Software Industri

Published on May 2016 | Categories: Types, Maps | Downloads: 55 | Comments: 0 | Views: 208
of 26
Download PDF   Embed   Report

Journal about software industri

Comments

Content



Globalisation, human capital and technological catch-up

Suma Athreye, Economics, Open University, UK.
John Cantwell, Rutgers Business School, USA.

Paper prepared for the ESRC research seminar series on International trade,
Technological Change and Labour, 5-6 April 2005

Preliminary draft: numbers may change. Please do not cite or quote without
authors’ permission

Abstract:
The interest in this paper is to observe over a long span of time (1950-2001) the
periods of technological catch-up in the sense of new countries contributing to technology
generation in the world economy. We also assess the role of globalisation (through trade,
and inward FDI) and human capital in explaining such technological catch-up. Our
empirical analysis shows that 1950-65 and 1992-2001, were periods of significant
technological catch-up in the world economy. However, despite the catch-up of the Four
dragons, the decades of the 1970s and 1980s were periods of overall technological
concentration when increases in world technology generating capacity came from a small
group of countries that had already begun with significant patent shares. We also find that
trade and inward FDI encouraged catch-up while the increasing concentration of the
world’s human capital tended to increase technological concentration.

1
Globalisation, human capital and technological catch-up

There is considerable debate on the issue of whether new countries in the
developing world are catching-up in technological capabilities and if they can emerge as
significant producers of technology. On one-hand countries like Ireland, Israel and India
have emerged as significant exporters of technologically sophisticated products and
significant multinational R&D in these sectors has moved to these countries. On the other
hand, there is evidence that some countries from Sub-Saharan Africa have shown
technological regress in recent years.
The issue of whether new countries are themselves emerging as generators of
technology is of importance for two kinds of reasons. The first reason, as an empirical
literature on twin peaks and convergence clubs has argued (Quah 1996), is because the
heterogeneity in technological diffusion ultimately determines the evolution of the world
distribution of incomes. However, secondly, the participation of new countries in the
production of technology is also an issue of interest in its own right from the perspective of
the provision of their own development needs. For one thing, there is concern about the
ability of developing countries to develop environmentally friendly technologies and drugs
to combat diseases that disproportionately affect poor country populations like AIDS and
Malaria. Can developing countries produce technologies appropriate for their needs? The
answer to this question could dictate fundamentally the policies that should be adopted. If
technological catch-up is slow it may be socially more useful to find ways to subsidise the
production of such technologies in developed countries. For another thing, the capacity of
developing countries to participate in the higher value creating parts of global production
networks, and thus to catch-up economically with established industrialised countries, both
2
depends upon and is reflected in the emergence of their own indigenous technological
efforts.
Our paper is more concerned with the second reason for studying technological
catch-up in the second sense than the first. The interest in this paper is twofold:
-To observe over a long span of time (1950-2001) the periods of technological
catch-up in the sense of new countries contributing to technological generation.
- Explain the relative importance of globalisation and human capital in influencing
technological catch-up in the world economy
We measure a nation’s contribution to technology generating capacity by observing
the patent shares attributable to the nation in all patents granted in the US. As more
countries start contributing to the world’s technological capacity we should observe a
relative dispersion of the origin of patents across the world in the dataset. We employ a
particular decomposition of the Herfindahl index of patent concentration that allows us to
track the aggregate influence of new patentees in the overall dispersion of patents. This
decomposition demarcates more clearly the characteristics of periods of technological
catch-up and periods of technological concentration.
We then use time series techniques to explain the movement of this catch-up term
due to globalisation and human capital build-up in the world economy. We pay attention to
different dimensions of globalisation in the world economy - openness to trade, share of
international production, growth of international patenting and the growth a and variance in
the stock of human capital.
Our empirical analysis shows that 1950-65 and 1992-2001 were periods of
significant technological catch-up. Despite the catch-up of the Four dragons, the decades of
the 1970s and 1980s were periods of overall technological concentration when increases in
world technology generating capacity came from a small group of countries that had
3
already begun with significant patent shares. In assessing the role of globalisation we find
that openness to trade and inward FDI played an important role in explaining technological
catch-up. However, increased international patenting by multinationals and the increasing
variance of the human capital encouraged the concentration of patents among a few
countries.
The remainder of the paper is organised as follows: A brief review of the literature
on technological catch-up in Section 1, is followed in Section 2 by an outline of the method
employed in our study, including a description of the method used to track technological
catch-up in the world economy. Section 3 describes our main results and Section 4
concludes.
1. A brief review of the literature on technological catch-up
The literature on technological catch-up has developed according to two rather
different traditions. The first is the growth accounting inspired studies of convergence at a
global macro level, and the second has been an almost parallel literature on technology and
development based on the more detailed study of historical episodes and successful
development of technology in new regions of the world. Very good reviews of both
literatures exist and our aim here is to stress the important and complementary conclusions
to which the two literatures come, albeit from different starting points.
In the growth accounting tradition, the rate of growth of technology is seen as the
ultimate factor constraining the long-term economic growth of nations. Early studies in
this tradition introduced the not so intuitive idea that the larger the technological gap the
faster would be the potential catch-up as countries converged to the same level of income
due to the free availability of technology through trade. They also provided considerable
support for the view that the G-8 countries had converged to US levels of income in the
4
post-war period. This convergence was aided by capital flows from the US to Europe and a
favourable trade and aid regime from 1950-65.
However, the view that this convergence in world incomes was general and
extended to all countries soon ran into the sand as it was clear that world incomes were not
converging to any one level. Instead it was pointed out that there were ‘convergence clubs’
of high and low-income countries. These ‘twin peaks’ in the world distribution of income
(Quah 1996) and membership of countries in either the richer or poorer club ultimately
hinged upon the heterogeneity/differences of technology across countries (Bernard and
Jones 1996). In a second development two seminal papers by Romer also introduced the
idea of cumulative causation in growth because of the public good aspect of technology
and the effects of learning on productivity. The important departure of these papers was to
amend neoclassical growth models to make the generation of technology endogenous to the
processes of investment and economic growth, and this provided a perspective on why such
twin peaks might exist at all.
In contrast to the growth accounting models, the literature on technology and
development had always recognised the essential heterogeneity of the technological catch-
up process as well as its endogenous character. Drawing on industrial history these
scholars offered mixed answers to the narrower question of whether new countries could
catch-up and become generators of new technology.
1
Historically, each fresh wave of
technological change did see some new countries catch-up technologically by exploiting
the new opportunities that occasionally emerged in the technological transitions between
waves. Outstanding examples of such a success were the cases of German industrialisation
in the late nineteenth century and later the catch-up of the US. Yet these countries made
significant complementary investments in infrastructure and developed unique institutions

