EJISDC (2011) 48, 4, 1-18

1

INFORMATION AND COMMUNICATION TECHNOLOGIES (ICT) AND MEXICAN MANUFACTURING EXPORTS Ana L. Valderrama Santibáñez School of Economics National Polytechnique Institute Mexico [email protected]

Omar Neme Castillo School of Economics National Polytechnique Institute Mexico [email protected]

ABSTRACT The paper analyzes the effects of changes in the purchase of computer equipment and peripherals (investment in information and communication technologies - ICT) on exports by classes of Mexican manufacturing industry using data from the Annual Industrial Survey and Industrial Stan OECD data base for the period 2003-2006. The paper distinguishes among industries with high and low technology. We used a cross-section methodology to evaluate the effect of such technologies in each year of the period. The results indicate that investment in ICT has a positive effect on manufacturing exports, particularly in industries with low technological content that can integrate ICT into their export activities. Keywords: exports; information and communication technologies; manufacturing industries. 1. INTRODUCTION In the last few decades, economic literature has been exposed to the concept of technology, which has evolved and no longer only refers to machinery and equipment, but also to all the “know-how”, information and knowledge. We are facing an increasing barrage of information ever seen. With the emergence of Internet, the information economy has changed the relationship among individuals as well as among countries. The changes created by a global economy based on knowledge, investment and use of information and communication technologies (ICT), usually defined as computer hardware, software and telecommunications equipment, has become an explanatory factor of advances in productivity, international trade and economic growth in industrialized countries. The determination of ICT impact has been globally studied by a number of authors since the late nineties. For instance, studies for industrialized countries indicate that increase in productivity and production in the second half of the nineties, are due largely to the adoption of ICT (Jorgenson et al., 2005). In general, studies of ICT effects on industries conclude that ICT-intensive industries are those whose product, employment and labor productivity has grown further (for example, Oliner and Sichel, 2000). Thus, there are numerous studies that found a positive relationship between ICT and economic activity, which has generated an expansion in ICT use in the economies of the world. Generally, they point out a positive and significant correlation between ICT and gains in productivity. In this regard, Papaioannou and Dimelis (2007) estimated a positive and significant ICT growth effect in developing and developed country, with the highest impact observed in the former. At the same time affirm that this effect stems entirely from the hardware and communication components. Kraemer and Dedrick (1994), found a positive correlation between growth in ICT investment and use and growth in both GDP and productivity in 12 Asia-Pacific countries. The implication of this finding is that countries with higher growth rates in ICT investment achieved consistently higher growth rates of GDP and productivity, confirming the hypothesis of ICT-led development. Finally, Oulton (2002), states, in a study of ICT and productivity in the United Kingdom, that the growth of ICT output has contributed about a fifth of GDP growth from

EJISDC (2011) 48, 4, 1-18

2

1989 to 1998 in that country. Besides, he highlights that ICT capital deepening accounted for 48 per cent of the growth of output in 1994-1998. This result are due both to production of information technology, as well as the complement factors, such as human capital and research and development (R&D) which link ICT producers with the rest of manufacturing sector. According to the Declaration of Principles adopted at the first phase of the World Summit on the Information Society in 2003, ICT is an important factor for growth because increases efficiency and productivity, especially among small and medium enterprises. Policies that promote productive investment and allow companies to make the changes necessary to benefit from ICT are probably the most advantageous (WSIS, 2004). In particular, in its efforts to improve productivity, output growth and export market shares, most developing countries, including Mexico, have established their development strategies considering ICT. In this sense, policies to integrate Mexican economy to new forms of interaction in global markets have implemented. The 2001-2006 National Development Plan has established the objective to raise the country's competitiveness by promoting the use and exploitation of ICT. It was noted that the incorporation and use of the latest science and technology advances should be based on a strategy that included both the production and promotion of the use of ICT. The Mexican government has made efforts to promote ICT growth, showing a strong intention for adaptation and improvement of this technology, however, does not indicate the mechanism by which investment in this sector will impact economic growth or exports. In other words, the question is can the growth of this sector generate technological spillovers that the manufacturing industry can exploit in its processes and ultimately gain international competitiveness? Additionally, it is worthy to note that the structure of manufacturing industry is differentiated by technological contents. The impact of ICT must be different in each group, for example because the dynamic of innovation existing in high technology industries demands new technologies. Consequently, it can be expected different effects of these technologies in each sector or industry. Thus, the aim of this paper is to demonstrate that there is an effect of ICT on Mexican manufacturing exports to the United States and also that this is different depending on the technology content of industries in the period 2003-2006. One contribution of this paper is the accounting of ICT impact at industry level for Mexico, filling the gap in empirical studies with respect to exports, since the limited analysis at industrial level in Mexico. The paper is structured as follows. The next section reviews concepts of New Economy and ICT, establishing how these technologies are represented in Mexico. Section three analyzes the use of these technologies in manufacturing subsectors. In fourth section, to verify the impact of these technologies on manufacturing exports, a model of cross section for four years is proposed; the analysis is done by industrial groups. Finally, concluding remarks are presented. 2. NEW ECONOMY OR ECONOMICS OF INFORMATION AND ICT From the technological revolution of computers and telecommunications in the late nineties, the term of new economy arises. It refers to the growth of ICT-producing sector and the spread of its use in all sectors of the economy. Machlup (1962) describes the New Economy with the term “knowledge-based industry”. Porat (1977) explains that an economy becomes an "economics of information" when the labor related to information overcome the labor in other sectors. He distinguishes two economic sectors. The primary sector is the generator of information and includes activities related almost entirely to creation or management of information (scientists, writers, librarians, etc.). The secondary sector includes industries not directly related to information,

