Staff Working Paper ERSD - 2003 - 02 August, 2003
World Trade Organization
Economic Research and Statistics Division
ICT, access to services and wage inequality
Hildegunn Kyvik Nordås: WTO
Manuscript date: August 2003
Disclaimer: This is a working paper, and hence it represents research in progress. This paper represents the opinions of individual staff members or visiting scholars, and is the product of professional research. It is not meant to represent the position or opinions of the WTO or its Members, not the official position of any staff members. Any errors are the fault of the authors. Copies of working papers can be requested from the divisional secretariat by writing to: Economic Research and Statistics Division, World Trade Organization, rue de Lausanne 154, CH-1211 Genève 21, Switzerland. Please request papers by number and title.
ICT, ACCESS TO SERVICES AND WAGE INEQUALITY
*Hildegunn Kyvik Nordås
This paper discusses how information and communication technology (ICT) affects the quality and reach of consumer services. These services need to be provided locally, but consist of several components, some of which can be digitised and transmitted over long distances. A general equilibrium model is developed and numerical simulations in a stylised two-factor, two-region, centre-periphery setting are presented. Trade in intermediate services improves the quality of consumer services enormously in the periphery, but may reduce the quality at the centre. Trade in intermediate services also has a dramatic impact on skilled workers’ wages in the periphery, both relative to unskilled workers in their own region and relative to skilled workers at the centre and leads to a more equal distribution of income both between the centre and the periphery and within the periphery.
JEL codes: D23, F12, F13, R12, R13.
Keywords: Trade in services, transaction costs, income distribution, ICT
* I am grateful to Hans Jarle Kind, Leo A Grünfeld, Carsten Eckel, Kåre Petter Hagen, Alexander Keck, a seminar audience at the University of Bergen, participants at the Nordic Conference in Development Economics 27-28 May 2002 and two anonymous referees for useful comments on this and previous versions of the paper. Funding from the Norwegian Research Council is also gratefully acknowledged
ICT, ACCESS TO SERVICES AND WAGE INEQUALITY
Hildegunn Kyvik Nordås
World Trade Organization
Does diffusion of information and communication technology (ICT) eliminate the relevance of distance? Or does the communication revolution lead to increased centralisation? And are we witnessing a growing digital gap that leads to increased income disparities both within and across countries? These questions have given rise to a lively debate among scholars as well as policy makers and the general public. The low cost of storing and transmitting huge amounts of information allows people to access information, search for suppliers/customers, enter into contracts and exchange services without much regard for distance, goes the argument. However, the information economy is also characterised by rapid technological change, a high rate of innovation and frequent introduction of new goods and services that are customised and often have short product cycles. In this environment of complexity, some essential information cannot be digitised but is entrenched in relationships and communicated through direct interaction (Leamer and Storper, 2001).
This paper explores the impact of the use of ICT in information-intensive consumer services such as health, education, financial services or entertainment. Health and education are important for social and economic development, account for a relatively large share of national expenditure and employ a relatively large share of a community's skilled workforce, while access to such services is important for where people choose to live. Entertainment is already a global industry and the most popular performances and performers are known worldwide, largely thanks to ICT. Audio and video appear to be reasonably good substitutes for live performances while locally provided entertainment can be combined with entertainment provided from a distance for example in sports-cafés, internet-cafés etc. Tele-banking and e-banking have rendered personal visits to the bank unnecessary for most transactions and provided customers in suburban and rural areas with a wide range of financial services.
The question analysed in this paper is whether and to what extent we will see similar developments in social services such as health and education. How will for example access to information on the Internet and interactive learning programs affect a) the quality of education; b) the role of the local teacher/lecturer; c) the demand for her services; d) her relative income? These questions are closely related and are therefore analysed in a general equilibrium framework where consumer services are provided locally but ICT allows the local provider to source intermediate inputs from suppliers outside the region.
Distance learning has existed for more than a century using correspondence by mail, adding TV from the 1950s and the Internet from the 1990s. In the past distance learning constituted a relatively small market for adult learning. However, according to the OECD (2001) more than a third of the member countries' population between the age of 25 and 60 participated in adult education and training in 1998.
Furthermore, demand for adult training is growing rapidly as the labour market becomes more flexible and technological progress creates the need for frequent skills upgrading and updating. Universities, regional colleges, large companies and a number of specialised suppliers have entered this market in recent years (OECD, 2003). In tertiary education e-learning, distance learning and traditional learning are increasingly mixed to provide more flexible course packages (OECD, 2001). In Norway small schools and schools in rural areas have fewer students per computer than large schools and schools in urban areas at all levels of education (Quale, 2000). This may indicate that ICT is used to compensate for less variety in the choice of courses, excursions etc. in smaller communities in addition to teaching computer skills.