1
For a survey of this literature see Athreye and Simonetti (2004).
5
that facilitated innovation and growth. Firms in these economies had also developed
unique strategies to meet the incumbent competition and exploit the opportunities provided
by the newly emerging electricity technology. Examples include the early R&D facilities
of German chemical firms and their close links with the university sector, and the invention
of joint stock companies in the US to pool financial risks.
The literature on technology and development ascribed the heterogeneity of
technological experiences of countries to two main factors: differences in the technological
capabilities of the firms of nations (see Bell and Pavitt 1997) and differences in the
institutional structures governing innovation by firms and linking them with a variety of
other actors in the economy, which is also sometimes referred to under the heading of
National Systems of Innovation (see Freeman 1997 and a somewhat different take by
Lundvall 1992 ).
Despite significant differences in their conception of technology and the role of
institutions in technological catch-up, both traditions share the importance they ascribe to
human capital and globalisation in the technological catch-up process. The post-war
convergence of incomes and technological catch-up involved a number of countries that
had had strong historical links through the migration of people. In a recent work O’Rourke
and Williamson (19XX) have shown that a large part of the catch-up of European wages to
US levels in the inter-war period was explained by migration and to a lesser extent by
capital flows. The post-war catch-up seemed to reverse this older trend with capital flows
and trade playing a more important role than migration. The technological catch-up of
Japan in the mid-1970s, and the Four Dragons in the late-1980s was also closely associated
with a globalisation of trade and production in the world.
Another important factor emphasised in both literatures is the role of human capital
and training to the technological catch-up process. Studies on the emergence of new
6
science based regions such as those by Bresnahan and Gambardella (2004) and Arora and
Gambardella (2005) also suggest that human capital variations have opened up the
possibility for new regions and nations to occupy distinctive technological niches in a
global market based upon variations in their stock of human capital. Recent examples of
technological catch-up such as those of Israel and Taiwan point to the important role of
openness and human capital investment in creating distinctive comparative advantage
positions for the countries often in global production chains.
2

However, the influences of the two dimensions of globalisation and human capital
on catch-up need more careful empirical study. It is well recognised in the literature on
international trade, foreign investment, human capital and economic growth that there is
considerable interrelation between the three and so disentangling their influence on growth
is very problematic. In this paper we construct a measure of technological catch-up that
does not depend upon growth measures in a direct way. This allows us to bypass some of
the endogeneity issues and examine the impact of the three factors on technological catch-
up and assess the direction of causality.