EJISDC (2011) 48, 4, 1-18

3

but whose activity involves information as a secondary aspect; it refers to industries that produce information for internal use in the production of goods1. On the other hand, Castells (1999) explains that New Economy has two characteristics: a) globalization, which implies more extensive and changing markets, free movement of goods and capital, integration of supranational business areas or multinational corporations, and b) development of ICT, leading to improvements in productivity, absence of rising costs due to product life cycle, greater relevance of innovation and knowledge against capital, the emergence of new professions, etc. Following Castells (1999), the New Economy affects all agents, especially in access to information. Firms have better access to information and location of raw materials, thus reducing costs and generating changes in the way of selling their products through electronic commerce. Consumers now can access to global market from home via Internet. Workers can perform certain tasks at a distance. In general, appear new products, services, and jobs based on information and knowledge that are used thanks computers and Internet. Internet opens up new spaces in the model of each business since it presents opportunities to reduce costs, enables new ways to serve customers and manage suppliers, adopts new marketing systems and represents the opportunity to create new products or services. Also, to Nordhaus (2001) the New Economy is the acquisition, processing, transformation and distribution of information. The major components are hardware, mainly computers that process information, and software, which runs the entire system. At socioeconomic level there is the widespread use of software; he also finds some economic characteristics such as good of experience2. On the other hand, according to Kelly, there are basic principles of the new economy, based on information, communications and intangibles. Physical resources are relatively less important, because now the information and services are central. Given this feature, individuals are the most important asset in the new economy, not only for their physical abilities but for their knowledge and skills. In an environment dominated by the Internet, the new economy does not consider geographic location in the decision to establish a business3. On the supply side, the firm has many more opportunities to sell and they have effects on efficiency; greater benefits are obtained through “infomediaries” who provide the information. In this sense, the intermediaries are part of the old economy and way of doing business. This flow of information reduces the cost of transactions; for instance, Internet allows transactions to be handled at high volume serving personal transactions without increasing the process. Thus, considering the three most commonly used definitions of new economy it can be said that the feature that defines it is the expansion of both production and use of information technologies, that is why sometimes the new economy is called economics of information. The new economy refers to an economy centered on information and knowledge, where information is considered as input, output and strength that move the economy and relates, through knowledge, economic agents. In consequence, information and knowledge 1

The primary information sector includes the production of knowledge and invention; distribution of information and communication; risk management; services for processing and transmitting of information; information goods (computers); selected government activities (education and postal service); support facilities; wholesale and retail products and services information. The secondary sector includes “all information services produced for internal consumption by government and firms outside the sphere of information”, except governmental activities in the primary sector of information such as education and printing, but including government activities such as planning, coordination, monitoring, control, evaluation and decision making. It also includes areas of firms involved in information tasks, such as internal services of data processing. 2 They are goods purchased when are relatively unknown and are of no longer interest once the buyer knows them. These goods presents high sunk costs of entry and marginal production costs than tend to zero. 3 http://www.wired.com/wired/archive/5.09/newrules.html?pg=2&topic

EJISDC (2011) 48, 4, 1-18

4

can be considered as two, related but different, key factors in productivity, production and, ultimately, with the expansion of domestic market to foreigners, in exports of economies. Regarding the economics of knowledge and the relationship with ICT, Dahlman and Andersson (2000), stated that in this type of economy, knowledge is created, acquired, transmitted and used effectively by firms, organizations, individuals and communities to promote economic and social development as well as firm benefits; so, the creation, distribution and use of knowledge is the greatest engine for growth, wealth and employment. According to the above, economics of knowledge exists around the idea of technological change and human capital and is generally accepted that economics of knowledge and economics of information are closely related. Another element necessary for the existence of economics of knowledge is people. The importance of human capital in innovation and income generation is highlighted. This participation occurs mainly through productivity increases, since it facilitates the adoption of technologies and production processes more sophisticated and flexible. Human capital contribution to growth is pointed by several studies which show that part of the country's economic growth is explained by total factor productivity and that a key element of this growth is the quality of education. In addition to the growth of ICT, Edwards (2002) explains that to take full advantage of ICT, it is necessary to invest in complementary areas such as education, infrastructure, public services and R&D. Therefore, the idea is that, for example, education generates people with autonomous thinking and ability to acquire knowledge the rest of their life. As mentioned, there are industries involved in production in the economics of knowledge and industries that generate and use information. The direct impact of the growth of ICT-producing sector is an increase in high technology investment in the same sector and in the sector that uses it, that is, in the export sector. It also generates changes in labor force (it becomes more skilled) and, thereupon improvements in productivity.Furthemore, information is a product whose creation requires a significant amount of physical and intangible inputs, but whose reproduction is virtually free, leading to an economy very different from traditional economy of physical products. Information products can be reproduced at virtually no cost (marginal production cost of virtually zero), which means that once created, the lack of this product can be eradicated virtually without cost. Since the new economy consists of a technological transformation and that it is of general purpose, there are two effects of ICT on economy. First, an increase of productivity in the ICT-producing sector which, in turn, increases quality, speed and capacity of hardware, software and communication equipments (Argandoña, 2001). The lower costs and prices lead to a higher demand and production so that the sector increases its importance in the whole economy. The second effect is derived demand. Since the costs of ICT have been reduced, the other sectors of the economy that use these technologies replace capital and labor by computers, software and communication equipments, and even create new products or services with those technologies. This dynamic tend to increase the productivity of other factors, mainly of the labor. Hence, changes in the demand of other factors are expected, mainly regarding the qualification of manpower, new management systems, infrastructure, etc., which will change the productivity of capital. Finally, to the extent that technological progress in ICT is spread from some sectors to other, the spillover effects increase factor productivity in sectors not directly related to ICT and throughout the economy. To achieve the second effect and to use ICT from a microeconomic view, it should mix different levels of use of such technologies. In general terms, these levels refer to three types: i) infrastructure, telecommunications network and Internet; ii) platform, which connects networks to the application level: computer, mobile

