Modern Economy, 2011, 2, 181-193
doi:10.4236/me.2011.23024 Published Online July 2011 (http://www.SciRP.org/journal/me)
Copyright © 2011 SciRes. ME
The Optimal Level and Impact of Internal Factors on
Growth
Kui-Wai Li
Department of Economics and Finance, and APEC Study Center, Col le ge of Business, City University of Hong Kong,
Hong Kong, China
E-mail: efkwli@cityu.edu.hk
Received May 5, 2011; revised June 15, 2011; accepted June 25, 2011
Abstract
This paper empirically uses data from the world economy to show that performance of domestic factors are
equally important to external factors when comes to growth. Various external and domestic factors are used
to construct two separate indices and the principal component method is applied in the analysis. The empiri-
cal results show that given a different level of performance in the economy’s external factors, a higher per-
formance in the internal factors will produce a higher growth rate. When the performance of an economy’s
internal factors is extremely low, it would be appropriate for that economy first to improve its internal factors.
Keywords: Economic Openness, Domestic Factors, Globalization Index, Principle Component Analysis, World
Growth
1. Introduction
In the debate on economic growth, the neo-classical
school argues that capital and labor are exogenous fac-
tors in production, while technological advancement em-
braces a number of non-measurable factors. In contrast,
the “new growth theory” advocates the importance of
endogenous factors that incorporates human capital and a
number of institutional and domestic factors, such as the
level of corruption and protection of property rights.[1-3]
Exogenous factors are quantifiable and are derived from
outside the economic system, and examples include such
external factors as export, foreign direct investment,
tourism and international transfers. Endogenous factors
are unquantifiable and are derived from within the eco-
nomic system; examples include such internal factors as
education spending, political stability, rule of law and
other institutional factors.
Empirical growth studies have produced a mixed re-
sult in the impact of different factors on growth and glo-
balization. While external factors are considered crucial
gains in the gains process of globalization, critics have
considered the costs of globalization in terms of domes-
tic factors. [4-14] Similarly, the construction of globali-
zation indices have popularly been based on a mixture of
external and domestic factors. [15-18] Other cross-coun-
try empirical studies have identified a great number of
domestic and geographical factors that have various de-
grees of impact on growth.[19-26]
Although the performance of the external factors con-
tribute to economic growth, the link between the external
factors and an economy’s growth performance in the
global community depend to a considerable extent on
how well internal factors have performed.[27] A more
matured capital market, for example, will facilitate a
greater capital flow, while a more transparent, corrup-
tion-free investment environment could attract more for-
eign direct investment. Successful performance of inter-
nal factors can complement the performance of external
factors in the growth and globalization process.
The performance of both external factors and internal
factors can impact on growth directly, but there is also an
indirect link between internal factors and growth. The
good performance in internal factors can exercise an ad-
ditional and positive influence on the performance of
external factors, which in turn can have a greater impact
on growth. It is probable that internal factors can influ-
ence economic growth directly, and can indirectly impact
on growth through a better performance in external fac-
tors. As such, the internal factors can be the more fun-
damental factors to growth than external factors.[25-26]
This paper empirically investigates the hypothesis that
internal factors are the more fundamental factors to
growth than external factors. Two separate indices for
K.-W. LI
Copyright © 2011 SciRes. ME
182
external factors and internal factors will first be con-
structed. Instead of examining the impact of a single
factor each of the two indices is constructed from a total
