Transcript Cross country comparison on the factors determining

Harokopeio University Athens

Innovative model of software development,
where software is open for inspection,
modification and exchange without
restrictions or discrimination.

High diffusion of OSS in different market
segments:
Linux, Android, Apache, Mozilla Firefox,
Open Office, PhP, Java, MySQL, Moodle,
Joomla, Apache OfBiz ...
Limited Knowledge regarding:

OSS diffusion curves

Factors that shape OSS diffusion among
countries

How is OSS diffused among countries?

What theoretical foundation can be used
to identify country level factors that affect
OSS diffusion?

What are the factors that impact the
diffusion process?

What is the impact of these factors at
different stages of the diffusion process?
Diffusion of Innovations Theory
+
Socioeconomic Theories
 Endogenous Growth Theory
 Exogenous Growth Theory
 Institutional Theory

A country is conceptualized as a socioeconomic system within which OSS growth
occurs.

The model is based on the idea that the
forces of growth to an economic system
comprise of institutional, endogenous and
exogenous factors :
OSSit = F(Xendog, Xexog, Xinst)

Economic growth is generated from
within a system as a direct result of
internal processes (Romer, 1986, 1994).
Hypothesis H1: Telecommunications
infrastructure has a positive impact on OSS
growth.
 Infrastructure such as the Internet and
broadband connections are key elements not
only for the diffusion but also for the existence
OSS, as the OSS development model is totally
based on virtual teams and remote
management and collaboration.

Hypothesis H2: Human capital skills and education
can impact OSS diffusion.
 Human capital is a key input to the
development of knowledge, new ideas and
products associated with technological
progress.
 Prior studies demonstrate that the average level
of education and the quality of human capital
are influential drivers for individual technology
adoption.
 Inexperienced and unskilled users would be very
reluctant to migrate to an OSS system, due to
the fear of lack of technological support and
long term.

Hypothesis H3: Innovation can be a ‘pushing’
factor for OSS diffusion.
 The growth of OSS is mainly due to the large
number of skilled and qualified developers
with willingness to create and innovate.
Thus, OSS is not only an innovative
technology, but also is a continuous source
of innovation.
 Prior studies have identified the close
relationship between OSS and innovation.

Hypothesis H4. Economic growth is an impacting
factor for OSS diffusion.
 Previous research has showed that economic
growth is directly related to technological
growth.
 Although OSS can be acquired at low or zero
price, the study tests the possibility that
economic development creates demand for
more technologically advanced products such
as OSS.

Growth is primarily determined by
external factors, such as the flow of
goods, ideas, capital and technology
innovations, rather than internal factors
(Solow, 1965).

Hypothesis H5: Technological openness can
leverage OSS adoption among countries.
 A country’s openness can be defined as the
degree to which a country is open to
business and economic influences through
trade activities. It can be perceived as the
external force that captures knowledge
spillovers among countries .
 The study considers ICT trade as the channel
for achieving such spillovers .

It attends to the deeper understanding
of the processes by which social
structures are maintained and
reproduced (Scott, 2004).
 The quality and operation of institutions may
have a severe impact on the way that social
systems behave towards a technology.
 Governments have the opportunity to
intervene and take actions affecting the
markets and the dissemination of new
technologies.
 Assuming that trade and innovation can
affect OSS diffusion process (H3 and H5), the
effects of government’s regulations and
policies in these sectors are further examined.

Hypothesis H6: regulation in trade could
have a negative impact on OSS adoption
among countries.

Hypothesis H7: regulation in IPR could have
a negative impact on OSS adoption
among countries.
Hypothesis
Endogenous growth theory factors
H1:Telecommunications infrastructure measured by the number of
broadband subscribers . (Source: World Bank Indicators)
H2: Education and Human capital measured by the Human Capital
Index. (Source: United Nations Development Program )
H3: Innovation measured by “charges for the use of intellectual
property” indicators. (Source: World Bank Indicators)
H4: Economic growth measured by GDP per capital.
(Source: World Bank Indicators)
Exogenous growth theory factors
H5: Technological openness measured by ICT trade .
(Source: World Bank Indicators)
Institutional theory factors
H6: Policy regarding trade is measured by the regulatory trade
barriers index . (Source: Economic Freedom Network -Fraser
Institute)
H7: Policy for the protection of property rights is measured by the
Protection of property rights index . (Source: Economic Freedom
Network -Fraser Institute)
Measure Variable
name
Natural
log
Ranged
in [0, 1]
Natural
log
Natural
log
BBS
Natural
log
ICTtrd
HCI
INNOV
GDP
Ranged RTB
in [1,10]
Ranged IPR
in [1,10]

Describes the process of the diffusion of
innovations through a social system,
seeking explanations in terms of “how”
innovations diffuse through it, by the
means of appropriate mathematical
models, the diffusion models.
The Bass diffusion model suggests that a new
product adoption decision is driven by two
factors:
 coefficient of external influence, the influence
that is independent of the existing number of
adopters
 coefficient of internal influence, the social
influence of the existing number of adopters.