EJISDC (2011) 48, 4, 1-18

5

phone, mobile Internet, iii) application, is the electronic commerce between suppliers and among firms and consumer, electronic data exchange and supply chain management. On the other hand, in an economy based on knowledge and information as the basis of production, it must be distinguished between the concepts of information and knowledge which tend to be used interchangeably (especially when codified knowledge is identified with the information), although there are differences between them. From the definition of Machlup (1983) that considers information as a stream of messages or meanings that adds, restructures or changes knowledge, information is differentiated of knowledge since other approaches. The main difference is perhaps that the information is descriptive whereas knowledge is predictive (Machlup, 1983). In this sense, information is a stream of messages, while knowledge is created precisely through this flow of information, depending on characteristics of the subject. Information is data assigned with certain meaning, and provides a new way to interpret events or objects, and consequently, information is a necessary means to acquire and construct knowledge (Kock et al., 1997). Moreover, the distinction between knowledge and information can be associated with the terms of stock and flow (Ancori et al., 2000). However, Cowan et al. (2000), indicate that knowledge cannot be interpreted as the accumulation of information flows, but as a very complex structure with interconnected parts. Also accepting that there is a difference between information and knowledge is due to issues related to the "conversion" of knowledge in information through the "codification" of it. On the contrary, since the reverse process is not symmetric, i.e., reproduce knowledge from the information; it cannot be regarded information and knowledge as synonymous (Ikujiro and Noboru, 1998). Thus, for purposes of this paper a review of export growth and structure of the ICTproducing sector for 2003-2006 for Mexico is done; it is measured the growth in demand for computers in Mexican manufacturing. The behavior of exports in manufacturing is assessed and the correlation among exports and the investment and use of ICT is checked. In the several studies that evaluate the impact of ICT on economy there are difficulties in measuring these technologies. Jorgenson and Stiroh (2000) define information technology including computers, software and communication equipments. Oliner and Sichel (2000) include computers and semi-conductor integrated to computers. Nordhaus (2001) defined this variable as machinery, electrical equipment, telephones, telegraphs, and software. He includes three major industries: machinery and equipment; electronics and other electrical equipment, telephone and telegraph, and software4. Although there is no consensus about what is the best indicator, there are certain indicators of ICT accepted by literature, which are generally grouped into four main categories (Mum et al., 2002): i) knowledge creation, which includes spending on R&D as a percentage of GDP, number of researchers, researchers per capita and patents per capita; ii) the acquisition/transfer of knowledge, for example, the share of imports with technological content, number of multinational corporations established in the country, and number of firms in ICT sector ICT, iii) dissemination of knowledge; which includes ICT spending, ratio of Internet access cost to per capita GDP, percentage of workforce with secondary education; iv) 4

At a general level, the International Standard Industrial Classification (ISIC) includes in the ICT sector: Manufacturing: office machinery, accounting and computer; wire and cable insulation; electronic valves, tubes and other electrical components; radio and television transmitters and apparatus for telephony and telegraphy lines; radio and television receptors; sound or video playback apparatus and other goods; instruments and devices for measuring; checking, analysis, and navigation; other goods, except industrial process equipment and control equipment of industrial process. Services: sales of machinery, equipment and consumables; telecommunications; rental of office machinery and equipment including computers; computing services and related activities.

EJISDC (2011) 48, 4, 1-18

6

knowledge application, proxied through the ratio of workforce with university and postgraduate education, percentage of knowledge-intensive workers in the labor force, range of business behavior. In any case, it must be recognized the difficulties in defining the precise content of these technologies within a changing world. Besides, according to the definition of ICT, the sector that falls under this idea is “Mass Media Information” which includes activities that transform information into a good, as well as the activities of distribution. Accordingly to North American Industry Classification System –NAICS-, within this sector 35 classes of products are included (annex Table A)5. Not all industries within the information sector of mass media can be considered as ICT. These technologies include tangible and intangible, so that ICT will cover 16 product classes, eight classes of manufacturing industry and eight included in mass media information directly related to computer and communication equipments. So, the products shown in Table 1 are considered as ICT. Table 1: ICT in Mexico Class Description 334110 Computers and peripherals manufacturing 334210 Telephone apparatus manufacturing 334220 Radio and TV broadcasting and wireless communication equipment manufacturing 334290 Other communications equipment manufacturing 334310 Audio and video equipment manufacturing 334410 Electronic components manufacturing 334519 Other measuring and controlling device manufacturing 334610 Magnetic and optical recording media manufacturing 511210 Software edition 513319 Telegraph and other wired telecommunications 513321 Mobile telephones 513329 Other wireless telecommunications, except satellite services 513330 Resale of telecommunications services 513340 Satellite services 514210 Electronic data processing 517111 Fixed telephony (traditional) 3. ICTS AND MEXICAN MANUFACTURING INDUSTRIES This paper explains how investment in new technologies, as well as physical capital and labor (including knowledge), are related to Mexican exports. To this end, technological structure of manufacturing is described, since differences in the needs of each productive factor in each industry, create differences in the use of new technologies. Therefore, Pavitt´s classification (1984) is used to divide the manufacturing sector and, in turn, to describe both the technological sector and the traditional sector. Industries are classified following Dutrénit and Capdeville (1993) and Unger (2002)6. In the Mexican case, production data are at product class, data of usage for different classes of manufacturing industries are proxied through investment. Thus, ICT sector in Mexico is described as well as its development and the use of ICT in manufacturing sector is discussed. According to the OECD (2002) ICT are those devices that capture, transmit and 5