of 17 factors.[22,24,28] Regression analysis shows how
the two groups of factors can independently impact on
growth. To show how the internal factors can exert in-
dependent influence on growth, a simulation analysis is
used to find the optimal level of performance in each
economy’s internal factors. Lastly, the sample world
economies are mapped according to their performance in
the external factors and internal factors. The empirical
result shows that economies will have to achieve a cer-
tain level in the performance of the internal factors be-
fore they can take advantage of the performance in ex-
ternal factors. The data sources of the 34 factors for the
62 world economies for the period 1998-2002 are given
in the Appendix.
Section II uses the principal component analysis me-
thod to work out the two indices for the external factors
and internal factors. Section III gives the regression es-
timates, while section IV compiles an optimal level of
performance in each economy’s internal factors and a
simulation study is conducted to show how the 62 world
economies performed in the two types of factors. Section
V concludes the paper.
2. The Two Indices
Instead of pulling different factors into a single globali-
zation index, this section constructs two separate indices
on the external factor and internal factors. Kearney [15]
grouped the external factors into the four categories of
economic integration, technological connectivity, per-
sonal contacts and international engagement. Kearney’s
selection of external factors and categories can be im-
proved by incorporating the inter-industry trade and in-
tra-industry trade indicators. In theory, trade statistics are
post-trade data that reflect the outcome of trade policies
and show the actual quantity of export and import. An
economy’s inter-industry trade is traditionally based on
comparative advantage. In intra-industry trade, eco- no-
mies export and import the same good or service in a
given period. Thus, intra-industry trade reflects more on
the varieties of goods the economy enjoys due to indus-
trial diversity and technological advancement than sim-
ply trade flows based on comparative advantages. The
calculation of the inter-industry and intra-industry trade
indicators is shown in the Appendix. The External Fac-
tors Index (EFI) is constructed from a total of 17 ex-
ternal economic openness factors grouped under six
categories.
The choice of internal factors used to construct the In-
ternal Factors Index (IFI) is chosen from the list in Dur-
lauf. [19] The 17 internal factors are classified into three
broad categories. While the first category of institutional
establishment is considered as proxy measures for civili-
ty, security and protection of individuals, the other two
categories provide indicators on the quality of life. Ap-
pendix Table 1 summarizes the categories of external
factors and internal factors. Both the external and inter-
nal factors are normalized on a yearly basis before they
are used to construct the two indices.1 [29]
The principal component analysis (PCA) is used to
construct the two indices. There are several advantages
in using the PCA method. First, the PCA is meant to give
weightings that maximize the variance of the indices.
Since the factors are likely to be correlated, the PCA
reduces the number of factors to capture the maximum
variation. Secondly, the PCA method can commensurate
on the different measurement units of these factors. Most
importantly, the PCA method selects the weights by the
data itself. [30] The principal components are extracted
from the correlation matrix of the variables, in a way that
they accounted for the highest percentage of variation.
The PCA is applied to each individual year instead of
applying one PCA to the whole sample period. This has
the advantage of incorporating various changes in the
sample period, and can eliminate the impact of a sudden
change in any particular year that could affect other
sample years.2
We adopt a latent variable model and postulate that the
index is linearly dependent on a set of observable factors
(V) and an error term. The principal components (PCs)
are computed from the following procedure:
1111 1
2211 2
11
,,
,,
,,
LL L
PC VV
PC VV
PC VV