2. Methodology employed
2.1 Measuring technological catch-up
This paper uses a USPTO patent database to construct an index of technological
catch-up in the world economy. The USPTO database has advantages and disadvantages in
the analysis of technological behaviour and these have been widely discussed in the
literature using patent data.
3
For our purposes a major advantage is that it helps us track

2
These case studies also emphasise the large and coordinated investments by numerous agents in the
economy required to achieve success in technological catch-up and the role of indigenous institutions
in imparting unique advantages to nations. It is beyond the scope of the aggregated level of analysis of
this paper to examine these aspects of technological catch-up, though we think such factors do affect
the inter-country differences in catch-up.
3
See e.g. Schmookler, 1950, 1966, Pavitt, 1985, 1988; Griliches, 1990; Archibugi, 1992.
7
contributions of countries to the world technology generating capacity directly. This
collective innovative capacity is sometimes viewed as representing what is contemporarily
a common world technology frontier, but we follow the modern evolutionary perspective in
supposing that technology is instead developed in an incremental, localised and
differentiated fashion at multiple different sites and following multiple different paths or
approaches to innovation. The US Patent share of countries thus represents an
underestimate of the true technological capacity of countries.
The number of foreign (non-US) countries actively patenting in the dataset rose
slowly from 42 in 1950 to a high of 60 in 1989, although not every country patented every
year. This is far fewer than the total number of countries we were able to collect economic
data for from sources like the Penn Tables and the World Development Indicators. Thus,
like with firms, only a very small proportion of countries patent and demonstrate
technological capabilities. To check on how good a measure of technological ability patent
shares represented as compared to TFP and other estimates we correlated the change in
patent shares with measures for Efficiency Index of countries reported in Russell and
Kumar (2002) for comparable years. The correlation coefficient between the two measures
was about 0.30.
A major drawback of the USPTO dataset however is that the US accounts for a
large proportion of all patents granted, though its own share of patents has been decreasing
over time. Thus, the patent share of the US alone was over 90% in 1950 and fell over time,
but was still high at 55% in 1995. To get a clearer picture about the role of new countries
in patenting, we consider all foreign patents issued by the USPTO – i.e. we exclude US
patents. Appendix 1 describes the main features of the data used in this study.
8
We compute the Herfindahl index of concentration of patent shares across countries
as a summary measure of the uneven technological ability of nations at any point in time.
By definition
H
t
= E S
it
2
, (1)
where S
it
is the share of the ith country in all (foreign) patents issued at time t.
We then exploit a particular decomposition of the Herfindahl index, which splits the
change in overall concentration into a turbulence effect and a regression effect.
4

H
t
= H
t-1
+ AH
t
(2)
Substituting (1) into (2)
AH
t
= E
i
(AS
it
)
2
+ 2 E
i
(S
it-1
AS
it
) (3)
In equation (3), the first term of the RHS measures patent share turbulence (the
concentration of the change in shares). Both positive and negative changes have the same
weight in this index and the larger the value of the turbulence the more changes there will
have been in patent shares. By construction the turbulence measure is always positive.
The second term is however, the more interesting one for tracking technological
catch-up by new countries. It measures the linear association between initial share and the
change in share, weighting large initial shares more than small ones. We call this the
Inverse Regression Effect, since negative values imply a regression of country shares
towards the mean.
5
Negative values of the inverse regression effect come about due to
those that had initially larger patent shares being predominantly also those with negative
values of AS
it
, which occur when these countries lose patent shares. When small patentees

