Report Study in Impact of ICT use in Business

Description
The report is the outcome of a joint project of UNCTAD with the National Statistical Office (NSO) of Thailand which builds upon the measurement of information economy statistics to enable the assessment of the economic impact of information and communication technology (ICT).

U N I T E D N A T I O N S C O N F E R E N C E O N T R A D E A N D D E V E L O P M E N T
Measuring the impact
of ICT use in business
THE CASE OF MANUFACTURING IN THAILAND
Prepared jointly by the UNCTAD secretariat and the Thailand National Statistical Offce
United Nations Conference on Trade and Development
Measuring the impact of ICT use in
business: the case of manufacturing in
Thailand
Prepared jointly by the UNCTAD secretariat and the
Thailand National Statistical Office
United Nations
New York and Geneva, 2008
Measuring the impact of ICT use in business: Measuring the impact of ICT use in business:
ii ii
Note Note
UNITED NATIONS PUBLICATION
Sales No. E.08.II.D.13
ISBN 978-92-1-112746-1
the case of manufacturing in Thailand
iii
Acknowledgements
This report was prepared jointly by the UNCTAD secretariat, under the supervision of Susan
Teltscher, Chief of the ICT Policy and Analysis Unit, ICT and E-Business Branch, and the
Thailand National Statistical Office, under the supervision of Ruamporn Sirirattrakul, Chief of
the Economic Statistics Analyzing and Forecasting Group, Statistical Forecasting Bureau. The
main contributors were Diana Korka (UNCTAD) and Areerat Kittisomboonsuk (Thai NSO),
who have worked closely together through 2007 to prepare the data sets, design the econometric
model, carry out the analysis and draft the results. The UNCTAD secretariat greatly
acknowledges the making available of statistical microdata by the Thai NSO. The work was
carried out under the overall direction of Anh-Nga Tran-Nguyen, Director, the division for
Services Infrastructure for Development and Trade Efficiency in UNCTAD and in Thailand
under the direction of Jirawan Boonperm, Deputy Secretary General of NSO and Wilas Suwee,
Director of the Statistical Forecasting Bureau.
During their internships with UNCTAD, Lidia Villalba contributed to the statistical analysis of
the 2005 ICT enterprise survey and Sirirat Kiatichaipaibul made useful inputs to the
interpretation of the Thai questionnaire.
Jose Luis Cervera Ferri delivered an econometric modelling training course to the staff of the
Thailand NSO as part of the project and in preparation of the ICT data analysis, and provided
helpful comments on the econometric analysis.
Useful comments were also received from Ugo Panizza, Marco Fugazza, Marcin Skrzypczyk, Albi
Tola, Adam Gross, Oluwatobi Osobukola, Chengetai Masango and staff members of the ICT
and E-Business Branch, in particular Angel Gonzalez-Sanz, Dimo Calovski, Scarlett Fondeur-Gill
and Cécile Barayre El Shami.
v
Contents
Page
Acknowledgements ............................................................................................................................. iii
Executive summary.............................................................................................................................. vii
1. Introduction.............................................................................................................................. 1
Thailand’s ICT strategy and policy................................................................................... 1
Background and objectives of the project ...................................................................... 1
2. Data and statistical methodology........................................................................................... 3
3. Overview of ICT use in business .......................................................................................... 4
General characteristics of the business sector................................................................ 5
Use of computers................................................................................................................ 8
Use of Internet and web presence.................................................................................... 11
Barriers to the use of ICT.................................................................................................. 18
4. ICT use in manufacturing firms ............................................................................................ 20
Use of computers................................................................................................................ 21
Use of Internet and web presence.................................................................................... 21
ICT use and economic performance ............................................................................... 21
5. Measuring ICT impact on labour productivity.................................................................... 23
ICT use and firm labour productivity.............................................................................. 23
Complementary factors explaining the ICT–productivity relationship...................... 24
Impact of specific ICTs on productivity......................................................................... 25
ICT investment, soft technologies and total factor productivity gains ...................... 27
6. Presentation of the model ...................................................................................................... 28
7. Results ....................................................................................................................................... 29
Differences between employment size groups............................................................... 31
Differences between age groups....................................................................................... 32
Regional differences............................................................................................................ 33
Industry differences............................................................................................................ 35
8. Conclusions and policy recommendations .......................................................................... 39
Annex 1. Summary of literature on ICTs and productivity at the firm level .............................. 41
Annex 2. Summary of the variables used in the regression analysis............................................. 43
Annex 3. Correlation coefficients between the regressors used in the analysis.......................... 44
References ....................................................................................................................................... 47
vii
Executive summary
The report is the outcome of a joint project of UNCTAD with the National Statistical Office
(NSO) of Thailand which builds upon the measurement of information economy statistics to
enable the assessment of the economic impact of information and communication
technology (ICT). This is one of the first studies to use official developing-country data to
measure the productivity impact of ICT use in business. The project is part of UNCTAD’s
capacity-building programme on measuring ICT to help developing countries to improve the
production and quality of their ICT statistics at the level of firms through an international
“Partnership on Measuring ICT for Development”. These data and the ensuing analysis on
measuring the economic impact of ICT use aim to provide policymakers with better tools to
design, monitor and evaluate their ICT strategies.
Information and communication technologies have received particular attention in Thailand
as enablers of economic and social development. In the context of the national ICT plan, the
NSO has collected a large amount of data on ICT use through its annual ICT surveys of the
business sector, ICT household surveys and surveys of specific industries such as
manufacturing and services.
This report shows a detailed analysis of trends in ICT use by the Thai business sector by looking
in particular at the use of computers, the Internet and the web. This is done against the
background of a continuous increase in the proportion of businesses using ICTs in Thailand. The
study also reviews the specialized literature estimating the productivity impact of ICT use at the
firm level in a number of developed countries. It then presents the results of the empirical
analysis measuring the impact of ICT use on productivity in manufacturing firms, both at a
general level and also by geographical region, industry branch, firm age and size.
The results indicate that the use of basic ICTs such as computers is important to firm
productivity, particularly in countries where a significant proportion of businesses are still not
using computers. The analysis also finds that, in addition to computer presence, Internet use and
web presence are also reflected in higher labour productivity. The study shows that small and
newly founded manufacturing businesses, especially the ones located in the north and north-east
of the country, should receive more support both in terms of facilitating their access to ICTs and
in terms of information campaigns on how ICTs can help to increase productivity, improve the
quality of products and better respond to demand. Technical information on how businesses
implement ICT solutions can provide additional guidance to set industry-specific ICT strategies.
1
1. Introduction
Thailand’s ICT strategy and policy
For more than a decade, the Government of Thailand has considered information and
communication technology (ICT) an important enabler for economic and social development
and for enhancing the competitiveness of domestic businesses. The establishment of the
National Information Technology Committee in 1992 was one of the first high-level policy
initiatives to promote ICT for development. The committee was chaired by the Prime Minister
and had members from both the public and private sector (Thuvasethakul and Koanantakool,
2002).
Currently, Thailand’s national ICT policy is based on the ICT Master Plan 2002-2006, which is
part of the broader National Information Technology Policy Framework 2001-2010 and the
Ninth National Economics and Social Development Plan. Government agencies, representatives
of the private sector, civil society and academia participated in the debate leading to the adoption
of the ICT Master Plan. Since 2002, the Ministry of ICT has been in charge of pursuing and
implementing the objectives and strategies set out in the Master Plan. The Master Plan had the
general goal of fostering Thailand’s development through ICT and focused on four main
objectives: to increase the country’s economic competitiveness; to develop a knowledge-based
society; to foster sustainable development through equitable access for all; and to develop the
ICT industry (NECTEC et al., 2003). Currently, the Ministry of ICT is in the process of
preparing the second National ICT Master Plan for the next five years (2007-2011).
1
Measuring statistically the access to and use of ICTs has become an important element of the
Thai national ICT policy. Producing ICT indicators is considered key to monitoring and assessing
progress in the implementation of national ICT plans, to compare ICT developments in Thailand
with those in other countries and to help in future policy making (Smutkupt and Pooparadai,
2005). As established in the national ICT plan, the National Statistical Office (NSO) of Thailand
is responsible for producing the necessary ICT data, conducting surveys and carrying out relevant
analysis. The first ICT household survey dates back to 2001. In its 2003 Manufacturing Industry
and Business Services Surveys the NSO included a number of ICT indicators and since 2004, the
NSO carries out an annual stand-alone ICT survey with businesses in Thailand’s municipal areas.
Background and objectives of the project
The collaboration between UNCTAD and the Government of Thailand on ICT-related matters
dates back to 2002, when the Asia-Pacific Regional Conference on “E-commerce Strategies for
Development” was held in Bangkok, under the auspices of the Thai Government. Since then,
UNCTAD has been cooperating closely with the Government in the area of ICT for
development, mainly through NECTEC (National Electronics and Computer Technology
Center).
Cooperation on ICT statistics started in 2004, when UNCTAD began its annual data collection
on ICT in business and the ICT sector. Since then, UNCTAD has been actively assisting
developing countries to improve the production and quality of their ICT data. In the context of
the Partnership on Measuring ICT for Development,
2
which was launched in 2004 at UNCTAD
XI in Brazil, UNCTAD has developed a capacity-building programme on ICT measurement that
1
Policy Statement by Mr. Kraisorn Pornsutee, head of Thai delegation to the ITU Plenipotentiary
Conference in Turkey, 2006,http://www.itu.int/plenipotentiary/2006/statements/thailand/index.html.
2
For more information seehttp://measuring-ict.unctad.org.
Measuring the impact of ICT use in business:
2
includes the delivery of training, courses and workshops, advisory services to countries and the
publishing of a methodological “Manual for the Production of Statistics on the Information
Economy”.
Measuring the impact of ICT using firm-level data has received increasing attention recently,
particularly by NSOs in OECD countries, which have carried out firm-level analyses on the
impact of ICT use on labour productivity using microdata (see section five). Based on the
research approach applied in such studies, UNCTAD in collaboration with the Thai NSO carried
out a research project to measure the impact of ICT use on labour productivity in Thai
manufacturing firms.
The objectives of the project were two-fold:
First, the project aimed to assist the NSO build capacity in the analysis of ICT statistics by
applying econometric modelling techniques. To start the project, in January 2007, UNCTAD
provided a one-week training in Bangkok to staff of the NSO, on applying econometric
techniques to ICT data analysis. This was followed by a period of in-depth data analysis by the
UNCTAD ICT Policy and Analysis Unit, in close cooperation with the Economic Statistics
Analyzing and Forecasting group of the NSO, during which further technical assistance was
provided via long distance. The objective was to allow the NSO staff to replicate the analysis
carried out in UNCTAD with a view to enable the NSO to apply similar analytical methods when
new data become available.
Second, from a substantive point of view, the project aimed to study the link between ICT use in
firms and labour productivity in a developing country setting. While previous studies on the
productivity impact of ICT use in firms have been carried out in developed countries, this is one
of the first comparable analysis based on official statistics from a developing country.
3
Several
recent studies
4
highlight the need to study the way in which ICT use by the business sector
translates into greater economic efficiency. Businesses investing in ICTs do not necessarily
acquire long-term competitive advantage positions, while those not investing in ICT are almost
certain to find themselves at a disadvantage in the market. Ultimately, higher gains from ICT
depend increasingly on identifying the efficient ways of using these technologies. Accordingly, the
Partnership on Measuring ICT for Development encourages research on estimating the
economic impact of ICT, and UNCTAD in particular has engaged in helping developing
countries to use micro data for measuring the link between ICT use by businesses and their
economic performance.
The report is structured as follows. First there is a short presentation of the main data sources
and the methodology used for obtaining ICT statistics. Section three provides an overview of the
use by Thai businesses of basic ICTs, such as computers, Internet and the web, comparing data
from the 2004, 2005 and 2006 ICT Business Surveys. It also includes qualitative information on
factors identified by businesses as barriers to ICT. Section four focuses on ICT use in
manufacturing firms, the sector chosen for the productivity analysis. Section five presents the
literature review and findings of studies on measuring the impact of ICT on labour productivity,
followed by the theoretical framework for this study in section six. A verified empirical model
(Cobb-Douglas production function) was used to quantify empirically the relationship between
ICT uptake and productivity as well as to identify differences based on geographic location,
industry sector, firm size and age in the Thai manufacturing sector. The Report concludes with a
detailed presentation of estimation results and with the formulation of policy recommendations
in sections seven and eight.