In the health sector, local doctors and nurses make use of distant laboratories for analysing tests and hospitals outsource routine tasks such as the typing of patient journals to lower-cost regions or countries. Finland and Norway have adopted a national telemedicine policy and are investing heavily in infrastructure and training of personnel (WHO, 2002). There are also joint Nordic projects in the northern counties of Finland, Norway and Sweden. Monitoring patients at home, providing health information on digital text-TV, supporting the parents of children with various chronic diseases, securing electronic links between patients and their practitioners, mental health related online discussion groups and many more applications have been developed and introduced (Norwegian Center for Telemedicine, 2003). Finland, Norway and Sweden are all sparsely populated, particularly in the northern counties and the telemedicine projects aim at providing quality services to these remote areas.
An additional recent example of how ICT can enhance health services and reduce the relevance of distance is quoted from the WHO (2003): "On 17 March 2003, WHO called upon 11 laboratories in 9 countries to join a collaborative multi-center research project on SARS diagnosis. This network takes advantage of modern communication technologies (e-mail; secure web-site) to share outcomes of investigation of clinical samples from SARS cases in real time. Daily assessment of research results supports immediate refinement of investigative strategies and permits instant validation of laboratory findings."
The rest of the paper is organised as follows: Section two briefly discusses related previous research. Section 3 develops the model in a closed economy and analyses its properties. The model is extended to a two-region setting with exogenously declining (iceberg) transport costs in section 4, where I also present numerical simulations. I take as given that there are communication networks in place in the same way as trade models take as given that adequate transport infrastructure exists. Section 5 summarises and concludes.
RELATIONS TO PREVIOUS RESEARCH
There is a growing body of research on the relationship between transaction costs, trade and location of
1 A major finding in this literature is that economic activities in which vertical linkages between firms.
firms are important tend to cluster. Furthermore, when transport costs are significant and labour immobile between regions, a centre-periphery pattern is formed where the centre pay the highest wages even when the regions have the same factor endowments. The model developed in this paper adds more realism to the standard geography model by accommodating two factors of production, a non-traded consumer services sector and a fixed cost of entering a distant market. I show that in this setting differentiated services will agglomerate in one region only if the regions are asymmetric. In addition the introduction of two categories of labour allows me to analyse income distribution within economies as well as between
A relatively early analysis of trade in business services over the Internet is a study by Harris (1998). He applied a small open economy setting where prices are exogenously given in the world market and where digitised information is a perfect substitute to directly communicated information. He shows that the skills premium increases with skilled labour supply and that the location of business services firms is indeterminate. The general equilibrium model developed here in contrast, allows prices to adjust to changing demand and trade costs. Then the skills premium declines with relative skilled labour supply and the industrial structure in each region is determined, although the model has multiple solutions.
Several studies have found that sunk costs of entering export markets are significant (Roberts and Tybout, 1997; Bernard and Bradford Jensen, 2001). Entering a distant services market through arms-length trade involves fixed costs that cover software, translation, organisational changes and skills
2upgrading. The cost of software for distance learning for example can be between $16 000 and $150 000 dollars per year (Anderson, 2003). Venables (1994) provided an analytical framework for analysing the impact of such fixed costs, introducing them in a Dixit-Stiglitz type model with symmetric countries and one production factor. He shows that trade liberalisation, modelled as a decline in variable iceberg trade costs increases the proportion of trading firms and reduces the total population of active firms. The latter effect is due to the fixed cost of exporting, which increases the break-even scale of exporting firms. The model developed here extends Venables' framework in order to assess income distribution and welfare implications in asymmetric markets.
To summarise, the main contribution of this paper is to provide a realistic framework for analysing how ICT affects reach and quality of essential services, income distribution and welfare in small communities. It shows that agglomeration does not arise spontaneously if countries initially have the same
1 See for example Fujita, Krugman and Venables (1999) for a comprehensive study of these issues. 2 See Bresnahan et. al. (2002) for a discussion of organisational restructuring following introduction of ICT.
factor endowments, and it spells out the trade-off between universal services and variety of services. Finally, as opposed to most models of trade where scale matters, it is the small trading partner that gains the most from trade, while welfare in the large trading partner may be reduced compared to autarky.