4
For an application of this decomposition to study the evolution of market shares and concentration see
Kambhampati and Kattuman (2003).
5
This very similar to the Galtonian regressions used in Cantwell (1991a), in which the variance of
shares is analogously decomposed into a mobility effect (measured by one minus the correlation
coefficient), and a regression effect (measured by one minus the slope coefficient on lagged shares).
Since H = (V/µ
2
+ 1) / N, where V and µ are respectively the variance and mean of the country shares
and N is the number of countries considered, while in our case the mean share µ = 1/N, it follows that
H = NV + (1/N). Thus, for a given N, H rises with V.
9
have gained or lost patent shares these are given a smaller weight and the cross term will
have a smaller positive value than if the same were to happen to large patentees. As new
countries begin to make small gains in patent shares they erode the shares of existing
nations and tend to cause lower positive values for the turbulence term (turbulence tends to
be greatest when it is the largest countries that make significant gains and losses against
one another, since at that end changes in shares tend to be higher in absolute terms), and a
negative value for the inverse regression index. When some already dominant existing
countries are increasing their patent shares both terms will be positive and higher. We plot
the two terms over time in an exploratory graphical analysis. The results are discussed
Section 3.1 below.
2.2 Explaining technological catch-up
We follow the exploration of the dependent variable with a time series analysis
where the changes in the inverse regression index are explained by measures of
globalisation and human capital in the world economy. The data for these are drawn by
aggregating the data over countries from well-known data sources.
We use three measures of globalisation:
(i) Openness to trade as measured by the ratio of exports and imports to total
world income.
(ii) The share of international production in world income
(iii) The ratio of domestic to international (MNC-owned patents in host
countries) patents in the world economy
We also included two measures of human capital:
(iv) The share of tertiary educated population in the world economy
(v) The variance in the share of tertiary educated population in the world
economy
10
In each case, we aggregated country data to obtain world averages. In order to
control for the effects of the internationalisation of the patent regime we introduced two
new variables the number of patents and the number of countries. The independent
variables used in the study, their data source and expected influence on technological
catch-up in the world economy is summarised in Table 1 below.
[Table 1 here]
Lastly, to assess causality between the IRI and each of the independent variables we
used Granger causality tests. In this exercise we ask the data to predict observed values of
the dependent variable using past lagged values. If the coefficient on the lagged
explanatory variables is significantly different from zero we infer that the explanatory
variable causes movements in the dependent variable. Since the explanatory variables are
all I(1) while IRI is I(0) we use the explanatory variables in their first differences.
3. Empirical analysis
3.1: Assessing periods of technological catch-up
Figure 1 below shows the overall trend in the Herfindahl index of foreign patents
granted by the USPTO. After a long period between 1954-1975 when overall
concentration hovered around 15%, the index rose sharply in the period between 1975-
1992, reaching a value of 28 % in 1992 but it fell again to levels close to 22%. The number
of countries over the entire period rose from about 40 to 60, with the bulk of the increase in
numbers coming in the decade of the 1950s.
6

[Figure 1 here]
Figure 2 below plots the three terms of equation (3). Since patent numbers vary
widely year-on-year they can cause individual patent shares to fluctuate widely. To smooth

6
See Appendix 1 for the number of countries by year.
11
the data for these variations we also plot a four period moving average for the two RHS
terms.
[Figure 2 here]
Figure 2 shows that overall there was relatively little turbulence in the cross-country
distribution, and so the changes in the Herfindahl index were mostly due to changes in the
Inverse Regression Index. This finding is consistent with the view (Cantwell 1991b,
Vertova 1999) that technological advantages of countries are strongly sector-specific and
takes a long time to change.
Through much of the 1950s and 1960s the inverse regression index values were
negative, reflecting a loss of patent shares to new patentees. The negative values were
somewhat larger in the 1950s than in the 1960s, when they hovered between 0 and 0.5%.
The period 1992- 2001 has also been one of catch-up, with smaller patentees gaining patent
share. Again this is consistent with observations of decreasing inequality of world incomes
across countries reported by many studies.
In the intervening period (1972-92) the index turned positive and continued to rise
in value up until the mid-1980s. Thus, for much of the period since the mid-1970s a small
number of countries consolidated their technological positions and accounted for a growing
share of world technology generating capacity. This view of the overall concentration in
technological activity in a few countries from 1975-92 is consistent with the results of a
recent study by Kumar and Russell (2002). Using data envelopment analysis on cross-
country data, that study decomposed labour productivity growth into three components:
technological change (movements of the supposed world frontier), technological catch-up
(movements towards the world frontier) and capital deepening (movement along the
frontier). They found that while technological change contributed positively to growth in
the period 1965-90, the pattern was very dissimilar to overall productivity growth in that
12
there were striking examples of technological regress for low-income countries. Further
they found larger than average contributions to growth for most high-income economies
suggesting technological change benefited the richer countries far more than poorer
countries.
3.2 Explaining movements in the inverse regression index
The order of integration of all the variables is reported in Table 2. The tests
indicate non-stationarity in all the main explanatory variables, but indicated stationarity in
the IRI. We thus used specifications based on first differences of the explanatory variables
explaining the level of IRI.
Table 3 reports some descriptive statistics of the variables we constructed. It is
worth noting the smaller number of observations for inward foreign investment and for
human capital. Table 4 reports the correlation matrix. The globalisation variables are quite
highly correlated with human capital and openness is highly correlated with human capital
and FDI variables.
Table 5 reports the results of the time series estimations. The first four columns
report the influence of the variables by themselves. We find that neither openness, nor
human capital, nor the ratio of domestic to foreign patenting by itself has an effect on
technological catch-up. However, the proportion of international production is by itself
negatively associated with technological catch-up.
7