(http://www.inegi.org.mx/est/contenidos/espanol/metodologias/censos/scian/menu.asp). The industries are classified as high and low technology. For availability of space this characterization is not shown, however, may be obtained by request via email. 6

EJISDC (2011) 48, 4, 1-18

7

display data and electronic information and that support economic growth and development of manufacturing and services. According to Haag et al. (2004) ICT comprise any computer that people use to work with information, support information and process the information needs of an organization. ICT includes computers, Internet, mobile phones, and anyone similar device (from now computers). From these definitions, computers are the protagonists in ICT because the use of other products or services in this sector depends on them. Therefore it is important to know the evolution of the production value of machines for processing, computers and peripherals. In this way it can be known the market size of these products and therefore to have a proxy to growth in the use of ICT. Thus, in 1998-2000 the production value of computers increased. From 2001 it shows a downward trend in both industries that seems to improve in 2006 (Figure 1). For reference, the exchange rate between Mexican pesos and US dollars is shown in Figure 2. Figure 1. Production value of ICT equipment (millions of pesos 2003) Computers and peripherals

60,000

Information processing machines

50,000 40,000 30,000 20,000 10,000

0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Source: own elaboration based on data from AIS, INEGI.

Figure 2. Exchange rate between Mexican pesos and US dollars 12 11 10 9 8 7 6 1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Source: own elaboration based on data from BANXICO. http://www.banxico.org.mx/sistema-financiero/estadisticas/mercado-cambiario/tiposcambio.html

On the other hand, since only the 2004 Industrial Census is available, it is done an analysis of ICT use in Mexican manufacturing at sub-sector level for the beginning of the

EJISDC (2011) 48, 4, 1-18

8

period7. Also, since ICT has effects on the economy mainly in two areas, the management or production processes, Table 2 shows the extent of ICT use (including computers and Internet) in different activities. The use of ICT in Mexican manufacturing is divided into four activities that are classified in management use or technical processes use. These activities are subdivided according to how ICT is integrated into different activities and depends on the complexity of its use. It is possible that firms use ICT in activities that belong to more than one group. In fact, the uses are cumulative, i.e., it is expected that those sub-sectors that uses ICT to develop software also use them in technical processes and likely in administrative processes and with customers and suppliers. However, the inverse argument is not true, because those sub-sectors that use ICT in management do not necessarily use them in technical processes. One reason is the difference in the technological requirements of each industry. It is expected that industries within diffusers technologies sub-sectors or innovation-producing sectors, uses ICT to develop programs or improvements in technical processes. On the other hand, more traditional industries do not need to develop these technologies, since they demand them from the other group. Another reason is that firms require greater investment in capital goods, infrastructure and human capital in order to use ICT in the technical processes. In general the high investments in these areas are limited by financial and technological structure of the traditional industries. The exceptions are related to market structure of each sector or to linkages between different industries. Besides, in what follows it is analyzed the impact of ICT in manufacturing exports. The variable used is “purchase of computer equipment” from the Annual Industrial Survey (AIS) of INEGI defined as the value of the purchase of computer equipment and peripherals by the economic unit that are not integrated to machinery and production equipment such as computers, communication networks, printers and scanners, among others8. Since ICT is formed by some manufacturing classes plus those in mass media information directly related to computer and communication equipments, it is observed that ultimately the variable computer equipment and peripherals is a proxy for the use of ICT in manufacturing. Also, to measure exports (X) data from the Stan Data Base from OECD according to ISIC Rev. 2 is used; this data are homologated to those obtained from the NAICS. Thus, the aim is to identify the relationship between exports and the level of ICT use in manufacturing industries. Table 3 shows the relationship between exports and purchases of computer equipment of manufacturing sector as a whole. The exports grew 19.7%, while the purchase of computer equipment and peripherals grew 14.7%. The ratio of computer equipment per unit of export grew at a small rate during the period for total manufacturing (3.8% growth). On average, for each Mexican peso of X generated in high-tech industries (HTI), 0.09 pesos of computer equipment was bought. In contrast, low technology industries (LTI) bought 0.0064 pesos of ICT for each peso exported. Also, HTI had the highest investment in ICT in 2004 and subsequent years registered a lower level but higher in comparison to total investment in manufacturing. LTI registered the biggest investment in 2003 and it has fallen since then. In both cases there are two blocks 2003-2004 with high investment, and 2005-2006 with low investment. However, it is always the high-tech sector who invests more in these technologies.

7 8

http://www.inegi.org.mx/prod_serv/contenidos/espanol/biblioteca/Default.asp?accion=1&upc=702825172817 http://dgcnesyp.inegi.org.mx/cgi-win/bdieintsi.exe/NIVR250130#ARBOL

EJISDC (2011) 48, 4, 1-18

9

Use of Internet in relationships with customers and suppliers

Use of computers in technical or design processes

Development of programs or software packages to improve its processes

Food manufacturing Beverage and tobacco product manufacturing Textile mills Textile product mills Apparel manufacturing Leather and allied product manufacturing Wood product manufacturing Paper manufacturing Printing and related support activities Petroleum and coal products manufacturing Chemical manufacturing Plastics and rubber products manufacturing Nonmetallic mineral product manufacturing Primary metal manufacturing Fabricated metal product manufacturing Machinery manufacturing Computer and electronic product manufacturing Electrical equipment, appliance, and component manufacturing Transportation equipment manufacturing Furniture and related product manufacturing Miscellaneous manufacturing Total Source: Innovation and Research Module. Economic Census 2004. INEGI.