3
For example, the Information for Development Programme (InfoDev) has launched in 2006 a project for
measuring the impact of ICT use in Poland, Russia and the Baltic States
(www.insme.info/documenti/InfoDevGlobalNet-web.pdf).
4
For example, Atrostic and Nguyen (2005), Bloom, Sadun and van Reenen (2005) and Farooqui (2005).
the case of manufacturing in Thailand
3
2. Data and statistical methodology
The statistics presented in this report are mainly based on two types of data sources made
available by the NSO of Thailand.
The 2004, 2005 and 2006 ICT Business Surveys were used to produce descriptive statistics on ICT
use in the Thai business sector. The surveys cover a reference period from April to March. Each
of the three survey covers more than 77,000 businesses engaged mainly in manufacturing and
services, including business trade services, construction, land transport and activities of travel
agencies and hospitals. Only firms located in urban (municipal) areas were covered by these
surveys. When information on a given indicator could be traced through the three ICT Business
Surveys, the report showed a comparison between results of the 2004 and the 2006 Survey.
Where data were not available, the report presents only data from the 2005 and the 2006 surveys.
The 2003 Manufacturing Survey was used for carrying out the productivity analysis using
econometric techniques. This dataset contains information on more than 8,800 manufacturing
businesses located both in urban (municipal) and rural (non-municipal) areas, with reference
period January - December 2002. In order to study the relationship between ICT uptake and
economic output it was necessary to link ICT and economic variables, which was only possible
with data contained in this Survey. The econometric study was therefore limited to the 2002
cross-section of Thai manufacturing firms. In the future, when more data will become available, a
more complex analysis can be carried out in order to establish the impact of the past levels of
ICT use on current economic performance.
The ICT statistics used in this research were collected by the NSO of Thailand through a
stratified sampling method that took into account businesses of different employment
5
size,
located in different geographical regions and from different industry branches. This method
made it easy to keep track of businesses with different characteristics and of the share of each
group within the whole business sector. Thus, all the descriptive statistics presented in this report
have been transformed with the help of corresponding sampling weights in order to reflect the
entire population of Thai businesses engaged in either services or manufacturing and with more
than 10 employees. The comparison between the 2005 ICT Business Survey and the 2003
Manufacturing Survey focusing on manufacturing businesses only, also refers to firms located in
urban (municipal) areas.
While both surveys include a large number of micro enterprises (with 10 or less employees), the
data and results presented in this study only include firms with more than 10 employees. This
approach allowed the study to focus on the performance of small, medium sized and large
enterprises, which are producers of most revenues and value added. In Thailand, micro
enterprises represent 94 per cent of all establishments in the market, but they have a small
contribution to revenue, value added and employment. For example, in the manufacturing sector,
firms with more than 10 employees represent only 8 per cent of the firms but account for 77 per
cent of the employment, 98 per cent of the revenues and 96 per cent of value added. Similar
characteristics can be found in other economies as well
6
and therefore usually research is
conducted separately for the two categories of firms. While recognizing that a separate analysis
studying the impact of ICT use on micro enterprises can also yield meaningful results for
policymakers, this was beyond the scope of the project.

5
Employment figures presented in this report refer to all employees, which include both paid and unpaid
labour.
6
For example 92 per cent of all European enterprises have less than 10 employees and they employ 34 per
cent of the workforce. In the US (2000) as well 94.2 per cent of the firms present in the market are micro
enterprises and they provide 21.5 per cent of the jobs (Eurostat, 2003).
Measuring the impact of ICT use in business:
4
3. Overview of ICT use in business
The Thai economy has experienced positive growth since 2001, with increases in both real GDP
and real per capita GDP of respectively 5.6 and 4.8 per cent (CAGR) annually.
7
In 2005, the
structure of the Thai economy expressed in GDP shares is composed of 43 per cent services, 37
per cent manufacturing and 10 per cent agriculture (World Bank, 2006). According to the Thai
2005 ICT Business Survey, 68 per cent of the business sector is made up of manufacturing,
wholesale & retail trade, hotels, restaurants and the food shop industry. After the contraction
suffered in connection to the 1997-1998 regional financial crisis, the volume of sales in the
manufacturing sector rebounded already in 1999, grew slowly in 2000-2001 and registered record
growth rates of 7.5 per cent annually in 2002-2006 (Economist Intelligence Unit, 2007). Strong
domestic demand and a competitive position in the export markets have been important factors
of growth in Thailand. The main manufacturing products exported are vehicles, electronics,
electrical goods and textiles. Output in the electronics sector in particular has recorded 30 per
cent annual average growth rate between 2002 and 2006, with two main types of products:
integrated circuits and hard disk drives (HDDs). Thanks to the success of foreign investment
promotion strategies global producers contribute with a high share to the output in this sector
(UNCTAD, 2005). The computer and related service sector, although increasingly important in
terms of sales, represented in 2005 only 1 per cent of the total number of firms with more than
10 employees.
The use of ICT in Thailand has expanded rapidly. According to NECTEC, the Internet started
being used in 1987. By 1994, all state universities were on-line, and commercial Internet Service
Providers started to operate in 1995 (NECTEC, 2003). Since then, the growth of ICT has been
astounding. The 2005 ICT Business Survey revealed that among Thai businesses with more than
10 employees, 79.7 per cent use computers, 55.4 per cent have access to the Internet and 26.2 per
cent have web presence (table 1).
Despite significant progress over the past years, ICT adoption in Thailand is still behind levels
achieved in some Asian economies with higher per capita income, such as the Republic of Korea,
Singapore and Hong Kong (China). In fact, as shown in table 1, 90.2 to 96.6 per cent of the
enterprises in the three above mentioned economies were using computers and 84.4 to 95.9 per
cent had access to Internet in 2005. Similar to these economies, Thailand has carried out efforts
to produce and disseminate on a regular basis statistics on the use of ICT by businesses. This is
indicative of the relevance placed by Thailand on the implementation and monitoring of ICT
policies.
Table 1. Proportion of businesses with computers, Internet and web
presence in selected Asian economies, 2005
2005
Proportion of
businesses using
computers
Proportion of businesses with
access to Internet
Proportion of businesses with
web presence
China .. 67.6 22.3
Hong Kong (China) 90.2 84.8 40.5
Republic of Korea 96.6 95.9 56.5
Singapore 92.8 91.0 68.3
Thailand 79.7 55.4 26.2
Note: Economies selected based on the availability of official statistics.
Source: UNCTAD information economy database, 2007, and the 2005 ICT Business Survey, businesses with more
than 10 employees.
This section presents detailed information on the use of basic ICTs such as computers, the
Internet and the web by Thai businesses. It draws mainly on data from the 2004, 2005 and 2006
ICT Business Surveys and covers only municipal (urban) enterprises with more than 10
employees, engaged in manufacturing and services activities.

7
Data source UNDESA 2007.
the case of manufacturing in Thailand
5
Over the analysed period of time, the proportion of businesses with computers, access to
Internet and web presence has increased slowly (chart 1). Improvements in the quality and price
of Internet connections resulted in a considerably faster rise in the proportion of businesses with
access to Internet than in the case of the other two technologies. The slightly smaller proportion
of businesses with web presence in 2006 as compared to 2005 is related to a decline of 3 per cent
in the number of large businesses with more than 200 employees present in the Thai market. The
proportion of businesses with computers grew continuously over the three years by 1 per cent
annually, thus increasing in three years by 3,346 the number of businesses with computers. The
highest growth rate is found in the proportion of businesses with access to Internet.
Chart 1. Proportion of businesses with computers, Internet and web presence
in Thailand, 2004-2006
0
10
20
30
40
50
60
70
80
90
P
e
r