When we control for the extent of human capital, the influence of both trade and
proportion of inward FDI have negative signs and are significant. (The same results hold
when we use the variance of human capital in the world economy rather than its proportion,
though we do not report those results in Table 5). When we consider the influence of

7
All equations display autocorrelation suggesting the need for additional lagged variables in the
specification. We take this into account when setting up the granger causality tests
13
openness, inward FDI and of human capital, we find that proportion of human capital is not
a significant variable in its own right.
However, if we looked at the variance in human capital we find that it is positively
associated with movements of IRI, when openness and inward FDI are controlled for.
Including the ratio of domestic to foreign patents in the estimation renders only openness
significant. Including inward FDI renders openness insignificant.
These results accord with what is observed in empirical case studies. Though
studies of the Four Dragons and Japan show the role of openness and foreign firms in
technology acquisition and the technological capability building process, the role of human
capital is less clear. Narula and Wakelin (199?) also find that human capital affects export
performance only for the group of very developed countries. Furthermore, our definition
human capital is a more general measure than human capital acquired through training,
which is firm specific and may be expected to raise the productivity of capital employed
within the firm. On the other hand, case studies of science-based industries such as those
contained in Bresnahan and Gambardella (2004) have shown that areas of relative
concentration of human capital attract domestic and foreign science-based firms. Such
firms also tend to be more global in their selling operations.
3.3. Assessing Causality
We performed Granger causality tests to infer the direction of causality between the
variables. Since the first four columns of table 4 showed autocorrelation we included three
lags of the independent variable. The results of the Granger causality tests are reported in
Table 6. The results show that the two aspects of globalisation- international trade and
foreign investment- have a different causal relationship to technological catch-up
(measured by the IRI). Openness to trade granger causes changes in the IRI. From the
regression equation in Table 5 we know that this relationship is negative and so increased
14
trade causes technological catch-up. Neither of the human capital variables appear to cause
catch-up.
However, changes in the IRI granger cause movements in both IFDI and
DOMINT. From Table 5 we know that this relationship is negative, and so we can
conclude that downward movements in IRI (technological catch-up) induce increases in
IFDI. Similarly increases in IRI cause decreases in the DOMINT ratio or conversely
increases in international patenting by MNCs. These findings are consistent with the
observation made by many scholars that inward foreign investment seeks global sources of
competitive advantage and will be drawn to regions of advantage. It is also consistent with
the observation from studies at the firm level (Vuegelars and Cassiman 1998), which have
found that the evidence for foreign firms transferring technology is weak when their
(better) access to technology is controlled for.

4. Summary and conclusions
In this paper we use a patent based measure of technological catch-up in the world
economy to try and assess periods of catch-up as well as assess the factors that seem to
cause changes in catch-up.
We find a mild increase in the concentration of innovation across countries in the
period from 1970-90, and the existence of technological catch-up in the 1950s and 1960s
and again in the period 1992-2001.
As with many empirical studies we find the degree of openness and inward foreign
investment are associated with catch-up. However, causality tests reveal that only
openness to international trade causes technological catch-up. Growth in inward FDI and a
decrease in domestic patenting are caused by technological catch-up.
15
The proportion of human capital in the world economy has no discernible effect on
technological catch-up, but increases in the concentration of human capital in the world
economy are associated with increases in the concentration of technology production.
However, the significance of both these human capital variables vanishes when we control
for the effect of international patenting by multinationals.

