Impact of ICT Diffusion We have in recent years been able to observe what can be described as an evolutionary process whereby economies at the national and the regional level and all their sectors are being transformed by the rapid development, adoption, and use of ICT innovations. In this respect, ICT functions as a new generic general purpose technology, which impacts these economies both broadly and deeply by generating a wide array of new products, production processes and services (Brynjolfsson & Kahin, 2000; Mowery & Simcoe, 2002). Carlsson (2003) takes this idea one step further arguing that ICT, which involve among other things a combination of digitalisation and the Internet, seem to have broader applicability than previous general-purpose technologies. It not only affects all manufacturing industries but also, and even more so, all different service industries, which account for an increasing and dominating share of the economy in developed economies. Furthermore, it has given rise to new industries within both the manufacturing and the service sector. However, one should observe that it is a common feature of new general purpose-technologies that it takes a long time before they are implemented (including the necessary organisational changes) and used in such a way that they could develop their abilities to the fullest (David, 1991). It is unquestionable that the effects of the development, spread, and use of ICT go much further than changing the industrial composition of developed economies. ICT are playing an increasing role in economic growth, capital investments, and other aspects of the macro-economy (Brynjolfson & Kahin, 2000). The emergence of new goods and services as well as changes in the characteristics of old goods and services due to the use of ICT, including the ways good and services are produced and distributed lead to changes in market structures and competitive conditions affecting and creating new opportunities for small firms and entrepreneurs. As ICT are routinely deployed in organisations to re-engineer processes, gain new strategic advantages, or network across organisational boundaries, they change both the internal organisation of companies and other organisations and the relationships between companies and organisations (OECD, 2002). The adoption of ICT allows for a reduction of transaction costs and leads possibly to more efficient markets (Malone, Yates & Benjamin, 1987; Lee & Clark, 1997). 9