c
e
n
t
Proportion of businesses with computers 79.3 79.7 81.4
Proportion of businesses with access to
Internet
46.8 55.4 55.9
Proportion of businesses with web
presence
23.0 26.2 25.7
2004 2005 2006
Source: 2004, 2005 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
General characteristics of the business sector
Small companies (11 to 50 employees) make up for the lion’s share of the Thai market (chart 2),
whereas middle-sized and larger firms represent the remaining one fifth of the businesses. From
2004 to 2006, the number of businesses in Thailand grew moderately from approximately 52,000
to 54,000. 2005 witnessed a significant pick up in the number of medium and large businesses
with more than 50 employees. This tendency was reversed in the next year with a 4 per cent
growth in the number of small businesses and a 3 per cent drop in that of large firms.
Chart 2. Thai businesses by size, 2004-2006
2004
83%
13%
4%
11to 50
51to 200
larger than 200
2006
81%
14%
5%
11to 50
51to 200
larger than 200
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
Measuring the impact of ICT use in business:
6
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
Most enterprises are located in and around the capital region (chart 3). About three quarters of all
municipal (urban) businesses are located in Bangkok, its vicinity and the Central region. These
regions also have the highest concentration of large and middle-sized enterprises. Whereas 73 per
cent of small enterprises are located in Bangkok, its vicinity and the Central region, this was the
case for 77 per cent of the medium enterprises and 83 per cent of the large enterprises. Such
concentration around the main municipal area is common in many economies and goes hand in
hand with the advantages of operating in the main conglomerate. Conglomerates provide
enterprises the possibility to benefit from, for example, cheaper and quicker diffusion of
technology and from access to a greater number of consumers and potential employees. From
2004 to 2006, Bangkok’s vicinity and the Central region showed the greatest increase in the
number of businesses, particularly the large ones. The share of middle and large enterprises also
increased in the northern, north-Eastern and Southern region over the same period of time,
confirming a slight tendency towards a less geographically concentrated business sector.
Chart 3. Thai businesses by geographical region, 2004-2006
Small businesses (11-50 employees)
0 5 000 10 000 15 000 20 000 25 000
Bangkok
Bangkok vicinity
& Central region
Northern region
North-eastern
region
Southern region
2006 2004
Medium businesses (51-200 employees)
0 1 000 2 000 3 000 4 000
Bangkok
Bangkok vicinity
& Central region
Northern region
North-eastern
region
Southern region
2006 2004
Large businesses (more than 200 employees)
0 200 400 600 800 1 000 1 200
Bangkok
Bangkok vicinity
& Central region
Northern region
North-eastern
region
Southern region
2006 2004
the case of manufacturing in Thailand
7
There are four industries, which combined account for more than two thirds of the total number
of businesses in Thailand
8:
manufacturing (29 per cent), retail trade (16 per cent), hotels and
restaurants including food shops (12 per cent), and wholesale and commission trade (11 per
cent). Industries differ in terms of average employment size of enterprises. There are many more
large firms (more than 200 employees) in manufacturing and hospital activities, whereas in retail
trade, wholesale and commission trade, hotels and restaurants small businesses are more
dominant (chart 4).
There has been little variation in the number of businesses by main activity over the three years
considered. The number of small businesses in wholesale trade, commission trade and retail trade
has dropped, along with that of large businesses in the real estate and recreational service sectors.
The manufacturing sector has seen an increase in the number of large enterprises over the three
years. The concentration of large firms in specialized sectors of activity remained the same over
time unlike shown before in the case of the geographical concentration.
Besides employment size, there are also differences regarding employment by gender. Generally,
the representation of female and male workers in services and manufacturing is balanced, with a
ratio of 11 women to 10 men. As in other economies, female participation is larger in hospital
activities (70 per cent) and considerably lower in sectors like construction (6 per cent).
Chart 4. Thai businesses by industry and size, 2004-2006
8
The agricultural sector has not been taken into account in the Survey.
Small businesses (11-50 employees)
0 2 000 4 000 6 000 8 000 10 000 12 000
Manufacturing
Retail trade
Hotels and restaurants, including food shops
Sale, maintenance and repair of motor
vehicles and motorcycles etc.
Wholesale trade and commission trade
Renting of machinery and equipment etc.
Construction
Recreational and other service activities
Other land transport and activities of travel
agents
Real estate
Computer and related activities
Hospital activities
Medium businesses (51-200 employees)
0 500 1 000 1 500 2 000 2 500 3 000 3 500
Manufacturing
Retail trade
Hotels and restaurants, including food shops
Sale, maintenance and repair of motor
vehicles and motorcycles etc.
Wholesale trade and commission trade
Renting of machinery and equipment etc.
Construction
Recreational and other service activities
Other land transport and activities of travel
agents
Real estate
Computer and related activities
Hospital activities
Measuring the impact of ICT use in business:
8
Chart 4. Thai businesses by industry and size, 2004-2006 (continued)
Note: No data available for hospital activities in the 2004 Survey.
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
Based on the 2005 Survey, 9 per cent of the businesses have some level of foreign ownership,
which - as shown later - is a factor linked to higher labour productivity. Foreign ownership also
varies with the type of industry. Whereas hospital activities, recreational and other services firms
are mainly based on domestic capital, the computer and the wholesale and commission trade
sectors have a higher share of foreign capital, in 18 per cent and 12 per cent of the businesses
respectively.
The following section presents detailed information on ICT use in the Thai business sector with
respect to computer and Internet use, and web presence. When the data was available the results
of the 2004 and 2006 ICT Business Surveys are compared.
Use of computers
Computers are an important pre-requisite for the development of the information economy.
Only 1 per cent of the Thai businesses with access to Internet used computers located elsewhere
than on their premises and all enterprises with web presence disposed of at least one computer.
9
The fact that mobile phone operators rely heavily on the GSM technology might explain the very
low proportion of businesses accessing Internet from mobile phones (Economist Intelligence
Unit, 2007). High speed Internet access through mobile technology may become available as
soon as the National Telecommunications Commission starts issuing licences for more advanced
third generation mobile telephony services. Results of the ICT Business Surveys indicate that
only a very small number of businesses, mostly microenterprises, use Internet cafés or telecentre
services. As long as alternative technology is not developed, computers will remain key for the
use of Internet and for web presence in the business sector.
The 2006 ICT Business Survey shows that 81.4 per cent of the Thai businesses have at least one
computer and each of those businesses has on average 14 computers available and 16 employees
9
Based on the 2005 ICT Business Survey.
Large businesses (more than 200 employees)
0 200 400 600 800 1000 1200 1400 1600
Manufacturing
Retail trade
Hotels and restaurants, including f ood
shops
Sale, maintenance and repair of motor
vehicles and motorcycles etc.
Wholesale trade and commission trade
Renting of machinery and equipment etc.
Construction
Recreational and other service activities
Other land transport and activities of
travel agents
Real estate
Computer and related activities
Hospital activities
2006 2004
the case of manufacturing in Thailand
9
using computers regularly. In other words, on average, there are more employees using
computers than available computers. This suggests that Thai businesses have, to a certain extent,
valuable human resources (i.e. computer literate staff) that would allow them to increase the use
of computers. Evidence from developed countries such as Japan (2001)
10
shows that there are on
average more than one computer per employee in the business sector.
Larger firms have more computers and more staff members using computers regularly in their
work (chart 5). For example, a firm with 11-50 employees has on average five computers,
whereas a firm with 501-1000 employees has on average 101 computers. In large firms with more
than 1,000 employees the number of staff members using computers was almost double the size
of the number of available computers in 2004. In 2006 the gap between the two use indicators
has been reduced to some extent. More recently, businesses, in particular the large ones allocate
resources more evenly between computers and the human resources who use them frequently.
Not only have large businesses more computer resources but also there is a large difference in the
proportion of businesses using computers between the different size categories. All businesses
with over 500 employees have at least one computer, while in smaller firms of 11 to 15
employees, this is the case for only three quarters (chart 8).
Chart 5. Average number of computers per business with computers,
2004-2006
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
The 2006 ICT Survey estimated that the business sector uses approximately 574,000 computers,
almost 50 per cent up from the results of the 2004 Survey. Similarly, the number of employees
using computers regularly in their work also increased by 43 per cent, now reaching 700,000. By
industry, the two sectors with the biggest share of middle-sized and larger firms - manufacturing
and hospitals - account for approximately half of the number of computers and computer-using
employees (chart 6). From 2005 to 2006, manufacturing businesses accounted for a slightly
smaller share in the total number of computers and computer-using employees, while other
sectors such as hospitals, wholesale and retail trade, and hotels and restaurants increased their
share.
Sharing computers among several users is easier in certain types of businesses. Computers used
by more than one employee (multi-user computers) are more frequent in hospital activities and
the manufacturing sector. From the 2004 to the 2006 Survey, the number of multi-user
computers also increased by more than 50 per cent, to reach 48,000 units.
10
Ministry of Economy, Trade and Industry, Survey on ICT Workplaces 2001.
2004
5
13
22
49
95
224
5
14
22
49
97
477
0
100
200
300
400
500
11 to 50
employees
51 to 100
employees
101 to 200
employees
201 to 500
employees
501 to
1000
employees
More than
1000
employees
Average number of computers per business with computers
Average number of employees using computers per business with computers
2006
5
17
32
101
338
6
17
33
127
470
66 64
0
100
200
300
400
500
11 to 50
employees
51 to 100
employees
101 to 200
employees
201 to 500
employees
501 to
1000
employees
More than
1000
employees
Average number of computers per business with computers
Average number of employees using computers per business with computers
Measuring the impact of ICT use in business:
10
Source: 2005 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
Chart 6. Share of computers, number of employees using computers and
number of multi-user computers by industry, 2005-2006
Industry share of businesses with computers
0 10 20 30 40 50
Manuf acturing
Hospital activities
Retail trade
Wholesale trade and commission
trade
Renting of machinery and
equipment etc.
Sale, maintenance and repair of
motor vehicles and motorcycles
Hotels and restaurants, including
food shops
Construction
Other land transport and activities
of travel agents
Real estate
Recreational and other service
activities
Computer and related activities
Per cent
Industry share of businesses with
employees using computers
0 5 10 15 20 25 30 35
Manufacturing
Hospital activities
Retail trade
Wholesale trade and commission
trade
Renting of machinery and
equipment etc.
Sale, maintenance and repair of
motor vehicles and motorcycles
Hotels and restaurants, including
f ood shops
Construction
Other land transport and activities
of travel agents
Real estate
Recreational and other service
activities
Computer and related activities
Per cent
If looking at the relative use of computers (chart 7), the computer and related activities together
with the real estate and the wholesale trade and commission trade industry have the largest share
of computers per employee, indicating a higher intensity of computer use in these industries.
Industry share of businesses with multi-user
computers
0 5 10 15 20 25
Manuf acturing
Hospital activities
Retail trade
Wholesale trade and commission
trade
Renting of machinery and equipment
etc.
Sale, maintenance and repair of
motor vehicles and motorcycles etc.
Hotels and restaurants, including
f ood shops
Construction
Other land transport and activities of
travel agents
Real estate
Recreational and other service
activities
Computer and related activities
Per cent
2005
2006
the case of manufacturing in Thailand
11
Chart 7. Intensity of computer use by industry, 2005
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Hotels and restaurants, including food shop
Manufacturing
Renting of machinery and equipment without
operator, research and development etc.
Construction
Recreational and other service activities
Sale, maintenance and repair of motor vehicles and
motorcycles etc.
Retail trade
Other land transport and activities of travel agencies
Hospital Activities
Wholesale trade and commission trade
Real estate activities
Computer and related activities
Number of personal computers per employee Number of multi-user computers per employee
Source: 2005 ICT Business Survey in Thailand, businesses with more than 10 employees.
Use of Internet and web presence
Slightly more than half of the Thai manufacturing and services businesses with more than 10
employees are connected to Internet (55.9 per cent in 2006). Business size matters in terms of
Internet access as it does in terms of computer and web presence use. Chart 8 shows how
computer, Internet and web presence penetration increases with employment size. Whereas 48
per cent of the small businesses have access to Internet, this is true of 79 per cent and
respectively 92 per cent of the medium and large businesses.
As shown before, in an overwhelming majority of cases (99 per cent), businesses access the
Internet from computers located on their premises rather than in Internet cafés or telecentres.
Only 385 businesses access Internet from outside their premises, despite the fact that more than
10,000 firms do not have computers and 23,530 do not have in-house Internet access. The lack
of computers, in small firms in particular, goes hand in hand with the lack of Internet access for
business purposes. For example all the large firms with more than 500 employees have
computers and only 3 per cent of them do not have Internet access. On the other hand, among
small firms with 11 to 15 employees only three quarters have computers, corresponding to 60 per
cent with no Internet access (chart 8). A positive development from 2004 to 2006 has been the
spectacular growth of around 10 per cent in the proportion of small and middle-sized businesses
(11 to 100 employees) with access to Internet. The increase is mainly due to small firms with
computers deciding to connect to the Internet.
Web presence as compared to Internet access is considerably less widespread in the business
sector in Thailand. Little more than a quarter of all businesses have web presence, whether on
their own websites or on sites belonging to other legal entities. Often businesses prefer to host
their own website (in 89 per cent of the cases) rather than having it hosted by a different agent.
Web presence is also more frequent among larger firms. The particularity is that even among the
largest establishments with almost full coverage in terms of computers and Internet access more
than a third remains without web presence. Whereas only a quarter of the small firms are present
on the web, this is true of 42 per cent of the medium and 62 per cent of the large businesses. The
proportion of businesses present on the web has grown mostly among those having 26 to 200
employees - by 5 per cent between 2004 and 2006.
Measuring the impact of ICT use in business:
12
Chart 8. Computers, Internet and web presence by size, 2004-2006
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
In certain economic sectors firms are more prone to using Internet and web applications due to
the inherent characteristics of the business (chart 9). For example, in the computer and related
services sector almost all businesses use computers and the Internet, and more than 80 per cent
have web presence. This confirms findings in other studies (Maliranta and Rouvinen, 2003)
whereby the ICT producing sector itself is a more frequent user of ICTs as compared to other
branches of the economy. There are three other industry sectors where the use of Internet and
web is more than proportional to their use of computers: renting of machinery and equipment,
wholesale trade and commission trade, and other land transport and activities of travel agents.
There was a sizeable increase in the proportion of businesses with Internet access from 2004 to
2006 in the real estate and the sale, maintenance and repair of motor vehicles sectors.
Furthermore, the web penetration rate grew considerably in real estate and computer and related
activities.
The recreational service sector ranks last in chart 9, with the lowest proportion of businesses with
access to any of the three basic ICTs surveyed. However, over time, ICT usage patterns may
change as businesses gain more experience and therefore it is critical to monitor such evolutions.
Although having a sizeable weight (32 per cent) in the total number of computers used in
business, the manufacturing sector recorded a below average penetration of computers, Internet
Proportion of businesses with Internet access
0 20 40 60 80 100
11-15 employees
16-20 employees
21-25 employees
26-30 employees
31-50 employees
51-100 employees
101-200 employees
201-500 employees
501-1000 employees
More than 1000
employees
Per c ent
2006 2004
Proportion of businesses with computers
0 20 40 60 80 100
11-15 employees
16-20 employees
21-25 employees
26-30 employees
31-50 employees
51-100 employees
101-200 employees
201-500 employees
501-1000 employees
More than 1000
employees
Pe r ce nt
2006 2004
Proportion of businesses with web
presence
0 20 40 60 80 100
11-15 employees
16-20 employees
21-25 employees
26-30 employees
31-50 employees
51-100 employees
101-200 employees
201-500 employees
501-1000 employees
More than 1000
employees
Per cent
2006 2004
the case of manufacturing in Thailand
13
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
and web. In 2006, only 58 per cent of the businesses in the manufacturing sector had access to
the Internet and only 25 per cent had web presence, while 78 per cent had at least one computer.
Since 2004 there has been a relatively positive evolution in the proportion of manufacturing
businesses with computers, Internet access and web presence.
Chart 9. Computers, Internet and web presence by industry, 2004-2006
Results of the 2006 Survey show that there are approximately 14,000 businesses present on the
web up from 12,000 in the 2004 survey. Despite the fact that there are fewer businesses with web
presence in manufacturing relative to services, the former sector still makes up for more than a
quarter of the total population of Thai businesses present on the web, followed by wholesale and
retail trade (chart 10). These results are in line with the composition of the Thai business sector
by industry affiliation (as shown previously in chart 4).
Proportion of businesses with web presence
0 20 40 60 80 100 120
Recreational and other service
activities
Hotels and restaurants, including
f ood shops
Manuf acturing
Construction
Retail trade
Other land transport and activities of
travel agents
Sale, maintenance and repair of
motor vehicles and motorcycles etc.
Wholesale trade and commission
trade
Renting of machinery and equipment
etc.
Real estate
Computer and related activities
Hospital activities
Per cent
2006 2004
Proportion of businesses with computers
0 20 40 60 80 100 120
Recreational and other service
activities
Hotels and restaurants, including
f ood shops
Manufacturing
Construction
Retail trade
Other land transport and activities of
travel agents
Sale, maintenance and repair of
motor vehicles and motorcycles etc.
Wholesale trade and commission
trade
Renting of machinery and equipment
etc.
Real estate
Computer and related activities
Hospital activities
Per cent
2006 2004
Proportion of businesses with access to Internet
0 20 40 60 80 100 120
Recreational and other service
activities
Hotels and restaurants, including
food shops
Manufacturing
Construction
Retail trade
Other land transport and activities of
travel agents
Sale, maintenance and repair of
motor vehicles and motorcycles etc.
Wholesale trade and commission
trade
Renting of machinery and equipment
etc.
Real estate
Computer and related activities
Hospital activities
Per cent
2006 2004
Measuring the impact of ICT use in business:
14
Source: 2005 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
Chart 10. Businesses with web presence by industry, 2005-2006
The type of Internet connection plays a crucial role in terms of the quality and speed, which
could impact on competitiveness and productivity at the firm level. In Thailand, a polarisation
seems to have occurred between two groups of firms: a large share of 46 per cent use slower dial-
up connections and a fast growing but lower share of 33 per cent use faster DSL connections.
Wireless connections remain exceptional, used in only 4 per cent of all businesses (chart 11).
Internet broadband connectivity is gaining importance worldwide and has been singled out by the
specialized literature as it can increase the capacity of enterprises to deliver through the Internet
by providing connectivity with higher capacity and speed. In Thailand broadband access almost
doubled from 2004 to 2006, to reach 47 per cent, which compares well with the 63 per cent
broadband penetration in businesses from the European Union. Exceptional cases in terms of
broadband penetration in Asia are the Republic of Korea, where in 2004 92 per cent of
enterprises had Internet broadband access, and Singapore, where in 2005 77 per cent of
enterprises with more than 10 employees had broadband access (IDA, 2005).
Chart 11. Type of connection to Internet in businesses, 2004-2006
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
2006
Dial Line
46%
Unknown
1%
Wireless
connection
4%
Other fixed
connection
1%
Other
1%
Leased Line
7%
Cable Modem
3%
DSL
33%
ISDN
4%
2006
Narrowband
51%
Broadband
47%
Other
broadband
or
narrowband
2%
2004
Narrowband
79%
Broadband
19%
Other
broadband
or
narrowband
2%
2005
Manuf acturing
25%
Wholesale trade
and commission
trade
16%
Retail trade
15%
Computer and
related activities
2%
Real estate
3%
Hospital activities
4%
Other land transport
and activities of
travel agents
3%
Construction
5%
Recreational and
other service
activities
2%
Sale, maintenance
and repair of motor
vehicles and
motorcycles etc.
7%
Renting of
machinery and
equipment etc.
8%
Hotels and
restaurants,
including f ood shops
10%
2006
Wholesale trade
and commission
trade
16%
Retail trade
13%
Manufacturing
27%
Hotels and
restaurants,
including food
shops
10%
Renting of
machinery and
equipment etc.
7%
Sale,
maintenance and
repair of motor
vehicles and
motorcycles etc.
7%
Recreational and
other service
activities
2%
Construction
4%
Other land
transport and
activities of travel
agents
3%
Hospital activities
5%
Real estate
4%
Computer and
related activities
2%
the case of manufacturing in Thailand
15
The type of Internet connectivity gives an additional dimension of the digital divide between the
different economic groups within the country. A large majority of small and medium-sized
enterprises prefer narrowband connections such as dial-up and ISDN (chart 12). In comparison,
a bigger proportion of the large enterprises used broadband leased lines to connect to the
Internet. From 2004 to 2006, there has been an overall shift from dial-up (narrowband) to DSL
(broadband) connections. Also, in medium-sized and large enterprises the proportion of
businesses using leased lines and wireless technology has increased slightly.
Chart 12. Type of connection to Internet by business size, 2004-2006
More than 200 employees
0 20 40 60 80
Dial line
DSL (ADSL, SDSL)
Leased lines
ISDN
Cable modem
Other fixed connection
Wireless connection
Other
Unknown
2006 2004
Source: 2005 ICT Business Survey in Thailand, businesses with more than 10 employees.
Since ICTs provide a wide range of advantages for doing business, such as faster and better
access to information, reductions in costs and time and improved business operations to name
only a few, the ICT Survey also inquired about the motivations for accessing the Internet and for
11 to 50 employees
0 20 40 60 80
Dial line
DSL (ADSL, SDSL)
Leased lines
ISDN
Cable modem
Other fixed connection
Wireless connection
Other
Unknown
2006 2004
51 to 200 employees
0 20 40 60 80
Dial line
DSL (ADSL,
SDSL)
Leased lines
ISDN
Cable modem
Other fixed
connection
Wireless
connection
Other
Unknown
2006 2004
Measuring the impact of ICT use in business:
16
establishing web presence. In the questionnaire, enterprise representatives had the option of
ticking several reasons for using the Internet or the web. For each response category, the
frequency of positive answers indicates the relative importance given by businesses to a specific
activity planned or carried out on the Internet or the web.
Results show that businesses access the Internet mainly for information search, which was ticked
by 50 per cent of the respondents. Sending and receiving e-mail is the second most popular
activity on the Internet (43 per cent), followed by monitoring the market (30 per cent). Although
other online activities such as purchasing and selling products online as well as e-banking and e-
finance are growing in importance in developed countries, businesses in Thailand do not consider
them yet among the main reasons for using Internet. In fact, only 6 per cent of the respondents
indicate that e-banking and e-finance are among the activities planned or already carried out by
their firms on the Internet (chart 13). From 2004 to 2006 there has been an increase in the
relative importance given by business to the three most popular activities mentioned above
(information search, e-mail and monitoring the market). There is scope for improving this part of
the questionnaire in order to define the response categories more precisely and avoid
imprecision.
Chart 13. Reasons for using the Internet (businesses with Internet access),
2004-2006
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
The reasons for establishing web presence are different from those for accessing the Internet.
Chart 14 shows that firms present on the web indicate that they plan to use it mostly for
marketing the products of the business (22 per cent) and as an inquiry and/or contact facility (17
per cent). Those activities would presumably expand the customer base as well as improve the
quality of final products. Other reasons such as receiving orders online (4 per cent) or receiving
online payments (3 per cent) were only pointed out by a very small share of the businesses
present on the web. Over the last surveyed years, there has been a strong increase mainly in the
already popular activities performed by businesses on the web, such as marketing the own
business and using it as an inquiry/contact facility.
0 10 20 30 40 50
Others
Banking and financial services
Other communication channel
Purchase/ sale of goods and services
and communicate with trading partner
Advertising of own goods and services
Monitoring the market
E-mail
Information search
Per cent
2006 2004
the case of manufacturing in Thailand
17
Chart 14. Reasons for web presence (businesses with web presence),
2004-2006
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
The Surveys also tracked information on the number of firms that receive orders online (chart
15). On average, only 7 per cent of the businesses received orders online, with some variation
among branches of activity between 13 per cent for the computer and related services sector and
1 per cent for the recreational and other activities industry. According to the 2006 ICT Survey the
sectors with an above average proportion of businesses receiving orders online belong to the
ICT-producing services industry, the retail and wholesale trade industry and hospital activities
sector. Regarding the hospital industry, the growth of medical and health tourism activities may
explain a relative specialization of businesses in this sector to receive orders online. The slight
changes from 2005 to 2006 indicate that businesses from industries less specialized in receiving
orders online have started to bridge the gap with businesses in the more specialized industries.
The biggest increase is observed for the renting of machinery and equipment sector.
0 5 10 15 20 25
Others
Online payment
Back office network information such as
warehouse logistics
Providing after sales services
Receiving purchased order
Inquiry/ contact facility
Marketing own products
Per cent
2004 2006
Measuring the impact of ICT use in business:
18
Chart 15. Share of businesses receiving orders online, 2005-2006
Source: 2005 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
Despite the relatively high share of Internet users among Thai businesses, e-commerce
11
remains
confined to a small proportion of companies, as seen above in the case of e-selling. The low rate
at which businesses adopt e-commerce is due to the lack of confidence in the quality of services
delivered, and to concerns regarding the sophistication of web technology, the lack of secure
servers and IT skills (NECTEC, 2003).
In order to understand the reasons for engaging in electronic commerce activities, the survey
asked the businesses already engaged in such activities why they did it. Again, the survey provided
respondents with a set of predefined response categories which they could evaluate as important,
very important, not important or unknown. 94 per cent of businesses deemed the use of e-buying
as important or very important for speeding up the business process and simplifying transactions.
In the case of e-selling, the most important reasons for going online were speeding up the
business process and expanding service beyond business hours. Other response categories
considered relevant for selling online were: improving the image of the company (89 per cent),
reducing business costs (87 per cent), expanding the customer base (87 per cent) and keeping up
with the competition (86 per cent). Overall, results show that both buyers and suppliers engage in
e-commerce because this provides them with greater efficiency.
Barriers to the use of ICT
Businesses were asked about the main barriers to the use of specific ICTs. Chart 16 presents a
summary of the responses to three different questions relating to perceived barriers to using
computers, accessing Internet and establishing web presence.
11
Electronic commerce stands for commercial transactions in which the order for a good or service is
made using some form of Internet-based communication. The delivery and payment may be performed
off-line in the physical world.
2006
1
2
3
3
4
4
4
8
9
10
12
13
0 2 4 6 8 10 12 14
Recreational and other service
activities
Real estate
Hotels and restaurants, including food
shops
Other land transport and activities of
travel agents
Renting of machinery and equipment
etc.
Construction
Manufacturing
Hospital activities
Wholesale trade and commission
trade
Sale, maintenance and repair of motor
vehicles and motorcycles etc.
Retail trade
Computer and related activities
Per cent
2005
1
2
3
4
2
3
5
7
8
11
11
13
0 2 4 6 8 10 12 14
Recreational and other service
activities
Real estate
Hotels and restaurants, including food
shops
Other land transport and activities of
travel agents
Renting of machinery and equipment
etc.
Construction
Manufacturing
Hospital activities
Wholesale trade and commission
trade
Sale, maintenance and repair of motor
vehicles and motorcycles etc.
Retail trade
Computer and related activities
Per cent
the case of manufacturing in Thailand
19
Chart 16. Barriers to the use of computers, Internet and web presence
(businesses not using ICTs), 2004-2006
Among businesses that do not currently use any of the technologies captured by the Survey
(computers, the Internet or the web), as much as 58 per cent or more do not perceive the use of
ICT as appropriate or necessary for their business. This problem is more often raised among
firms which do not use computers. From 2004 to 2005 there was a small (1 per cent) drop in the
share of businesses considering computers, Internet or the web as unnecessary for their business
or as having no perceived benefits.
Two factors should be singled out as important barriers: affordability and the lack of skills and
trained staff. High expenditure seems to be a barrier of greater importance in terms of accessing
the Internet or establishing web presence as compared to using computers. More ICT-related
skills and training are needed in firms without Internet connection and web presence (compared
Barriers to computer use
0 20 40 60
Others
Technology changes too fast
Lack of skills or appropriate
training
Lack of perceived benefits
Expenditures too high
Existing personal reluctant to
use or lack of skill
Not appropriate/ unnecessary
business
Per cent
2004 2006
Barriers to the use of Internet
0 20 40 60
No Internet service available in
the area
Others
Security concerns (e.g. fraud,
hacking, viruses)
Lack of skills or appropriate
training
Lack of perceived benefits
Existing personnel reluctant to
use or lack of skill
Expenditures too high
Not appropriate/ unnecessary
business
Per cent
2004 2006
Barriers to the use of web presence
0 20 40 60
Others
Technology changes too fast
Lack of skills or appropriate
training
Lack of perceived benefits
Existing personnel reluctant to
use or lack of skill
Expenditures too high
Not appropriate /unnecessary
business
Per cent
2004 2006
Note: Multiple answers possible.
Source: 2004 and 2006 ICT Business Survey in Thailand, businesses with more than 10 employees.
Measuring the impact of ICT use in business:
20
Source: 2003 Manufacturing Survey and 2005 ICT Business Survey in Thailand, businesses with more than 10 employees.
to using computers). Since the lack of skills has been identified as one of the factors hindering
further ICT development, Thailand has included training and education in information
technology as a main pillar in its national ICT plan.
Results of the survey also show that Internet security and service failures are not the most
important perceived entry barriers for using the Internet. An additional issue of interest may be
to investigate if security concerns represent a major difficulty for firms already connected to the
Internet.
This section of the report has described the main patterns of computer, Internet and website use
by the businesses sector in Thailand. What follows is a closer look at the specific characteristics
of ICT use in manufacturing firms, which is the focus group of the impact analysis carried out
under this project. The next section is based on a comparison of data from the 2003
Manufacturing Survey (reference year 2002) and the 2005 ICT Survey (reference year 1 April
2004 - 30 March 2005).
4. ICT use in manufacturing firms
Manufacturing firms account for a substantial share of the business sector in Thailand (almost a
third), they tend to have a larger than average number of employees per establishment and have
moderate users of ICTs. There is a larger concentration of manufacturers in Bangkok, its vicinity
and the Central region of the country, while services are considerably less concentrated.
In terms of ICT uptake, the manufacturing sector has a below average penetration of computer,
Internet and web presence compared to services. However, due to the considerable size of the
sector, manufacturing businesses still account for almost a third of the number of computers and
that of computer-using employees in the country.
From 2002 to 2005, there has been a positive evolution in all the available indicators of ICT use
in manufacturing. As shown in chart 17, the share of manufacturing businesses using computers
has increased by 4 per cent, from 68.2 per cent to 71.9 per cent. Internet penetration has grown
most (by 12 per cent), followed by the share of firms with web presence (9 per cent). These
figures show a faster evolution of ICT use in 2002-2005 in manufacturing than the current trend
experienced by the entire business sector in 2004-2006 (see chart 17).
Chart 17. Computers, Internet and web presence in the manufacturing sector,
2002-2005
0
10
20
30
40
50
60
70
80
P
e
r