, (1)
where 11 121
,,,

are elements of eigenvector
111 1
,,
 
, and there are a total of L eigenvectors,
which are determined by the data. A total of L principal
components are computed using successive eigenvectors
elements, 12
,,,
L

, corresponding to the largest L
eigenvalues, 12
L

, of the factor correlation
matrix. The first principal component, PC1, of the linear
combination with maximal variance becomes our EFI,
1The normalization formulas for the high and low value variables are:
 
111
min, ,max(, ,)min(, ,)
itiNNN t
Vv vvvvvv , and
 

111
max, ,max, ,min, ,
itNiNN t
Vvvv vvvv . Vit is
variable V of economy i at time t.
2This is seen as an improvement to Andersen and Herbertsson[16] who
m
use a single principle component for all years, and to Dreher[18]
whom uses weightings of year 2000 for the calculation of indices fo
r
all years from 1970 to 2000.
K.-W. LI
Copyright © 2011 SciRes. ME
183
Table 1. Pooled-GLS Estimates of 62 World Economies,
1998-2002.
Coefficients k = 3 k = 4 k = 8 k = 10
α 7.5159
(0.0722)*
7.3161
(0.0861)*
7.5144
(0.0967)*
7.5269
(0.0956)*
β1
0.2904
(0.0270)*
0.3591
(0.0360)*
-0.0324
(0.0911)
-0.0868
(0.0920)
β2
0.3036
(0.0073)*
0.2260
(0.0163)*
0.3593
(0.0729)*
0.3916
(0.0739)*
β3 0.3690
(0.0097)*
0.3472
(0.0174)*
0.4956
(0.0730)*
0.5224
(0.0731)*
β4
0.3421
(0.0188)*
0.5961
(0.0750)*
0.5561
(0.0749)*
β5
0.6334
(0.0762)*
0.6447
(0.0759)*
β6
0.7027
(0.0766)*
0.6757
(0.0770)*
β7
0.6847
(0.0777)*
0.7346
(0.0771)*
β8
0.6894
(0.0779)*
0.7523
(0.0782)*
β9
0.7342
(0.0787)*
β10
0.7427
(0.0788)*
F-test
Adj. R2
Wald Test
0.0000
0.999704
0.0000
0.0000
0.999624
0.0000
0.0000
0.999670
0.0000
0.0000
0.999745
0.0000
Notes: Figures in parentheses are standard errors. * and † = significance at
1% and 5% levels, respectively.
which is then normalized or scaled.3 The scaled EFI will
take a value of unity when an economy has the best per-
formance in its external environment. The same proce-
dures are applied to the construction of the Internal Fac-
tors Index (IFI).
Appendix Table 2 gives the five-year (1998-2002)
average of the EFI and IFI. The ranking based on the
five-year average shows that the top 10 economies in the
two indices are mainly advanced economies in North
America and Western Europe. Most of the remaining
European Union economies are included when the scores
are extended to the top 20. Singapore and Hong Kong are
the only two Asian economies in the top 20 of both indi-
cators. We observe that an economy can vary between
the two indices. For example, Japan ranked 18th in the
IFI, but ranked 26th in the EFI, while Indonesia ranked
44th and 55th in EFI and IFI, respectively. Economically
weaker economies tend to rank lower in the two indica-
tors. Effectively, economies that ranked below 30th are
all developing economies.
3. Regression Estimates
The hypothesis that economies with strong performance
in internal factors enjoy a higher rate of per capita GDP
growth at different level of performance in the external
factors is examined. The IFI is divided into k portions
using percentiles, shown in Equation (2), with N being
the number of economies.
For example, the IFI of year t is divided into three
portions, so k = 3, with 33.33 percent of the economies
in each portion. The first portion is made up of the min-
imum IFI in year t to the 33rd IFI in year t. A dummy
variable, D
, where 1,k
, is assigned to each of
the last (k-1) portions of IFI, namely2,k
DD. The D
dummy takes a value of unity if IFIit falls into the
th
portion, otherwise it takes a value of zero. An economy
with 1D
has a better internal environment than an
economy with 11D
.
The following model is used to examine how internal
factors can affect the outcome of external factors:
122,
,
lnlnln *
ln *,
ititit it
kitk itit
yEFIEFI D
EFI D
 

 
  (3)
where yit is the real GDP per capita deflated by the pur-
chasing power parity of economy i at time t. For economy
i who has the dummy 1D
, the regression equation
become:
1
lnln .
itit it
y EFI
 
 (4)
For another economy j has the dummy 1,
c
D
for
any c > 0. In other words, when economy j’s internal
environment is not as good as economy i’s, the regres-
sion equation become:
1
lnln .
j
tcjtjt
y EFI
 
 (5)
If a higher performance in internal factors brings a


100 100100
min, ,%*,1%*, ,2%*,,
100 100
11 %*,,%*.
th th
th
titt
th th
t
IFIIFIN IFIN IFINIFI
kk k
kNIFIkNIFI
kk