Use of computers in administrative processes

Subsector

Total establishments

Table 2. ICT Applications by Manufacturing Establishments, by Economic Activity Subsector, 2003

116,303 7,005 4,127 13,353 22,429 7,491 17,729 2,654 14,521 280 3,073 4,361 25,037 1,250 49,650 2,467 791 918 1,978 21,569 11,732 328,718

5,032 1,511 657 674 3,317 1,612 903 851 5,792 258 2,062 2,569 1,856 557 4,927 1,367 683 679 1,204 2,183 1,882 40,576

3,050 916 464 538 2,357 967 585 698 5,121 220 1,739 2,002 1,385 463 3,453 1,157 634 582 1,029 1,516 1,522 30,398

2,214 758 373 599 2,099 721 468 572 7686 172 1,327 1,477 1,022 350 2,963 1,015 567 508 931 1,329 1,410 28,561

1,941 64 268 257 1,256 580 289 362 2451 114 986 981 729 242 1,642 594 384 346 632 763 727 16,185

EJISDC (2011) 48, 4, 1-18

10

Table 3. Computer equipment purchases of manufacturing industry (total, and by technological intensity groups) Total manufacturing High technology Low technology Year COMP/X COMP/X COMP/X 0.00782 0.15301 0.06125 2003 0.00858 0.18714 0.05941 2004 0.00736 0.01184 0.00576 2005 0.00812 0.01323 0.00640 2006 COMP: value of computers purchases and X: value of manufacturing exports, in millions of pesos (2003). Source: Own Elaboration Based on AIS 2003-2006 It should be noted that if the use of ICT is concentrated in few industries, will be more difficult for these technologies to generate sustained growth in the economy, while if the use is diversified there is greater potential for growth. When the analysis is centered in purchasing per year, the 12 industries with the highest levels of investment over the period are identified (Table 4). Table 4: Product Classes with Highest Buys of Computer Equipment (2003-2006) Class Description 335991 Electrical products of coal and graphite 334610 Magnetic and optical recording media manufacturing 311613 Sausages and other preserved meat and poultry 336370 Motor vehicle metal stamping 323119 Other Commercial Printing 333299 Machinery and equipment for other manufacturing industries 312141 Manufacturing rum and other distilled beverages of cane 332211 Cutlery and flatware (except precious) manufacturing 327420 Gypsum product manufacturing 333220 Plastics and rubber industry machinery manufacturing 313230 Nonwoven fabric mills 311612 Meat processed from carcasses Source: Own Elaboration Based on AIS 2003-2006 Some stylized facts -that arise from Table 4 and the above data- are: 1) industries with greater use of ICT in management and production processes are: food, chemicals, petroleum products, plastics and rubber, clothing, printing, basic metals, machinery and equipment, equipment of computer, communication and measurement, power generation equipment, electrical equipment and appliances and transport equipment; 2) At the level of product class, who use ICT more are: i) in the food subsector, sausages and preserved, which are within IBT; ii) in printing and related products, sub-sector that more uses ICT in respect to total manufacturing, are production and reproduction of magnetic and optical media, labeled as HTI; iii) in the subsector of machinery and equipment, production of machinery and equipment for the plastics industry, corresponding to HTI; iv) in the transport subsector, production of stamped metal parts for motor vehicles, identified as LTI. Thus, there are certain manufacturing industries that may have a greater impact from the use of ICT. It is expected that the inclusion of computer equipment in the export process by these industries generate a higher level of exports along the time. So far, two things can be

EJISDC (2011) 48, 4, 1-18

11

noted. First, there is an effect of ICT on exports and, second, it seems that the group that benefits most of it is the LTI. 4. EXPORTS AND ICT: AN ECONOMETRIC ANALYSIS Considering the previous sections and since we are interested in determining the effect of ICT on Mexican manufacturing exports to the US, differentiated by industrial technological contents, we analyze this relation based on an econometric model. 4.1 Research Methodology A cross-section econometric analysis for the period 2003-2006 is performed. The aim is to determine the impact of ICT investment on manufacturing exports. It is proposed a model to evaluate the participation of computer equipment purchases -as a proxy for ICT- on exports. The present study considers the ICT as special types of knowledge and technology capital introduced in the production process. Consequently, the regression analysis will be carried on by decomposing the overall effect of total capital to that of its individual physical and ICT components. Additionally, other variables are considered such as the buy of machinery and equipment9 for export and worked hours,10 because along with the computer equipment are complementary production factors that allow accessing successfully to foreign markets. The joint use of these factors and the computer equipment generate increases in production that, in turn, affects the exported level. The traditional way to estimate the export function is the Cobb-Douglas specification, which with three factors of production is defined as: Xd=f(MAQ, POTHH, COMP) or in logarithmic terms: x=α + β1maq + β2pothh + β3comp + u. Where x is natural logarithm of exports of each product class of the manufacturing industry, in thousands of Mexican pesos; maq is natural logarithm of the purchase of machinery and equipment for production by product class of the manufacturing industry, in thousands of pesos; potthh is natural logarithm of hours worked by product class of the manufacturing industry, in thousands of hours; comp is natural logarithm of the purchase of computer equipment by product class of the manufacturing industry, in thousands of pesos. The parameters to be estimated are all in terms of elasticities and the expected signs of the elasticities are all positive. The above equation is estimated following cross-section techniques. These equations can be presented as an econometric model with a general form: yi= α+βXi+Ui, where i are industries (i=1,2,…,174), and error term denotes un-observed individual specific effect. This model considers heterogeneity in error term at industry level. It implies one coefficient for each industry. Intuitively, it is accepted differences among industries are mainly generated by own structural characteristics of each industry. The analysis takes into account observations from 2003 to 2006, but a temporal effect is not estimated because the interest was centered in sectoral effects. Thus, estimations were done by ordinary least squares with cross-section data by industries. This type of methodologies allows to control by potential endogeneity of explicative variable related to ICT (Wooldridge, 2001). Nevertheless, specification showed above was estimated including dummy variables (related to industrial technological levels) to controlling such endogeneity. In none of the estimations the dummy variable was significant. Additionally, to deal with this problem the Hausman (1978) test of endogeneity was performed. Thus, the model is 9

Purchases of machinery and production equipment are increased, through purchases, in the value of machinery and mechanical equipment, electrical, computer or other type directly linked to their production processes and auxiliary tasks. 10 The worked hours are the total hours actually worked by workers and employees. It includes the number of normal and overtime hours actually worked by the remunerated workers and employees, of base and temporary, considering the sub-hired or provided by another corporation.