c
e
n
t
Proportion of businesses with
computers
68.2 71.9
Proportion of businesses with access
to Internet
38.9 50.9
Proportion of businesses with web
presence
13.6 22.9
2002 2005
the case of manufacturing in Thailand
21
Use of computers
Computer penetration has reached a high level among manufacturing businesses, with 72 per
cent using at least one computer in 2005.
A firm with 100 employees has on average 12 employees using computers and 11 computers
(table 2). The intensity of computer use has improved during the two years considered, with the
number of computer-using employees remaining slightly superior to that of available computers
per business unit. However, in manufacturing as well as in the rest of the business sector the
number of computers grew faster than the number of employees using computers. Thus figures
show that since 2002 businesses invested more in computers and that the new investment has
resulted in a reduced ratio of employees per computer. More investment in computers has
brought a change in the type of tasks performed with computers, but it is also likely to determine
a growing demand for more computer literate employees.
Table 2. Intensity of computer use in the manufacturing sector,
2002-2005
2002 2005 Growth rate (%)
Average number of employees using computers per 100
employees
9.0 12.2 35.6
Average number of computers per 100 employees
7.8 10.7 37.2
Average number of employees using computers per 100
employees in businesses with computers
14.7 17.6 19.5
Average number of computers per 100 employees in
businesses with computers
11.5 15.3 33.3
Source: 2003 Manufacturing Survey and 2005 ICT Business Survey in Thailand, businesses with more than 10 employees.
Use of Internet and web presence
In 2005 just over half (51 per cent) of the manufacturing businesses had access to the Internet,
compared to 71 per cent of the firms with computers.
Almost a quarter (23 per cent) of the manufacturing firms had web presence in 2004. This figure
corresponded to a 45 per cent proportion of the businesses with computers connected to
Internet and suggests that, unlike Internet, web presence is in an earlier stage of adoption. None
of the manufacturing businesses was found to have web presence in absence of Internet access
and computers on its premises.
The only question on the type of activity carried out over the Internet included in the 2003
Manufacturing Survey relates to e-commerce activities - placing and receiving orders online.
Accordingly, in 2002, 7 per cent of the manufacturers (located in urban areas, with more than 10
employees) sold goods and services online. The proportion of manufacturers which placed orders
online (i.e. buying over the Internet) was much higher - close to 20 per cent
12
in 2002.
ICT use and economic performance
Economic performance is stronger in firms that use ICT and even more so in firms that use a
combination of several ICTs. Manufacturers using computers have on average 10 times higher
sales per employee than manufacturers without computers (chart 18). The order of magnitude is
higher when comparing sales per employee in firms with and without web presence.
In 2002, exporting firms generated three quarters of total sales in the Thai manufacturing sector,
while firms with foreign capital participation accounted for 62 per cent of total industry revenues.
There were very few Thai manufacturing firms without computers that exported or attracted