 
 

 
  
 








 
 

  
 




(2)
3Scaled
 
min maxmin
itiii
it
EFI EFIEFIEFIEFI
K.-W. LI
Copyright © 2011 SciRes. ME
184
Table 2. The EFI – IFI matrix of world economies, 1998-2002 average.
Range
Internal Factors Index (IFI)
0.00 - 0.20 0.21 - 0.40 0.41 - 0.60 0.61 - 0.80 0.81 - 1.00
External Factors Index (EFI)
0.00 - 0.20
Uganda (4.049)
Bangladesh
(3.025)*
Senegal (2.322)
N
igeria (1.575)*
Indonesia (1.408)
Pakistan (1.398)
Kenya (–1.343)
China (6.749)
Russian Fed. (6.381)
Ukraine (5.692)
India (3.287)
Romania (3.071)
Egypt (2.932)
Iran (2.786)
Sri Lanka (1.928)
Philippines (1.239)
Brazil (1.229)
S. Africa (1.227)
Mexico (1.001)
Peru (0.768)
Turkey (–0.096)
Colombia (–0.807)
Venezuela (–3.697)
Botswana (8.615)
Tunisia (3.198)
Thailand (2.911)
Chile (1.072)
Morocco (0.720)
Saudi Arab. (–0.938)
Argentina (–5.887)
0.21 - 0.40 Croatia (3.654)
Korea (5.957)
Greece (4.207)
Slovak Rep. (3.341)
Poland (2.981)
Malaysia (2.945)
Panama (0.661)
Hungary (3.869)
Slovenia (3.858)
Czech Rep. (3.354)
Spain (2.671)
Portugal (1.945)
Italy (1.590)
Japan (0.477)
Israel (–0.096)
0.41 - 0.60 Hong Kong (3.346)
France (2.201)
N
ew Zealand (3.150)
Canada (2.829)
Australia (1.821)
N
orway (1.374)
Germany (1.175)
0.61 - 0.80
Singapore (4.082)
Sweden (2.500)
Finland (2.161)
U.K. (2.102)
Denmark (1.788)
Austria (1.723)
N
etherlands (1.617)
USA (1.455)
Switzerland (1.095)
0.81 - 1.00 Ireland (9.737)
Note: Figures in parenthesis are the percentage growth rates of the average 1999-2002 GDP per capita (purchasing power parity in constant 2000
price). *Countries with IFI < 0.
higher marginal effect of external factors on economic
growth, we expect to see 11c

 
. Thus, ge-
neralizing all the k dummy variables, and if a better in-
ternal environment has a positive impact of external fac-
tors on growth, we expect to see 112


13 1k
 
 , suggesting that a strong perfor-
mance in an economy’s internal factors enables an
economy to benefit more from performance in external
factors.
Two Wald tests are conducted to show the signific-
ance of the coefficient estimates. The first Wald test is to
see if a low performance in the internal factors will con-
strain economic growth. An alternative hypothesis with
10
implies that if an economy has an extremely
weak performance in its internal factors (reflected in the
IFI value falling into the first partition of the index), ex-
ternal factors would bring negative effects on economic
growth, namely:
1
1
1
1
:0
:0.
Ho
Ha
(6)
The second Wald test shows that an economy’s IFI
can significantly affect the marginal effect of an econo-
my’s external factors on its real per capita GDP growth
rate:
2
1
2
111
:02,,.
:3,,.
Hofor k
H
afork

 
 
 
 