References:
Archibugi D. (1992): Patenting as an indicator of technological innovation: a
review, Science and Public Policy, Vol. 19, pp. 357-68.
Athreye, S. and R. Simonetti (2004): “Technology, Investment and Growth” in W.
Brown, S. Bromley and S. Athreye (eds): Ordering the International: History, Change and
Transformation, Pluto Press: London, August 2004.
Bell, M and K Pavitt (1997): “Technological accumulation and industrial growth:
contrasts between developed and developing countries”, in Archibugi, D and J. Mitchie
(eds) Technology, Globalisation and Economic Performance, Cambridge: Cambridge
University Press.
Bernard and Jones (1996): Technology and Convergence, Economic Journal
Vol.106 (6), pp.1037-1043.
Cantwell, J.A. (1991a): The international agglomeration of R&D, in Casson, M.C.
(ed.) Global Research Strategy and International Competitiveness, Oxford: Basil
Blackwell.
Cantwell, J.A. (1991b): Historical trends in international patterns of technological
innovation, in Foreman-Peck, J. (ed.), New Perspectives on the Late Victorian Economy:
Essays in Quantitative Economic History, 1860-1914, Cambridge and New York:
Cambridge University Press.
16
Freeman, C (1997): The ‘National system of innovation’, in Archibugi, D and J.
Mitchie (eds) Technology, Globalisation and Economic Performance, Cambridge:
Cambridge University Press.
Griliches, Z. (1990): Patent statistics as economic indicators: a survey, Journal of
Economic Literature, Vol. 28, pp. 1661-707.
Kumar, S. and Russell, R.R. (2002): Technological change, technological catch-up
and capital deepening: relative contributions to growth and convergence, American
Economic Review, Vol…
Lundvall, B- A. (1992): National Systems of Innovation. London: Pinter.
Nelson, R.R. and Rosenberg, N. (1993): Technical innovation and national systems,
in Nelson, R.R. (ed.), National Innovation Systems: A Comparative Analysis, Oxford and
New York: Oxford University Press.
Pavitt K.L.R. (1985): Patents statistics as indicators of innovative activities:
possibilities and problems, Scientometrics, Vol. 7, pp. 77-99.
Pavitt K.L.R. (1988): Uses and abuses of patent statistics, in van Raan, A.F.J. (ed.),
Handbook of Quantitative Studies of Science and Technology, Amsterdam: Elsevier
Science Publishers.
Quah, D. (1996): Twin peaks: Growth and convergence in models of distribution
dynamics, Economic Journal Vol.106 (6), pp.1045-1055.
Schmookler J. (1950): The interpretation of patent statistics, Journal of the Patent
Office Society, Vol. 32, pp. 123-146.
Schmookler J. (1966): Inventions and Economic Growth, Cambridge, Mass.:
Harvard University Press.
17
18
Vertova, G. (1999): Stability in national patterns of technological specialisation:
some historical evidence from patent data, Economics of Innovation and New Technology,
Vol. 8, pp. 331-354.
TABLES
Table 1: Explanatory Variables used in the econometric analysis

Variable Description Data Source Span Expected
sign
OPEN
(Import + Export) / real
GDP (1996 constant)
Unit: %
Penn Data
http://pwt.econ.upenn.edu/

1950-
2000
-
IFDI Inward FDI Stock / GDP,
Unit: %
World Investment Report
2004 (UNCTAD)
1980-
2000
-
DOMINT Ratio of domestic firm
patents to MNC patents
in host countries
USPTO data 1950-95 +
HUMCAP Tertiary enrolment /
Population
Unit: %
World Development
Indicators 2002
(World Bank)
1970-
2000
-
HCAPCV Coefficient of variation
of human capital
Unit: %
World Development
Indicators 2002 CD-Rom
(World Bank)
1980-
2000
+
Control
variables

PATENTS Total number of patents
in USPTO
USPTO data 1950-
2001

NUMBER Total number of
countries patenting
USPTO data 1950-
2001

Table 2: Unit Root Tests
Variable Order of Integration Data Span
HERF I(1) 1950-2001
PATENTS I(1) 1950-2001
OPEN I(1) 1950-2000
HUMCAP I(1) 1970-2000
DOMINT I(1) 1950-2001
HUMCAPCV I(1) 1970-2000
NUMBER I(1) 1950-2001
IFDI I(2) 1980-2000
Note: All tests are significant at the 5% level of significance
T
a
b
l
e

3
:

D
e
s
c
r
i
p
t
i
v
e

s
t
a
t
i
s
t
i
c
s


H
E
R
F







M
O
B
2
O
P
E
N
I
F
D
I
D
O
M
I
N
T
H
U
M
C
A
P
H
U
M
C
A
P
C
V

M
e
a
n

0
.
1
9

0
.
0
0
0
.
2
9
9
.
9
6
2
3
.
2
9
1
3
.
0
4
1
.
1
3

M
e
d
i
a
n

0
.
1
6

0
.
0
0
0
.
2
8
9
.
0
9
1
7
.
3
1
1
3
.
0
2
0
.
9
3

M
a
x
i
m
u
m

0
.
2
8

0
.
0
2
0
.
5
2
1
9
.
3
1
6
1
.
0
8
1
9
.
2
9
1
.
7
7

M
i
n
i
m
u
m

0
.
1
4

-
0
.
0
2
0
.
1
6
6
.
6
0
6
.
8
0
7
.
8
1
0
.
8
5

S
t
d
.

D
e
v
.