12
This figure is based on the positive answer to the sub-question regarding activities on Internet (identified
as shopping activity among Internet users in the 2003 manufacturing survey).
Measuring the impact of ICT use in business:
22
foreign capital investment. On average, of 100 exporting Thai manufacturers, 91 had at least one
computer, 77 were connected to the Internet and only 35 had a web presence.
Chart 18. Average sales per employee in the manufacturing sector
($ per employee), 2003
Source: Thai Manufacturing Survey 2003, businesses with more than 10 employees.
Chart 18 illustrates average sales per employee in firms with computers, Internet and web
presence in terms of their export position and foreign capital participation. A typical
manufacturer with at least one computer receives revenues of $ 35’185 per employee yearly. A
total of 33 per cent of businesses with computers are exporters and 19 per cent benefit from
foreign capital participation. This confirms that computers are distributed throughout Thai
manufacturing businesses and are not used exclusively by foreign-owned firms or by exporting
firms.
Among firms with Internet access and web presence, exporting and foreign-owned firms had a
higher participation. A typical manufacturer with at least one computer connected to the Internet
earns on average more than firms with computers - ($ 44,288 per employee); but 48 per cent of
Internet-connected businesses are exporters and 28 per cent receive foreign capital. The shares of
exporting and foreign-owned firms are even higher among web-present manufacturers (61 per
cent and 31 per cent). This suggests that web presence is used more frequently by exporting
firms, possibly for the purpose of following foreign market developments (information search).
Manufacturers with foreign capital participation do not establish web presence as frequently as
exporting firms. Foreign capital participation seems to make more difference in terms of Internet
access and less in terms of web presence, possibly owing to the existence of different
management practices and requirements in foreign firms. The high share of exporters among
manufacturers present on the web could be related to the language used or the relative scarcity of
Internet content in the Thai language.
13
However, available data provided no information on the
language used by firms on websites or about the characteristics of their target customers.
13
The Economist Intelligence Unit (2007) also shows that there are relatively few websites with content in
the Thai language.
Without web presence
18'783
With web presence
69'978
Without Internet
14'744
With Internet
44'228
Without computers
10'097
With computers
35'185
-10%
0%
10%
20%
30%
40%
-5% 5% 15% 25% 35% 45% 55% 65%
Share of exporting firms
S
h
a
r
e