(7)
The alternative hypothesis, 2
H
a, states that econo-
mies that have a better performance in their internal fac-
tors should benefit more from performance in external
factors.
The pooled-GLS with White-Heteroskedasticity con-
sistent standard error and covariance is applied to esti-
mate Equation (3), which is estimated with k = 3, 4, 8
and 10. Table 1 shows the empirical estimation of the
K.-W. LI
Copyright © 2011 SciRes. ME
185
pooled-GLS results for the 62 countries for the sample
period of 1998-2002. All estimates with k = 3 and k = 4
in Tab le 1 are significant at 1 percent level. In these two
cases, the estimate for
1 is not negative, but is signifi-
cantly different from zero, suggesting that a low perfor-
mance in internal factors does not adversely affect the
effect on economic growth, though this may be due to
the small size of k. When the size of k is small, the mar-
ginal effect of internal factors on economic growth may
not be obvious. The F-tests reject the null hypothesis of
Equation (3), suggesting that as economies improve their
performance in internal factors, the marginal effect on
growth increases.
For estimates with k = 8 and k = 10, and with the ex-
ception of the insignificant estimate for
1, all the esti-
mates are significance at 1 percent level. For these esti-
mated values of k, the estimate of
1 is negative, which
means that a low performance in internal factors of an
economy can adversely affect growth. Similarly, the
F-tests also reject the null hypothesis as in the cases of k
= 3 and k = 4.
4. Optimal Performance in Internal Factors
This section uses a simulation method to work out the
optimal performance in the internal factors in order to
achieve a maximum gain in economic growth. From the
estimation result of k=4, 8 and 10 in Table 1, we first
examine economies with top scores in IFI to see if there
is diminishing returns in the external factors. Hypotheti-
cal economies are compared in order to see how their
growth performs given a different level of performance
in internal factors.
Two hypotheses are postulated. First, given two ex-
ternally homogeneous economies (namely, economies
with same performance in the EFI), heterogeneity in the
performance of IFI will lead to differences in economic
growth. Secondly, given homogeneity in the perfor-
mance of IFI among different economies, those econo-
mies with a better performance in EFI will result in
higher economic growth.
The empirical result with k = 10 in Table 1 is used to
simulate the growth of per capita GDP for a total of 100
hypothetical economies with an incremental change of
0.01 in the IFI that ranged from zero to one. The differ-
ent values of the EFI are either below or above the me-
dian value. A simulated series of per capita GDP figures
are generated from the empirical results with k = 10 in
Table 1.4 The simulated per capita GDP growth rates are
plotted against the IFI, and a step function is presented
separately for the four values of EFI (at 0.25, 0.45. 0.75
and 0.95) as shown in Figure 1.
The first observation in Figure 1 is that economies
with a higher performance in external factors (with high-
er EFI) produce a higher level of per capita GDP growth
at all level of IFI above 0.1. In economies with IFI below
the median, a higher performance in EFI always produc-
es a higher economic growth, except when IFI is below
0.1. The second observation is that, when the IFI is
above median, economic growth keeps rising regardless
of the performance in the EFI until an economy’s IFI
reaches the range of 0.7 and 0.8, beyond which the
growth rate of per capita GDP declines. This suggests
that the 0.7 to 0.8 range of the IFI is the optimal level,
and economies will reach their highest possible growth
rates given their EFI. When the value of EFI lies be-
tween 0 and 1, the marginal contribution of IFI to the per
capita GDP growth of an economy is positive if the value
of IFI lies between 0 and the optimal level. When the
value of IFI is above its optimal level, the marginal con-
tribution of IFI to an economy’s GDP per capita growth
is negative.5
In short, if an economy has an IFI value below 0.1, a
lower value of EFI actually produces a higher per capital
GDP growth. So long as the value of IFI lies above 0.1,
the marginal contribution by the different level of EFI to
per capita GDP growth is positive. On the contrary, when
IFI lies between 0 and 0.1, the marginal contribution of
EFI to per capita GDP growth is negative.6
The marginal effect of both EFI and IFI can be ex-
amined from plotting the change in the per capita GDP
growth rate against the IFI at different level of the EFI,
Figure 2 shows that a higher EFI can lead to a larger
change in the growth rate of per capita GDP at different
level of IFI.7 However, as shown in Figure 3, the mar-
ginal effect of IFI on the change in growth rate of per
capita GDP at different level of EFI is increasing at a
decreasing rate. Furthermore, Figure 3 shows that when
the EFI value is below the median, its marginal contribu-
tion to growth is larger than that when EFI is above the
5This can also be seen if Equation (3) is modeled as a continuous o
r
differentiable function, where 0< i < 1, and IFI* represents the optimal
value:
**
0, ; ;
lnlnln
0; 0; 0
iii
IFI MedianEFI EFIIFI IFIEFI EFIIFI IFIEFI EFI
yyy
IFIIFI IFI
 