0
.
0
5

0
.
0
1
0
.
0
9
3
.
2
4
1
6
.
3
3
3
.
7
2
0
.
3
3

S
k
e
w
n
e
s
s

0
.
6
8

-
0
.
2
6
0
.
7
7
1
.
6
5
0
.
7
0
-
0
.
0
3
1
.
1
2

O
b
s
e
r
v
a
t
i
o
n
s

5
2

5
2
5
1
2
1
5
2
2
8
2
9
T
a
b
l
e

4
:

C
r
o
s
s

c
o
r
r
e
l
a
t
i
o
n

m
a
t
r
i
x




D
O
M
I
N
T

I
F
D
I

O
P
E
N
N
E
S
S

H
U
M
A
N
C
A
P

H
U
M
A
N
C
A
P
C
V
D
O
M
I
N
T


1
.
0
0

-
0
.
7
4

-
0
.
6
3

-
0
.
7
4


0
.
5
3

I
F
D
I

-
0
.
7
4


1
.
0
0


0
.
9
3


0
.
9
4

-
0
.
7
6

O
P
E
N
N
E
S
S

-
0
.
6
3


0
.
9
3


1
.
0
0


0
.
9
8

-
0
.
8
8

H
U
M
A
N
C
A
P

-
0
.
7
4


0
.
9
4


0
.
9
8


1
.
0
0

-
0
.
8
6

H
U
M
A
N
C
A
P
C
V

0
.
5
3

-
0
.
7
6

-
0
.
8
8

-
0
.
8
6


1
.
0
0


T
a
b
l
e

5
:

D
e
t
e
r
m
i
n
a
n
t
s

o
f

c
h
a
n
g
e

i
n

t
h
e

i
n
v
e
r
s
e

r
e
g
r
e
s
s
i
o
n

i
n
d
e
x


(
1
)









(
2
)
(
3
)
(
4
)
(
5
)
(
6
)
(
7
)
(
8
)
(
9
)
C
O
N
S
T
A
N
T










0
.
0
0
1
0
.
0
0
6
*
0
.
0
0
4
-
0
.
0
0
0
.
0
0
7
*
0
.
0
0
7
*
*
0
.
0
1
0
*
*
*
0
.
0
1
1
*
*
*
0
.
0
0
6
*
*
*
O
P
E
N
N
E
S
S





















































































-
0
.
1
2
2

-
0
.
4
1
2
*
*



-
0
.
4
5
3
*
*

-
0
.
4
3
7
*
*

-
0
.
3
7
1
*
*


I
F
D
I
-
0
.
0
0
8
*
*
*
-
0
.
0
0
7
*
-
0
.
0
0
7
*

-
0
.
0
0
8
*
*
*
D
O
M
I
N
T
-
0
.
0
0
-
0
.
0
0
1

H
U
M
A
N

C
A
P
I
T
A
L

-
0
.
0
0
1
0
.
0
0
8
-
0
.
0
0
4
0
.
0
0
0
H
C
A
P
C
V
0
.
2
1
9
*
0
.
0
0
9
N
U
M
B
E
R
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
-
0
.
0
0
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
0
.
0
0
1
0
.
0
0
0
P
A
T
E
N
T
S
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
A
u
t
o
c
o
r
r
e
l
a
t
i
o
n

Y
e
s
N
o
Y
e
s
Y
e
s
N
o
N
o
N
o
N
o
N
o
R
-
s
q
u
a
r
e
d
0
.
0
6
5
0
.
4
4
9
0
.
0
6
1
0
.
0
6
1
0
.
3
6
6
0
.
3
1
0
0
.
5
6
8
0
.
6
7
2
0
.
3
5
7
o
b
s
e
r
v
a
t
i
o
n
s
5
0
2
0
2
7
5
1
2
7
1
7
1
7
1
8
2
8


Table 5: Tests of Granger causality for pairs of variables (lags included=3)

Variable pair Null Hypothesis Obs F-statistic Probability
(IRI, ǻOPEN) ǻOPEN does not
Granger Cause IRI
47 2.926 0.045
IRI does not Granger
Cause ǻOPEN
0.364 0.779
(IRI, ǻFDI) ǻIFDI does not Granger
Cause IRI
17 2.629 0.108
IRI does not Granger
Cause ǻ IFDI
4.040 0.040
(IRI, ǻDOMINT) ǻ(DOMINT) does not
Granger Cause IRI
48 0.415 0.74
IRI does not Granger
Cause ǻ (DOMINT)
2.640 0.062
(IRI, ǻHUMCAP) ǻHUMCAP does not
Granger Cause IRI
24 0.658 0.589
IRI does not Granger
Cause ǻHUMCAP
0.352 0.788
(IRI, ǻHUMCAPCV) ǻHUMCAPCV does not
Granger Cause IRI
25 0.863 0.47
IRI does not Granger
Cause ǻHUMCAPCV
0.265 0.850



A
p
p
e
n
d
i
x

1
:

D
e
s
c
r
i
p
t
i
o
n

o
f

t
h
e

d
a
t
a
s
e
t

T
h
e

U
S
P
T
O

d
a
t
a

s
e
t

w
e

h
a
v
e

u
s
e
d

h
e
r
e

i
s

s
i
m
i
l
a
r

t
h
a
t

u
s
e
d

i
n

C
a
n
t
w
e
l
l

(
1
9
9
1
a
)
.