o
f

f
i
r
m
s

w
i
t
h

f
o
r
e
i
g
n

c
a
p
i
t
a
l

p
a
r
t
i
c
i
p
a
t
i
o
n
the case of manufacturing in Thailand
23
So far, the Report has presented an overview of ICT use in the Thai business sector, and in
particular in Thai manufacturing. The next section will turn to the question of how ICT use may
impact on firm productivity.
5. Measuring ICT impact on labour productivity
ICT use and firm labour productivity
At the firm level, analysis assessing the impact of ICT use on productivity can yield a complex set
of results. Variables on ICT use have different productivity effects in conjunction with other
control measures such as firm age, the share of foreign capital participation or access to more
skilled human resources. For example, the estimated impact of a larger proportion of employees
using computers was found by some studies to be more pronounced in young manufacturing
firms (Maliranta and Rouvinen, 2003). Similarly, other studies focus on the specific impact of
ICT in small and medium sized enterprises, as compared to larger businesses. To make it easier
to keep track of the different empirical applications of firm level productivity models, this section
groups the many dimensions into three main categories, by components of the analysis (table 3):
variables measuring labour productivity, ICT variables and complementary factors likely to
influence the ICT-productivity relationship.
Table 3. Key variables for measuring the impact of ICT on labour productivity
Labour productivity variables ICT variables Complementary/control variables
• Sales per employee
• Gross output per employee
• Value added per employee
• Or recalculations of the above
variables based on effective
hours worked by employees
Binary (dummy) variables: take on
value 1 if firm has access to a specific
technology and 0 otherwise.
Numerical variables:
• Spending on specific ICTs
• ICT capital stock
• Share of employees using ICTs
• Number of computers available in
the firm
• Firm age
• Ownership
• Affiliation to a multi-unit firm
• Skill mix (share of employees working
directly in production)
• Level of education
• Industry sector of activity (corresponding to
ISIC codes)
• Geographical region
• Factors of Cobb-Douglas production
functions (ordinary capital stock,
employment, cost of materials)
Labour productivity is commonly measured either as value added per employee or as sales per
employee. Criscuolo and Waldron (2003) derive results for both measures of productivity and
find that the impact of e-commerce
14
was slightly stronger on value added than on sales.
Conversely, Atrostic and Nguyen (2002) rely on the findings of Baily (1986) to argue that using
value added as a measure of labour productivity yields systematically biased estimates of the
theoretically correct growth model. Outside the context of empirical models, value added is a
more precise measure of labour productivity since it subtracts from the value of sales the costs
incurred with intermediate consumption. For those considerations, analysing the impact of ICT
on sales per employee was considered more appropriate for this study.
The empirical models presented here draw on several types of variables for describing the use of
ICT. Binary variables, which distinguish between firms with and without access to, for example,
Internet, are easy to collect and provide input for analyses of differences between the haves and
the have nots. Empirical studies also test for the effect of intensity in ICT uptake – for example,
the share of capital devoted to computer investments – and intensity of ICT use, such as the
number of computers available or the share of employees using e-mail. From a theoretical point
of view, findings based on numerical rather than binary variables are more powerful. Maliranta

14
Measured as placing or receiving orders on line.
Measuring the impact of ICT use in business:
24
and Rouvinen (2006) proposed a slightly different modelling structure in which they estimate the
net effects of several complementary features of computers: processing and storage capacity,
portability and wireless and wireline connectivity. In their model, the positive labour productivity
effect associated to the portability of computers is complementary to that of the basic processing
and storage capacity of any computers whether portable or not.
The different variables of intensity in ICT use by enterprises analysed by the specialized literature
are not the ideal measures. As some have emphasized,
15
ICT use becomes increasingly relevant to
productivity when combined with soft skills such as good management and superior marketing
abilities. Unfortunately, such soft skills and soft technology inputs cannot be quantified directly
and therefore their effect is hard to assess. Empirical research usually corrects for this unknown
effect by accounting for different economic results in foreign-owned firms, in exporting
companies, in establishments belonging to multi-unit corporations or simply in more experienced
firms. Therefore, policy implications derived from such research do not directly recommend that
the intensity of ICT use be scaled up (for example by increasing the number of computers per
employee). Rather they recommend investigating how the combined use of ICT and superior
managerial capabilities can account for variations in ICT gains between firms with different
characteristics.
ICTs can generate higher market shares either by reducing input costs and thus allowing firms to
produce more of the same products, or by improving the quality of products or product
packages, with, as a result, additional sales or higher-priced products. Empirical results presented
here cannot distinguish between those two effects. More information on the evolution of prices
in different sectors is needed in order to assess which effect prevailed in defined periods of time.
In accordance with the framework presented in table 3, the following section shows the most
common complementary factors to the ICT-productivity relationship highlighted by the
specialized literature.
Complementary factors explaining the ICT–productivity relationship
Control variables are additional elements likely to contribute to explaining productivity variation
between firms. They also give a different dimension to results relating to ICT use and
productivity when used in conjunction (interacted) with measures of ICT.
In several studies, firm age has proved to be an important element explaining productivity effects.
The European Commission’s Enterprise and Industry Directorate General showed in a report
(Koellinger, 2006) that the dynamic evolution of new firms is a source of economic growth and
employment and that new firms also contribute significantly to the diffusion of e-business
applications in Europe. In terms of econometric results, Maliranta and Rouvinen (2003) estimate
that young manufacturing firms in Finland, unlike older ones, have 3 per cent higher productivity
gains from the use of computers. Also, young Finnish services firms appeared to be 1 per cent
more productive thanks to access to the Internet.
In a different study, Farooqui (2005) runs four different growth models on young and older
British firms in manufacturing and services taken separately. Results show that ICT indicators
such as investment in IT hardware and software and the share of ICT-equipped employment
have a more pronounced impact on young manufacturing firms as compared with older ones.
The same finding did not apply to young British services companies, however, on that issue,
Atrostic and Nguyen (2005) draw attention to the fact that the measure of capital input used in
most papers – the book value of capital – is a more accurate proxy in the case of the new firms.
Older firms’ capital input is not properly captured by book values because this measure is
evaluated at initial prices when capital assets were acquired as opposed to current asset prices.
The first best proxy to use would be the current value of the capital stock computed by means of