EJISDC (2011) 48, 4, 1-18

12

estimated obtaining the residuals (μ) which capture the possible endogenous effects within exports and ICT relationship, identifying whether are related to unobservable errors. Then, in order to test endogeneity, it is necessary to estimate the structural form including the residual.11 Since μ are not statistically significant(p-value 0.0781), it is not possible to reject the hypothesis. 5. RESULTS AND DISCUSSION Before analyzing the econometric results it is worthy to say that the association between exports and ICT does not mean that causality runs from one direction. The direction of causation may run either way. To taking account of this matter, we calculate the correlations between ICT investment and manufacturing exports before running the regressions. Thus, Figure 3 shows the relationship between exports and computer equipment; clearly, it is positive. Furthermore, the correlation coefficient is 0.80 which implies a high positive correlation. Consequently, it is assumed that movements in computer buys affect manufacturing exports. The cross-section analysis takes information of all product classes of the manufacturing industry at a particular point in time and examines the variations in information across industries. Thus, we estimated a production function Cobb-Douglas for four points in time 2003 to 2006. The observation units correspond to 174 product classes included in AIS.12 The results are presented in Table 5. The expected results were obtained, in each year there is a positive relationship between exports and spending on computer equipment, the expenditure on machinery and equipment and hours worked also have positive relationship. Considering the value of the parameters, there is a structure where contribution on exports of labor and machinery is larger than the contribution of ICT.

Figure 3. Correlation among exports and computer equipment buys (2006)

lnx

12 10 8 6 4 2 0 5

7

9

11

13

15

17

19

lncomp When analyzing the percentage contribution of each factor to production in 20032006, it seen that the percentage contribution of machinery has increased 40%, the contribution of worked- hours fell by about 20%, while the percentage contribution of ICT doubled even though it registered a drop of 17% from 2005 to 2006. Considering the entire period, the percentage expenditure of manufacturing industries in machinery contributed on 11

The hypothesis of the test are the following H0:λ(ICT,ε)≠0, which implies exogeneity and the alternative hypothesis is H1:λ(ICT,ε)=0, which implies endogeneity. 12 The AIS includes 231 manufacturing industries, however, for estimating purposes, the data were used in logarithms, which led to loss of observation units.

EJISDC (2011) 48, 4, 1-18

13

average to a 0.51 percentage change in X, while the spending in man-hours contributed in 0.36 and the spending in computers in 0.24. Table 5. Cross-Section Estimates of Manufacturing (2003-2006) Coefficients 2003 2004 2005 2006 C 4.169 7.710 6.348 4.82 (8.59) (15.49) (14.73) (10.59) LMAQ 0.315 0.569 0.725 0.443 (7.69) (10.80) (8.40) (6.21) LPOTHH 0.354 0.378 0.387 0.281 (5.74) (5.16) (6.19) (5.17) LCOMP 0.114** 0.136 0.311 0.257 (1.82) (4.47) (5.29) (7.6) 2 R 0.816 0.805 0.811 0.813 N 189 * Variables are significant at 95% confidence, ** Insignificant variables Source: Own Elaboration Based on AIS 2003-2006 On the other hand, to verify whether the technological content of manufacturing is different and to determine whether the impact of ICT use is higher in industries with greater technological content than in the low, estimates were made by industry groups depending on the technological degree. Table 6 presents the results of cross-sectional analysis for each group in the years of study. It is found that computer equipment is significant for high technology sectors in the last three years, which is due to, compared with LTI, they spend more per unit exported. The worked hours are the major contributors to X, which can be explained by the criteria used by Pavitt (1984) to classify industries. This group is the main diffuser of technology in manufacturing, have a high expenditure on R&D, and high engineering capability, this gives great weight to the content of the worked hours which are of highly skilled labor. For lowtechnology sectors in 2003, buys of ICT are not significant at 95% of confidence; however, in subsequent years it became significant and increased its contribution to X, particularly in the last two years. In 2005, for every 1% increase in the buys of computer equipment, X grew 0.58%, while in 2006 it grew at 0.57%. In this group, machinery lost six points of share in the period and worked hours increased their contribution, which is due to the complementarity of computer equipment with worked hours. Additionally, although IBT has relatively less investment, it seems that the IBT uses computer equipment as a tool to increase their efficiency. In these sense, it is likely that computers are used in processes that facilitate the management of industries which is reflected in export shares. On the other hand, to validate the estimates, tests for structural hypotheses were performed. Correlation matrices were obtained, where the correlation between the three factors in some cases exceeded 0.82, indicating a linear association. For this reason it was calculated the square root of coefficient of determination (R2) estimated by OLS, which is higher than the coefficient of correlation between the regressors, indicating absence of multicollinearity. Thus, auxiliary regressions between explanatory variables were estimated and there are not signs of multicollinearity in any of these models. Also, it is analyzed whether the error vector components have equal variance, that is, that homoskedasticity assumption is met. This problem is common in cross-sectional models

EJISDC (2011) 48, 4, 1-18

14

when observations show a heterogeneous behavior. One possibility of the violation of this assumption is a model specification error, as the omission of some relevant variable. In this manner, it is done a graphical verification of the residuals and the White test is applied. In all three cases it is accepted the hypothesis of homoskedasticity.