I
t

c
o
n
s
i
s
t
s

o
f

t
h
e

n
o
n
-
U
S

p
a
t
e
n
t
s

i
s
s
u
e
d

b
y

y
e
a
r

a
n
d

c
o
u
n
t
r
y
.


T
h
e

t
o
t
a
l

n
u
m
b
e
r

o
f

p
a
t
e
n
t
s

i
n

e
a
c
h

y
e
a
r

a
n
d

t
h
e

n
u
m
b
e
r

o
f

c
o
u
n
t
r
i
e
s

i
s

l
i
s
t
e
d

i
n

T
a
b
l
e

A
1

b
e
l
o
w
.

T
a
b
l
e

A
1
.


U
S
P
T
O

d
a
t
a

u
s
e
d

i
n

t
h
e

a
n
a
l
y
s
i
s

Y
e
a
r











P
a
t
e
n
t
s
N
u
m
b
e
r
Y
e
a
r
P
a
t
e
n
t
s
N
u
m
b
e
r
Y
e
a
r
P
a
t
e
n
t
s
N
u
m
b
e
r
Y
e
a
r
P
a
t
e
n
t
s
N
u
m
b
e
r
1
9
5
0



4
2
9
5
2
4
1
1
9
6
6
7
2
6
9
9
5
2
1
9
8
2
5
7
8
8
8
5
2
1
9
9
8
1
4
7
5
1
8
5
8
1
9
5
1



4
4
3
2
6
4
2
1
9
6
7
6
5
0
5
6
5
2
1
9
8
3
5
6
8
6
1
5
4
1
9
9
9
1
5
3
4
8
5
5
8
1
9
5
2



4
3
6
1
4
4
7
1
9
6
8
5
5
4
0
5
5
1
1
9
8
4
6
7
1
9
7
5
7
2
0
0
0
1
5
7
4
9
5
5
8
1
9
5
3



4
0
4
6
8
4
4
1
9
6
9
6
7
5
5
7
5
1
1
9
8
5
7
1
6
5
9
5
5
2
0
0
1
1
6
6
0
3
7
5
7
1
9
5
4



3
3
8
0
8
4
1
1
9
7
0
6
4
4
2
9
5
4
1
9
8
6
7
0
8
5
8
5
7
1
9
5
5



3
0
4
3
1
3
9
1
9
7
1
7
8
3
1
5
5
5
1
9
8
7
8
2
9
4
9
5
9
1
9
5
6



4
6
8
1
0
4
7
1
9
7
2
7
4
8
0
8
5
5
1
9
8
8
7
7
9
2
2
5
6
1
9
5
7



4
2
7
4
3
4
4
1
9
7
3
7
4
1
4
3
5
4
1
9
8
9
9
5
5
0
5
5
9
1
9
5
8



4
8
3
3
0
4
5
1
9
7
4
7
6
2
7
6
5
6
1
9
9
0
9
0
2
7
8
5
7
1
9
5
9



5
2
4
0
7
4
6
1
9
7
5
7
2
0
0
0
5
5
1
9
9
1
9
6
5
1
3
5
7
1
9
6
0



4
7
1
7
0
4
3
1
9
7
6
7
0
2
2
6
5
4
1
9
9
2
9
7
4
4
1
5
4
1
9
6
1



4
8
3
6
8
5
0
1
9
7
7
6
5
2
6
9
5
5
1
9
9
3
9
8
3
4
2
5
6
1
9
6
2



5
8
9
5
0
5
0
1
9
7
8
6
6
1
0
1
5
3
1
9
9
4
1
0
1
6
7
5
5
5
1
9
6
3



4
2
4
0
7
5
2
1
9
7
9
4
8
8
5
4
5
5
1
9
9
5
1
0
1
4
1
9
5
6
1
9
6
4



4
7
3
8
1
5
0
1
9
8
0
6
1
8
1
8
5
6
1
9
9
6
1
0
9
6
4
5
5
6
1
9
6
5



6
2
8
5
7
5
3
1
9
8
1
6
5
7
7
1
5
5
1
9
9
7
1
1
1
9
8
3
5
7



Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close