Due to the use of time series, a small and a
large country that have exactly the same
number of adopters over time will have the
same diffusion parameters.
The use of fixed time periods for all countries in
a data set creates higher risk of left-hand
truncation bias, where the estimates of the
intercept of the diffusion curve are inflated for
those countries that start their adoption
process earlier than the time frame
represented by the data.
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
ni,t is the number of adopters at time t,
 Ni,t-1 is the cumulative number of
adopters up to time t-1.
 Si is the social system size,
 Ci is the long-run penetration ceiling.
 Bi is the diffusion growth rate.
 Ai1 = ni,1/CiSi is the diffusion curve
intercept.


Logistic models test covariates that affect the first
year adoption (Ai1) and diffusion growth rates (Bi):

d1 and d2 are vectors of parameters and Xi is the
vector of the covariates for country i.
Xi is the vector of the covariates for country i.





Vector of covariates Xi
Xi = (GDP, BBS, HCI, INNOV, ICTtrd, RTB, IPR)
Social system size Si is assumed each country’s
population.
Long-term penetration ceiling Ci is defined as
the proportion of the population aged
between 15 and 65, as this age range should
include the majority of OSS users.
Alignment of introduction timing of OSS: The
model sets time t=1, the first year of adoption
of each country, so that all countries’ diffusion
will set off from a common point of the time
axis.

The analysis is based in data from 25 countries
over the period 2003-2008.

OSS penetration: cumulative number of
subscribed users/developers per country in the
SourceForge.net. Data queried out of the
University of Notre Dame (UND) database
which keeps SourceForge.net data for
academic and scholarly research purposes.

Ci parameter: derived by the World Bank
Indicators (population ages 15-64, % of total).

Si, parameter: extracted by the United Nations
database
Country listings according to:
 Number of per country registered users in the
SourceForge.net in 2009 (I)
 Number of per country registered users, taking
into account the parameter CiSi of each
country in 2009 (II)
 User rates the first year of adoption, Ai1 (III)
 Diffusion Growth rate, Bi (IV)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
(i)
(ii)
(iii)
(iv)
2009
According to CiSi. (2009)
ΗΠΑ
Γερμανία
Ινδία
Καναδάς
Γαλλία
Κίνα
Βραζιλία
Αυστραλία
Ιταλία
Βρετανία
Ολλανδία
Ρωσία
Ισπανία
Ιαπωνία
Βέλγιο
Μεξικό
Τουρκία
Αργεντινή
Ρουμανία
Φιλανδία
Κορέα (Δημ.)
Νορβηγία
Ν. Αφρική
Ελλάδα
Τυνησία
Ολλανδία
Φιλανδία
Νορβηγία
Αυστραλία
Καναδάς
ΗΠΑ
Βέλγιο
Γερμανία
Γαλλία
Ιταλία
Ελλάδα
Βρετανία
Ισπανία
Ρουμανία
Αργεντινή
Βραζιλία
Κορέα (Δημ.)
Ρωσία
Τουρκία
Ν. Αφρική
Τυνησία
Ιαπωνία
Μεξικό
Ινδία
Κίνα
According to Αi,1
(1999)
Νορβηγία
Φιλανδία
Αυστραλία
ΗΠΑ
Καναδάς
Ολλανδία
Γερμανία
Βέλγιο
Βρετανία
Γαλλία
Ισπανία
Ιταλία
Ρουμανία
Ελλάδα
Αργεντινή
Ρωσία
Βραζιλία
Μεξικό
Ιαπωνία
Κίνα
Ινδία
Ν. Αφρική
Κορέα (Δημ.)
Τυνησία
Τουρκία
According to
Βi
Ολλανδία
Γαλλία
Ιταλία
Νορβηγία
Ελλάδα
Φιλανδία
Κορέα (Δημ.)
Γερμανία
ΗΠΑ
Βραζιλία
Αυστραλία
Καναδάς
Βρετανία
Βέλγιο
Αργεντινή
Ρουμανία
Ν. Αφρική
Ρωσία
Τουρκία
Ισπανία
Τυνησία
Ιαπωνία
Ινδία
Μεξικό
Κίνα
First year penetration (Αi,1)
R2 = 0.92
Constructs
BBS
HCI
INNOV
GDP
ICTtrd
RTB
IPR
Diffusion rate (Βi)
R2 =0.87
coefficient
stand error
coefficient
stand error
0.23***
0.03
0.14***
60.059
0.03***
0.016
0.08***
0.035
0.11***
0.025
0.21***
0.051
0.01***
0.007
0.02***
0.025
0.09***
0.029
0.08***
0.029
0.012***
0.006
0.02***
0.015
0.12***
0.038
0.02***
0.011
Significance level * = p < 0.10, ** = p < 0.05, *** = p<0.01

Factors that determine OSS diffusion:
 Innovation
 Quality of human capital
 Telecommunications infrastructure
 ICT trade
 Regulation protecting IPR
OSS evolves in social systems where knowledge,
education and creativity is highly developed.
 Economic growth and less regulation barriers in
trade affect OSS only in the beginning of the
diffusion process.
 Differences in the weight of impact at the two
diffusion stages. Factors related to the quality of
human capital and innovation are more critical
in the growth phase of the diffusion.
 Close relationship between OSS and innovation.