Table 6. Cross-Section Estimates for High and Low Technology Classes, 2003-2006 High Technology Classes Coefficients 2003 2004 2005 2006 C 4.170 4.556 5.059 4.993 (5.27) (5.23) (5.64) (6.14) LMAQ 0.232 0.159 0.225 0.280 (2.89) (5.18) (4.72) (4.05) LPOTHH 0.718 1.033 0.934 1.033 (2.91) (4.27) (4.08) (4.23) LCOMP 0.089** 0.118 0.212 0.308 (1.68) (2.76) (2.47) (3.44) R2 0.7616 0.7522 0.7716 0.7535 N 25 Low Technology Classes Coefficients 2003 2004 2005 2006 C 6.642 7.007 5.047 6.76284 (13.50) (13.32) (11.94) (12.84) LMAQ 0.759 0.764 0.621 0.696 (12.50) (12.52) (4.72) (6.23) LPOTHH 0.432 0.448 0.624 0.607 (3.28) (3.38) (5.06) (5.72) LCOMP 0.248** 0.313 0.582 0.571 (1.27) (2.49) (3.06) (3.18) 2 R 0.7474 0.7283 0.7042 0.6957 N 149 * variables are significant at 95% of confidence, ** nonsignificant variables Source: own elaboration based on AIS 2003-2006 Thus, it is concluded that ICT, measures as purchase of computer equipment, have a positive elasticity over manufacturing exports as a whole of 0.24%. However, contrary to expectations, exports from HTI are explained in a lesser extent by the purchase of these technologies, which depends more on skilled labor hours included in their processes. In contrast, purchase of ICT has positive effects on LTI exports in 2004-2006 and shows an ICT-elasticity at the end of the period of 0.57%. 6. CONCLUSIONS The rapid diffusion of ICT in various sectors of the economy and the evidence that it boosts the efficiency of several processes, attracted the economic literature for measuring this effect at the aggregate level, however, due to lack of data, the empirical assessment at industrial level is limited. In this sense, this paper contributes to fill this gap for Mexican case and to analyze the effects of changes in the purchase of computer equipment and peripherals (ICT investment) on manufacturing exports in Mexico for the period 2003-2006, distinguishing

EJISDC (2011) 48, 4, 1-18

15

between industries with high and low technology. It is used a cross-sectional methodology to evaluate the effect of such technologies. It is established a production function with three inputs (capital, labor and ICT-computer equipment-) in terms of elasticities. According to this analysis, the response to a change in ITC purchase is quantitatively different for the two types of industries considered. For HTI, the statistical significance of computer equipment is lower at any of the years, although in comparison with LTI, spends more on ICT per unit exported (2.5 times on average for the period). For HTI, which are characterized by a relatively high R&D spending and high engineering capability, the worked-hours are the major contributors to exports because these industries use highly qualified labor. By contrast, for LTI in 2003, ICT investment was not significant, however in subsequent years increased its contribution till reached an elasticity of 0.35%. Also, there is a positive impact of ICT on exports, but the effect is greater in LTI, which seems contrary to economic intuition. Additionally, it is stated that the knowledge and skills that make different the worked hours in each group of industries in HTI sector are more viable explanation for the difference in exports than the use of ICT. In qualitative terms, although LTI invests in ICT to a lesser extent, it seems that they use computer equipment as a tool that increases their efficiency. Computers are used in less complex processes that facilitate the management of industries, reflected in export shares. It can be concluded that investment in ICT, measured as computers, boosts manufacturing exports, particularly for less technology-intensive industries. Finally, it is found, that Mexico has failed to efficiently include these technologies; the workforce remains as a key factor for exports. In this regard, it is important for firms to increase investment in training in order to enhance human capital, which in turn, allows passing from an use of ICT in simple processes, to an integration in a more complex processes and as a consequence, that exports can achieve sustained rates of growth. Compared to previous evidence in developing countries, which is relatively scarce and generally focused in aggregate analysis, it can be stated that the global enhancing effect of ICT on exports of manufacturing industries, estimated in the present study is also found in some other papers. For instance Portugal and Wilson (2010) estimate the impact of ICT infrastructure on the export performance of 101 developing countries over 2004-2007, determining an increasingly relevance especially for richer countries. Nevertheless, the effect of ICT factors on export growth is slower in countries like Chile, Brazil and in lesser extent in Mexico and Argentina for whom the ICT-elasticity of exports oscillates in a range between 0 and 5%. Also, our results are comparable with that of Indjikian and Siegel (2005), who in a quantitative and qualitative research of the impact of IT on economic performance in developed and developing countries, pointed out that improving market access to ICT, export performance of selected developing countries (such as Brazil, Argentina, Chile and Mexico) is equally boosted. Finally, Baliamoune-Lutz (2003), examines links among ICT diffusion and several indicators of economic performance. Indirectly, she determines, using two different measures of trade (openness to international trade and FDI), a positive relation between ICT diffusion and trade. The results are similar to Mexico, Brazil and Chile. In a globalized economy, knowledge and information are factors with increasingly relevance in international global chain. Thus, ICT is a efficiently mechanism to encourage export sector through efficiency gains derived from data processing, provide a range of services and acquisition of parts and components, among others. The relevance of ICT, together other production factor such as knowledge, education and physical capital, on export flows represents useful guidance to policymakers, since establishing policies towards ICT will increase foreign sales.