THE MODEL - AUTARKY
The economy has three sectors, labelled X,Y and Z. The X-sector produces complex consumer services and its technology is assemblage of differentiated intermediate service inputs. These inputs are produced by the Z-sector, which consists of n firms each producing one service using skilled labour only. I apply the standard Dixit-Stiglitz framework where there is a fixed cost involved in setting up each differentiated service. The Y-sector represents an aggregation of perfectly competitive industries producing goods and services employing both unskilled and skilled workers. The number of unskilled and skilled workers in the economy is given exogenously. Equations (1)-(3) describe the technology in the three sectors:
；1，；Y，LS (1) y
1/?n??? (2) X，z：i??i，??1
1 (3) z，f？siib
Equation (1) is a standard Cobb-Douglas constant returns to scale production function where L represents unskilled and S skilled labour. Equation (2) represents the familiar Dixit-Stiglitz framework where production increases both with the quantity of each input, z, and the number of inputs, n. The elasticity of
?，1/(1，?)substitution between any pair of intermediate services is given by and is assumed to be
larger than unity. Each firm in the Z-sector produces its service subject to an increasing returns technology given by equation (3) where f is a fixed cost in terms of skilled labour. As usual in Dixit-Stiglitz models, only one firm produces each input and each firm produces only one product. The number n thus represents
both the number of firms in the Z-sector and the number of differentiated services being supplied to the X-sector. I apply the standard mark-up price rule:
bvq， (4) ?
where v is the unit cost of skilled labour. I further assume that there is free entry of firms into the Z-sector. Firms will in that case enter until the profit of the last firm entering is zero. The non-zero profit condition is the same for all firms due to symmetry and reads . Inserting this in (4) yields the unique q，fvz？bv
size of the service producing firm:
?fz， (5) b(1，?)
The X-sector is competitive, implying marginal cost pricing:
2~~PPbv(?，1)/?(?，1)/?xxP，qn，n！0;(0 where (6) x2?~n~n
Consumers have identical preferences described by a Cobb-Douglas utility function, which implies that consumers spend a fixed share of their income on each good:
Market equilibrium in the X-sector can now be determined by the following demand and supply conditions:
?f1/??PX，(1，)(wL？vS) and X，n (8) xb(1，?)
We finally turn to the labour market in order to close the model. Employment of skilled labour in the Z- and Y-sectors, and the skilled labour market equilibrium are given by:
nfS，n(f？bz)， (9) z1，?
PY；(1，)wLvS(1，；)?(？)yS (10) ，，yvv
Inserting (9) and (10) in (11), using (8) yields the allocation of skilled workers between sectors Y and Z:
；??(1，)(1，)，;， (12) SS SSyz1，；?1，；?
Allocation of skilled labour depends on the technology in the Y-sector and consumer preferences. The more skills intensive the Y-sector and the higher the income share consumers spend on Y-products, the less skilled workers are employed in the Z-sector. The number of services produced in the economy is determined by (9) and (12):
??，，(1)(1)， nS (13) ，；?f(1)
The number of Z-sector firms is larger the larger the stock of skilled workers, the smaller the fixed cost of producing services and the smaller the elasticity of substitution between producer services in the assembly industry. The intuition behind the latter point is that when producer services can be easily substituted, there is little to gain from having additional varieties. The linkage between the elasticity of substitution in the X-sector and the number of firms in the Z-sector is an externality between the two sectors. However, since ? also appears in the mark-up rate on the marginal cost of z, the externality is internalised by the market and it is therefore termed a pecuniary externality.
The skills premium, defined as the income earned by skilled workers over and above that of unskilled workers, can be found by using (10) and (12).
vL；?()1， (14) ，w；?S
We notice that the skills premium declines with the relative endowment of skilled labour, but it does not depend on the size of the total labour force as long as the ratio L/S is constant. In order to assess the relation between market size and real income, I deflate the wage rates by the price index
?1，?;；;；P，P?P(1，?) where P is the marginal cost in the Y-sector given by yxy
；1，；;；;；P，w；v(1，；) and Pis given by (6). Setting unskilled labour as the numeraire (w = 1), using x y
(13) and (14) and collecting parameters, the price index reads:
；?1，L，?(?，1)(1，?)/? (15) P，；S??S??
where ； is a positive constant. We notice that the price index declines with a proportionate increase in L and S. This is because an increase in the absolute number of skilled workers increases the number of differentiated intermediate services (equation 13) and thus lowers the price of X (equation 6). Consequently real wages increase while the skills premium remains constant. A proportionate increase in the stock of skilled and unskilled labour can be interpreted as the integration of two regions with the same relative factor endowments. If the two regions differ in size, the smaller region will experience a larger increase in the variety of intermediate services assembled in the X-sector, P will decline more in the x
smaller region and real wages will catch up with the larger region. Full integration of the services markets is, however, not technologically feasible, and I therefore analyse the impact of partial integration in the next section.
THE TWO-REGION CASE
In this section I explore the welfare and income distribution effects of trade between two regions labelled Big and Small, which can be interpreted as a rural and an urban region within a country, a small and a large
3country or a rich and a poor country. In all cases the market size is defined by the size of the labour force.
3 In poor countries the labour force may be interpreted as those working in the formal sector, or efficient