15
For example, Brynjolfsson and Hitt (2002).
the case of manufacturing in Thailand
25
the perpetual inventory method by using information on yearly capital investments, depreciation
and current asset prices. But in many cases, data are not available on all the above-mentioned
variables. Regression results using the book value of capital assets are likely to give biased results
for older firms and more accurate results for younger ones.
Firms with foreign capital participation seemed to have higher labour productivity. With regard
to developed countries, Bloom, Sadun and van Reenen (2005) estimated that in their large sample
of UK firms from all business sectors, US-owned establishments had significantly higher
productivity gains from IT capital than other foreign-owned firms or domestically owned firms.
This result can be linked with macro-level findings which indicated that the United States had
acquired greater labour productivity from investment in ICT than all other developed countries,
especially since the mid-1990s. More productive US-owned firms appear to be better managed or
have access to more efficient ICT solutions.
Similarly, firms belonging to multi-unit networks of affiliates may have greater labour
productivity since they dispose of additional resources to draw from in the subsidiary–
headquarters management structure. A multi-unit corporate configuration may justify benefits
from network effects (a success story replicated in several subsidiary branches) and access to
superior management resources.
The skill mix of production and non-production workers and the level of education in the
regions where companies are located were considered by some studies to be complementary to
the measures of ICT use by firms. Better-skilled workers are more likely to be able to develop,
use and maintain more advanced technology. Maliranta and Rouvinen (2003) comment that
growth models need to control for the human capital characteristics of employment and labour
because these variables are essentially complementary to ICT uptake and omitting them would
inflate the labour productivity gains from ICT.
Last but not least, when quantifying the relationship between ICTs and labour productivity one
needs to control for differences in demand and supply factors. For example, in many countries
businesses located in the vicinity of the capital benefit from higher demand than those located in
isolated provinces simply because there is a high concentration of the population in capitals. In a
similar way, different industries have distinct labour productivity averages owing to both demand
and supply factors. For example, an oil-producing company is very likely to have higher sales per
employee than a light industry manufacturer specialized in food and beverages of the same size
(because of industry characteristics such as price, labour intensity and type of consumer good). It
is therefore necessary to take into account regional and industry-specific characteristics when
accounting for the contribution of ICT to labour productivity growth.
The ICT-producing sector itself benefited from ICT use that considerably exceeded domestic
industry averages. Maliranta and Rouvinen (2003) estimate that in Finland firms belonging to the
ICT-producing sector had 3 to 4.5 per cent higher labour productivity gains from ICT use than
the rest of the manufacturing and services companies in the sample. This may be because ICT
producers have a know-how advantage over other ordinary users in terms of how to best put to
work specific technology to enhance labour productivity.
Impact of specific ICTs on productivity
Use of computer networks (such as the Internet, intranet, LAN, EDI and Extranet) had an
estimated 5 per cent positive effect on labour productivity in a large sample of American
manufacturing businesses (Atrostic and Nguyen, 2002). The model considered a theoretical
framework in which use of computer networks made a “disembodied” contribution to
technological change other than that of capital and labour. Atrostic and Nguyen (2005) take up
again the impact of computer networks in a slightly modified empirical model. The novelty of
their approach consists in using two different computer-related measures in the labour
productivity regression: computer capital, as distinct from ordinary capital, and the computer
network binary variable used previously. In their view, having separate measures for the presence
Measuring the impact of ICT use in business:
26
of computers (computer investment) and for how computers are used (computer networks) is
crucial for estimating accurately the two effects on labour productivity. When using a sample
composed only of newly registered US manufacturing firms, they find that the contribution of
computer networks added 5 per cent to labour productivity while investment in computers added
12 per cent. Within the entire data set of older and younger US manufacturing firms, the
contribution of computer capital dropped to 5 per cent and there was no evidence of a positive
effect on computer networks any more. However, as mentioned before, most empirical studies
tend to find that ICT use has less impact on older manufacturing firms, and this may be due to a
measurement bias as explained in Atrostic and Nguyen (2005).
E-commerce also has a significant impact on labour productivity in firms, with a marked
difference between businesses that buy and those that sell online. Criscuolo and Waldron (2003)
analyse a panel of UK manufacturing firms and find that the positive effect of placing orders
online ranged between 7 and 9 per cent. On the other hand, they estimate that firm labour
productivity was 5 per cent lower for those that used e-commerce for receiving orders online
(online sellers). Lower labour productivity associated with selling products online is likely to be
due to price effects. The prices of products sold online are considerably lower than the prices of
similar goods sold through different channels. Additionally, firms which specialize in selling
online may have difficulties in finding suppliers from which to buy online as much as they would
want. Larger firms with a stronger position in the market may be able to better cope with
balancing the extent of e-buying and e-selling. In a larger and updated UK data set of
manufacturing and services firms, Farooqui (2005) finds again that e-selling negatively impacts on
labour productivity in manufacturing, while e-buying has a larger and positive effect. In
particular, Farooqui (2005) finds that in distribution services e-buying boosts labour productivity
by 4 per cent.
Several studies show that measures of ICT use by employees are also reflected in enhanced firm
productivity. Within a large panel of Finnish firms Maliranta and Rouvinen (2003) compare the
impact of computer use on labour productivity in manufacturing and services sectors. A 10 per
cent increase in the share of computer-equipped labour raises productivity by 1.8 per cent in
manufacturing and 2.8 per cent in services. On the other hand, a higher share of employees with
Internet access was found to have a significant impact only on services firms (2.9 per cent). The
study considers Internet use as a proxy for external electronic communication and LAN use as a
measure of internal electronic communication. Findings show that manufacturing firms benefit
more from better internal communication, while services firms gain more from improved
external electronic communication. A 10 per cent higher share of employees using LAN in the
manufacturing sector results in 2.1 per cent higher labour productivity.
A similar study on a mixed sample of Swedish manufacturing and services firms estimated that a
10 per cent higher share of computer-equipped labour boosts productivity by 1.3 per cent (Hagén
and Zeed, 2005). The Swedish study also estimates productivity effects deriving from access to
broadband of 3.6 per cent. With the help of a composite ICT indicator, Hagén and Zeed (2005)
show that adopting an ever-increasing number of ICT solutions has positive but decreasing
effects on labour productivity in Swedish firms. Each additional level of ICT complexity seems to
add less to firm productivity.
Farooqui (2005) also identifies the use of computers and the Internet by employees as a proxy of
work organization and skills. In the United Kingdom, a 10 per cent increase in the share of
employees using computers raised productivity by 2.1 per cent in manufacturing and 1.5 per cent
in services. These effects are additional to the impact of ICT investment, also accounted for in
the Farooqui (2005) model. Similar estimates for Internet use by British firm employees showed
2.9 per cent for manufacturing and no significant impact for services.
Maliranta and Rouvinen (2006) estimate the impact of different complementary computer
features on labour productivity in a 2001 sample of Finnish services and manufacturing firms.
Their computer variables are measured in terms of share of employees using computers with one
or several of the following features: processing and storage capabilities, portability and wireline or
the case of manufacturing in Thailand
27
wireless connectivity. They find that a 10 per cent higher share of labour with access to basic
computer attributes such as providing processing and storage capabilities increases labour
productivity by 0.9 per cent. In addition, computer portability boosts output per employee by 3.2
per cent, wireline connection to the Internet adds 1.4 per cent, while wireless connectivity adds
only 0.6 per cent.
There is an emerging literature estimating the impact of broadband on firm productivity. Gillett
et al. (2006) were the first to quantify the economic impact of broadband and found that there
were positive and significant effects on the number of workers and the number of businesses in
IT-intensive sectors.
ICT investment, soft technologies and total factor productivity gains
A different empirical question was addressed by Brynjolfsson and Hitt (2002). They explore the
impact of computerization on total factor productivity and output growth in a panel data set of
527 large US firms over a period of eight years (1987-1994). They compare the short-run and
long-run effects of computerization on total factor productivity growth by taking first and five-
to seven-year differences of the log-linear output function. The aim of their analysis is to
understand the mechanism through which private returns from computerization accrue, since
they are the ultimate long-run determinants of decisions to invest in ICTs.
They model production as a function of computer capital and other inputs, and assume that in
the presence of computers, the efficiency of employees, internal firm organization and supply-
chain management systems are improved. They find that, in the short run, investments in
computers generated an increase in labour productivity primarily through capital deepening, and
found little evidence of an impact on total factor productivity growth. However, when the
analysis is based on longer time differences, results show that computers have a positive effect on
total factor productivity growth. In the long run, the contribution of computer capital to growth
rises substantially above computer capital costs, and this is then reflected in terms of total factor
productivity. Their interpretation is that computers create new opportunities for firms to
combine input factors through business reorganization. Brynjolfsson and Yang (1999) had
previously estimated that computer adoption triggers complementary investments in
“organizational capital” up to 10 times as large as direct investments in computers.
To conclude, there is a large variety of results derived for the impact of specific ICTs on firms
from different countries and industries. See annex 1 for a summary of the results captured by the
literature review. A combination of soft technologies and smart ICT use can lead to multifactor
productivity gains but this process takes time and requires complementary spending on other
resources additional to ICT investment. Most measures of ICT use had a positive effect on
productivity across the board - with the exception of exclusively selling online. However, the top
most fortunate firms in terms of gains from Internet use, for example, could have slightly
different characteristics from country to country. They may belong to different industries or they
may have been in the market for a longer time. Therefore such results cannot always be
generalized. Accordingly, it is useful for developing countries to conduct similar empirical studies
to reveal the particularities of the way in which local firms gain from ICTs.
More research, based on developing country data, is needed in order to ascertain how and when
ICT use increased production efficiency in firms and which ICT was used. A comparison of
estimation results across different countries, industry sectors and technologies can provide
policymakers with additional information for fine-tuning ICT policy master plans. As an
illustration, the next section presents the results of a firm-level productivity model applied to the
manufacturing sector in Thailand.
Measuring the impact of ICT use in business:
28
6. Presentation of the model
The primary goal of the analysis was to quantify the relationship between ICT use and labour
productivity in Thai manufacturing firms. In the analysis, which built on methods employed in
similar studies, firm productivity was modelled on the assumptions of a Cobb–Douglas
production function with three input factors: capital, labour and spending on materials (see
box 1).
Box. 1. The empirical model
The regression equation is based on a linearized version of the Cobb–Douglas function
16
(equation 1). Labour productivity is
regressed on factor inputs (capital, labour, spending on materials), one or several ICT variables and a set of controls for
industry and regional attributes of demand and supply, the presence of foreign capital participation and the activity of multi-unit
firms (head offices or branches).
j
i
j i
r
j r j
j j
j
j
j
j
j
j
j
u i Industry r region capital Foreign
unit Multi L
L
M
L
K
e ICTVariabl
L
sales
+ + + +
+ + +

+

+ + =

_ _ _
_ ) ln( ln ln ln
6
5 4 3 2 1 0
? ? ?
? ? ? ? ? ?
,where K is capital, L is employment and M is spending on materials. (equation 1)
With the data available from the Manufacturing Survey 2003 it was possible to run two similar models based on the same
Cobb–Douglas framework: one using total employment as a common denominator and the other using total effective
employment. Effective employment is total employment adjusted to reflect the declared number of hours effectively worked
during 2003. Results derived from the two models could be used as a check for the robustness of estimates. From a theoretical
point of view, the effective labour productivity model is more accurate since it accounts for variations in the hours effectively
worked by employees rather than assuming that all employees worked an equal number of hours. Empirically, the accuracy of
results depends on the quality of data. The model employed for effective labour productivity is described by equation 2.
j
i
j i
r
j r j j
j
j
j
j
j
j
j
j
u i Industry r region capital Foreign unit Multi
effective L
effective L
M
effective L
K
e ICTVariabl
effective L
sales
+ + + + +
+

+

+ + =

_ _ _ _
_ ln(
_
ln
_
ln
_
ln
6 5
4 3 2 1 0
? ? ? ?
? ? ? ? ?
,where K is capital, L_effective is effective employment and M is spending on materials. (equation 2)
White (1980) heteroskedasticity consistent standard deviations were calculated.
When the ICT variable employed is a dummy variable, the 1 estimate is transformed into an elasticity coefficient equal to
1
1
?
?
e
(Halvorsen and Palmquist, 1980). When the ICT variable is expressed as a share in total employment, the 1
estimate is interpreted as a semi-elasticity coefficient in the sense that a unitary increase in the ICT variable is associated with
1 percentage change in the amount of sales per employee.
17
The analysis also identified differences based on geographical location, industry sector, firm size
and age and their influence on the ICT use–productivity relationship. This was done by
estimating interactions of ICT measures with control variables within the same Cobb–Douglas
theoretical framework.
By looking at the performance of firms with and without specific ICTs the analysis could
quantify the extent to which during 2002 Thai manufacturers
18
with similar characteristics had
higher productivity when using ICTs. Further research with comparable data on several years
could investigate the impact of past levels of ICT use on current labour productivity. The analysis
could not establish whether beyond correlation, there is a causal relationship between ICT use

16
See for example Atrostic and Nguyen (2002) for the theoretical derivation of the empirical model.
17
See for example Maliranta and Rouvinen (2003) for the theoretical derivation of the model using the
share of employees using computers.
18
The source of the data is the 2003 Manufacturing Survey.
the case of manufacturing in Thailand
29
and firm labour productivity. To deal with that shortcoming, other specialized papers applied
instrumental variable estimation techniques in data sets covering several years.
19
In the light of the discussion in Atrostic and Nguyen (2002) the preferred dependent variable for
measuring labour productivity was total sales per employee.
A set of additional variables was used in each equation to control for the effect of foreign capital
participation, for the multi-unit organizational aspect and also for unknown disparities in demand
and supply across industries and regions. Information on firm age was used for verifying if
among ICT users more experienced firms had superior labour productivity as compared with
younger firms. Most regression estimates showed that businesses with foreign capital
participation have on average 7 to 8 per cent higher sales. Firms belonging to multi-unit
organizational structures have 2 to 4 per cent higher sales, presumably because they can more
easily gain access to a larger pool of resources. Also, there appeared to be decreasing returns to
scale: larger businesses had on average 0.5 to 3 per cent lower labour productivity given the set of
controls. For a summary of the variables used see annex 2.
Regression results are valid only for the available sample of manufacturing businesses. Sampling
weights corresponding to actual employment size, regions and industries were not used in the
regression analysis and thus regression results cannot be extrapolated to the entire manufacturing
sector.
To address concerns related to collinearity between regressors, annex 3 shows pair wise
correlation coefficients for the first model specifications. They are all below the 0.8 threshold. As
expected, a higher correlation of 0.67 is measured between computer presence and Internet
access. However, when dropping one of the computer or Internet variables, estimated
coefficients remain largely the same.
7. Results
Firstly, the study evaluated the relationship between computer, Internet and web presence and
the value of sales per employee. The three measures of ICT were regarded as progressive steps
adding to the complexity of ICT uptake since all firms using the Internet also had computers on
their premises and all firms present on the web also had Internet access. Accordingly, estimates
on the Internet factor are interpreted as additional to computer-related gains; similarly, web
presence-related gains are complementary to those from Internet and computer use. Results are
shown in table 4.
After controlling for a series of firm-specific economic characteristics, as well as industry and
regional aspects of demand and supply, estimated results showed that firms with a combined use
of computers, the Internet and the web had on average 21 per cent higher sales than firms
without any of the ICTs considered. Among the three ICTs considered, computers contributed
with 14 per cent, Internet access with 3 per cent and web presence with 4 per cent. Similar
estimates were also obtained with the effective labour productivity variant of the model (table 4,
second column). Atrostic and Nguyen (2005) estimated that in 1999 computer networks (such as
the Internet, intranet, LAN, EDI, extranet or other) had a 5 per cent positive impact on labour
productivity in a large sample of United States manufacturing firms. In comparison, the results
derived in this study show that in Thailand computer presence was more closely associated with
labour productivity than Internet connectivity.