EJISDC (2011) 48, 4, 1-18

16

7. REFERENCES Ancori, B., Bureth, A. and Cohendet, P. (2000) The Economics of Knowledge: The Debate about Codification and Tacit Knowledge, Industrial and Corporate Change, 9, 2, 255-287. Argandoña, A. (2001) La Nueva Economía y el Crecimiento Económico, División de Investigación, Universidad de Navarra, Documento de Trabajo 437, p. 17. http://www.iese.edu/research/pdfs/DI-0437.pdf Baliamoune-Lutz, M. (2003) An Analysis of the Determinants and Effects of ICT Diffusion in Developing Countries, Information Technology for Development, 10, 3, 151-169. Castells, M. (1999) Information Technology, Globalization and Social Development, United Nations Research Institute for Social Development, Discussion Paper No.114. http://www.rmportal.net/library/content/dp114.pdf Cowan, R., David, P. and Foray, D. (2000) The Explicit Economics of Knowledge Codification and Tacitness, Industrial and Corporate Change, 9, 2, 211-253. Cumbre Mundial sobre la Sociedad de la Información, (2004) Declaración de Rio de Janeiro, http://lac.derechos.apc.org/wsis/cdeclaraciones.shtml Cuthbertson, K., Hall, S. and Taylor, M. (1992) Applied Econometric Techniques, The University of Michigan Press, Dahlman. Dutrénit, G. and Capdeville, M. (1993) El Perfil Tecnológico de la Industria Mexicana y su Dinámica Innovativa en los Ochenta, El Trimestre Económico, 47, 239, 46-68. Edwards, S. (2002) Openness, Productivity and Growth: What Do We Really Know?, The Economic Journal, 108, 447, 383-398. Haag, S., Cummings M., and McCubbrey, D. (2004) Management Information Systems for the Information Age, (4th Edition), New York, McGraw-Hill. Nonaka, I. and Konno, N. (1998) The Concept of “BA”. California Management Review, 40, 3, 40-54 Indjikian, R. and Siegel, D. (2005) The Impact of Investment in IT on Economic Performance: Implications for Developing Countries, World Development, 33, 5, 681-700. Jorgenson, D., Ho, M. and Stiroh, K. (2005) Productivity: Information Technology and the American Growth Resurgence. Cambridge, MA: The MIT Press. Jorgenson D. y Stiroh K. (2000) Information Technology and Growth, American Economic Review, 89, 2, 109-115. Kock, N., McQueen, R. and Corner, J. (1997) The Nature of Data, Information and Knowledge Exchanges in Business Processes: Implications for Process Improvement and Organizational Learning, The Learning Organization, 4, 2, 70-80. Kraemer, K., and Dedrick, J. (1994) Payoffs from Investments in Information Technology: Lessons from the Asia–Pacific Region, World Development, 22, 12, 1921-1931. Machlup, F. (1962) The Production and Distribution of Knowledge in the United States, Princeton University Press. Machlup, F. (1983) Semantic Quirks in Studies of Information, in Machlup F. and Mansfield U., (Eds.), The Study of Information, Interdisciplinary Messages, New York, Wiley. Mum, H., Chin, T., and Choo, A. (2002) Mapping Singapore’s Knowledge-Based Economy, Economic Survey of Singapore, Economics Division, Ministry of Trade and Industry, Third Quarter. Nordhaus, W. (2001) Productivity Growth and the New Economy, National Bureau of Economic Research, Documento de Trabajo W8096. OECD (2002) Reviewing the ICT Sector Definition: Issues for Discussion, Working Party on Indicators for the Information Society, Stockholm.

EJISDC (2011) 48, 4, 1-18

17

Oliner, S. and Sichel, D. (2000) The Resurgence of Growth in the 1990´s: Is Information Technology the Story?, Journal of Economic Perspectives, 14, 4, 33-49. Oulton, N. (2002) ICT and Productivity Growth in the United Kingdom, Oxford Review of Economic Policy, 18, 3, 363-379. Papaioannou, S. and Dimelis, S. (2007) Information Technology as a Factor of Economic Development: Evidence from Developed and Developing Countries, Economic Innovation and New Technology Journal, 16, 3, 179-194. Porat, M. (1977) The Information Economy, Office of Telecommunication, US Department of Commerce, Washington. Working Papers. Stiglitz, J. (2000) The Contributions of the Economics of Information to Twentieth Century Economics, Quarterly Journal of Economics, 115, 1441-1478. Unger, K. (2002) Estructura Industrial y su Evolución: 4 Etapas en Busca de la Integración Industrial y Tecnológica de México, Mimeo, Aula Siglo XXI, España. Wooldridge, J. (2001) Econometric Analysis of Cross Section and Panel Data, Second Edition, The MIT Press.

EJISDC (2011) 48, 4, 1-18 Annex: Product Classes of Mass Media Information Class Description 511110 Newspaper publishers 511120 Periodical publishers 511130 Book publishers* 511140 Directory and mailing list publishers 511191 Greeting card publishers 511199 All other publishers 511210 Software publishers* 512110 Motion picture and video production 512120 Motion picture and video distribution 512131 Motion picture theaters (except drive-ins) 512132 Drive-in motion picture theaters 512191 Teleproduction and other postproduction services 512199 Other motion picture and video industries 512210 Record production 512220 Integrated record production/Distribution 512230 Music publishers 512240 Sound recording studios 512290 Other sound recording industries 515110 Radio broadcasting* 515120 Television broadcasting* 515210 Cable and other subscription programming* 516110 Creation and dissemination of content only via Internet 517111 Fixed traditional telephony 517119 Telegraph and other wired telecommunications 517211 Mobile telephones 517219 Other wireless telecommunications, except satellite services 517310 Resale of telecommunications services 517410 Satellite services 517510 Distribution of television programs by subscription 517910 Other telecommunication services 518110 Internet access providers and search services on the web 518210 Data processing, hosting, and related services 519110 News agencies 519121 Libraries and archives of private sector 519190 All other information services * Except through Internet; Source: North American Industrial Classification System 2002, INEGI

18