19
See for example, Atrostic and Nguyen (2002).
Measuring the impact of ICT use in business:
30
Table 4. Results for computer, the Internet and web presence
OLS (White heteroskedasticity – consistent standard errors)
Independent variables Dependent variable: log (sales per employee) Dependent variable: log (effective sales per employee)
R
2
= 0.923877
Number of observations included: 5 877
R
2
= 0.893852
Number of observations included: 5 651
Computer presence 0.1359*** 0.1346***
Internet access 0.0322* 0.0365*
Web presence 0.0418** 0.0491***
Note: The regression included controls for employment size, capital, costs incurred with materials, foreign capital participation, multi-unit firms and
industry-specific characteristics. (Level of significance at *** 1 per cent, ** 5 per cent and * 10 per cent).
Studies based on developed country data rarely estimate the impact of computer presence on
labour productivity because in many developed countries computer penetration rates in the
business sector have already reached levels of as much as 95 per cent. However, in developing
countries the share of firms that use at least one computer for business purposes has remained
lower (60 per cent in manufacturing Thailand in 2002). Furthermore, data from the 2003
Manufacturing Survey in Thailand show that in absence of computers firms cannot access the
Internet and establish web presence. This explains why in developing countries computer
presence in firms is more closely related to economic performance than in developed countries.
Because the use of at least one computer seemed to account for a large share of the variation in
sales per employee, it was interesting to estimate also the relationship between the intensity of
computer use and labour productivity (tables 4 and 5). Results show that an increase of 10 per
cent in the share of employees using computers is associated with 3.5 per cent higher sales per
employee in Thai manufacturing firms. For the same variable, the estimated coefficient in
Maliranta and Rouvinen (2003) was only 1.8 in a panel of Finnish firms (1998–2000). Since the
variation in the intensity of computer use in Thailand was greater – with many firms not having
computers – computer use was associated with larger differences in labour productivity in that
country. Similarly, a 10 per cent improvement in the number of computers available per
employee was correlated with a 4.5 per cent increase in sales per employee (table 5).
Computer intensity in firms, as captured by the number of physical computers per employee, can
be interpreted as a measure of investment in computer capital. Similarly, the share of employees
using computers also represents a proxy for computer capital investment, including investment in
human capital and training for work with computers. As estimates in tables 5 and 6 show,
Internet access contributed as a significant factor additional to computer intensity in explaining
differences in labour productivity among firms. At the same time, when accounting for the
intensity of computer use, it is noted that web presence is no longer significantly contributing to
higher sales per employee. This suggests that a greater intensity of computer use and Internet
access are factors facilitating the decision to establish web presence in the businesses analysed.
Both the labour productivity and the effective labour productivity models produced similar
estimates for the overall use of computers, the Internet and the web in Thai manufacturing firms.
Table 5. Results for the share of employees using computers, the Internet and
web presence
OLS (White heteroskedasticity – consistent standard errors)
Independent variables Dependent variable: log (sales per employee) Dependent variable: log (effective sales per employee)
R
2
= 0.923260
Number of observations included: 5 863
R
2
= 0.893054
Number of observations included: 5 637
Share of employees
using computers
0.3492*** 0.3969***
Internet access 0.0561*** 0.0582***
Web presence 0.0160 0.0229
Note: The regression included controls for employment size, capital, costs incurred with materials, foreign capital participation, multi-unit firms and
industry-specific characteristics. (Level of significance at *** 1 per cent, ** 5 per cent and * 10 per cent).
the case of manufacturing in Thailand
31
Table 6. Results for the share of computers per employee, the Internet and
web presence
OLS (White heteroskedasticity – consistent standard errors)
Independent variables Dependent variable: log (sales per employee) Dependent variable: log (effective sales per employee)
R
2
= 0.923570
Number of observations included: 5 871
R
2
= 0.893028
Number of observations included: 5 645
Number of computers per
employee
0.4466*** 0.5185***
Internet access 0.0545*** 0.0562**
Web presence 0.0147 0.0213
Note: The regression included controls for employment size, capital, costs incurred with materials, foreign capital participation, multi-unit firms
and industry-specific characteristics. (Level of significance at *** 1 per cent, ** 5 per cent and * 10 per cent).
The study also estimated coefficients relating labour productivity to the different modalities of
accessing the Internet (ISP subscribers, prepaid Internet package, Internet café, etc.), as well as to
the different activities carried out on the Internet (e-mailing, information search, placing orders
online, business promotion, etc.) and on the web (advertising own business, receiving orders
online). However, results did not show significant differences in the way in which those factors
were reflected in the value of sales per employee. More analysis would be needed to assess the
role of the different modalities of Internet access and of the different activities carried out online
in explaining firms’ economic performance. This was beyond the scope of the present analysis.
In the second stage, the analysis aimed at identifying firms’ characteristics that influenced the
relationship between specific ICTs and labour productivity. For that purpose, the regression
equation was slightly modified to estimate the effect
20
of ICT uptake in groups of firms with
different size, age, geographic location and industry branches. The next subsection analyses
whether firm size contributes to explaining the ICT–labour productivity relationship.
Differences between employment size groups
To analyse the implications of firm size in determining the ICT–productivity relationship, three
employment size groups were considered: small firms (11 to 50 employees), medium-sized firms
(51 to 200 employees) and larger firms (more than 200 employees). The groups chosen
correspond to the national classification of Thailand into small, medium and large enterprises.
Table 7 summarizes the results of three different empirical specifications. The estimation model
shown in the first rows took into account information on the presence of computers in firms,
Internet access and web presence. In the second and third models additional information was
included on the intensity of computer use as measured by the share of employees using
computers (the second specification) and the number of computers per employee (the third
specification). Accordingly, results in the first model indicate that computer presence is correlated
with higher productivity in all size groups but more so in large firms (more than 200 employees)
and in small firms (the group with 11 to 50 employees in particular). Results further suggest that
Internet access matters most to labour productivity in small firms, while web presence makes
most difference in large firms. Overall, the link between ICT use and labour productivity is
strongest in large firms. Medium-sized firms (51 to 200 employees) appear not to benefit as much
from Internet access as small firms and also lag behind large firms in terms of gains from web
presence.
The second and third models in table 7 show the estimated relationship between the intensity of
computer use and productivity and confirm the results obtained with the first model. However,
in accounting for the intensity of computer use, web presence no longer had a significant effect.

20
Further econometric tests show that differences between the group estimates are not systematically
different in all cases.
Measuring the impact of ICT use in business:
32
This indicates that in general firms with greater intensity of computer use also had web presence.
In all the size groups most productivity gains from ICT are associated with the intensity of
computer use. A 10 per cent higher share of employees using computers leads on average to an
estimated 4 per cent higher sales in small firms, 1.7 per cent in medium-sized firms and 3.8 per
cent higher sales per employee in large firms. Similarly, 10 per cent more computers per
employee are correlated with 4.5 per cent higher sales per employee in small firms, 2.5 per cent in
medium-sized and respectively 5.2 per cent in large firms. As shown before, access to Internet
seems to make most difference in small firms where it is correlated with 8 per cent higher
productivity, additional to gains from the use of computers. Estimates taking into account the use
of computers indicate again that medium-sized firms did not achieve significantly higher
productivity in correlation with the use of Internet and web presence. Additional analysis showed
that among the group of small businesses, firms with 25 to 50 employees seemed to have higher
gains from ICTs.
Table 7. Results by employment size
OLS (White heteroskedasticity – consistent standard errors)
Dependent variable: log (sales per employee)
Small firms
(11 to 50)
Medium-sized firms
(51 to 200)
Large firms (>200)
Computer presence 0.1365*** 0.1034*** 0.1806***
Internet access 0.0528** 0.0331 -0.0297
Web presence 0.0460 0.0432 0.0455*
R
2
= 0.923573, Number of observations included 5873
Share of employees using computers 0.4014*** 0.1723** 0.3821***
Internet access 0.0809*** 0.0391 0.0082
Web presence 0.0189 0.0310 0.0157
R
2
= 0.922656, Number of observations included 5858
Number of computers per employee 0.4453*** 0.2549** 0.5166***
Internet access 0.0839*** 0.0342 0.0074
Web presence 0.0183 0.0314 0.0136
R
2
= 0.923411, Number of observations included 5867
Note: The regression included controls for employment size, capital, costs incurred with materials, foreign capital
participation, multi-unit firms and industry-specific characteristics. (Level of significance at *** 1 per cent, ** 5 per
cent and * 10 per cent).
Differences between age groups
Several studies reviewed in the previous section found that firm age was an additional factor
explaining how much enterprises gain from ICT. To analyse whether the same effect appeared in
manufacturing Thailand, this study grouped businesses according to their founding year and
hence experience in the market. There are three groups with an equal number of firms: young
(founded between 1997 and 2002), middle-aged (founded between 1991 and 1996) and old
(founded before 1991). The applied estimation technique was the same as in the case of firm size.
Results are shown in table 8 with the first model showing the effect of computer presence,
Internet access and web in the different age groups, while in the second and third models the
analysis also takes into account the intensity of computer use.
In young firms computer presence is associated with the greatest value of gains in terms of sales
per employee, which suggests that young firms use computers more effectively. However, older
firms seem to gain most from the combined use of computers, the Internet and the web. In older
firms the presence of computers also matters, albeit less than in the younger ones, while there is
an additional contribution to productivity from Internet access and the web.
With regard to accounting for the intensity of computer use, results indicate again that for larger
firms, with more experience in the market, there was a stronger correlation between ICT uptake
and labour productivity. Results also show that younger firms with a lower intensity of computer
use seem to achieve higher sales per employee when they have access to the Internet.
the case of manufacturing in Thailand
33
Table 8. Results by firm age
OLS (White heteroskedasticity – consistent standard errors)
Dependent variable: log (sales per employee)
Young firms (1997-2002)
Middle-aged firms (1991-
1996)
Old firms (