6When the function is a differentiable, the results can be summarized as
follows:
0.5 10.10.500.1
ln lnln
0; 0;
0
ln lnln
IFI IFIIFI
yy y
EFI EFIEFI
 
 
 
 
7The marginal effect can be summarized as follows when a differentia-
ble equation is used:
0.25 0.450.95
0.75
ln ln ln ln
EFI EFIEFI
EFI
yyyy
IFI IFI IFI IFI
 



4For example, when EFI = 0.25, and with 3, 1
it
D (namely, the range
of IFI is between 0.2 and 0.3, and other dummies take a zero value),
the simulated GDP per capita growth is 8.92904 (i.e. 7.52687 +
(–0.08675)*ln(0.25*100) + (0.522359)*ln(0.25*100)* 1).
K.-W. LI
Copyright © 2011 SciRes. ME
186
Figure 1. Effect of external factors on growth.
Figure 2. Marginal effect of EFI on growth.
K.-W. LI
Copyright © 2011 SciRes. ME
187
median.8
With the construction of the two indices that look sepa-
rately at internal factors and external factors, the regres-
sion and simulation results can provide additional evi-
dence to support other studies that internal factors can
have independent influence on growth. [17] Various pol-
icy recommendations can be suggested from the empiri-
cal and simulation analysis. Firstly, a more global eco-
nomy indicated by the higher performance in the external
factors does not always lead to higher economic growth.
Instead, those economies with 0 < IFI < 0.1 should im-
prove their IFI in order to reap additional gain from
economic openness and globalization. Secondly, econo-
mies whose IFI is above 0.1, but below the optimal range
(0.7 to 0.8), should aim to improve the performance of
the internal factors.
A summary pattern of relationship between economic
growth and the performance in the external factors and
internal factors seems to have emerged from the simula-
tion analysis. Figure 4 shows that once the performance
in the internal factors has reached a minimum level, im-
provement in internal factors will lead to a larger per
capita GDP growth rate at every higher level of EFI.
Thus, at a high level of external factors, EFI3 for exam-
ple, a higher level of per capita GDP growth rate can be
achieved.
To see how the 62 world economies perform in the
1998-2002 period, Table 2 maps out the sample period
average in five different ranges of EFI and IFI. Individu-
al economies can consider their own positions in the
ranking of the two indices, and compare their perfor-
mance with other economies, including the periodic av-
erage in the per capita GDP growth rates. There are sev-
en mainly poor developing economies (Bangladesh, In-
donesia, Kenya, Nigeria, Pakistan, Senegal and Uganda)
that have the lowest rankings in both indices. On the
contrary, those economies that performed strongly in
both EFI and IFI are mainly developed economies (Aus-
tria, Denmark, Finland, Netherlands, Singapore, Sweden,
Switzerland, United Kingdom and USA). Most devel-
oped economies have performed stronger in IFI than in
EFI. Ireland is the only economy that has a stronger per-
formance in EFI than in IFI in the sample period.9
One observation from Table 2 is that performance of
internal factors is the relevant constraint in the growth of
any economy. Most economies that are strong in the
performance of IFI are also strong in the performance of
EFI, but not the reverse. In other words, it would be ap-
propriate for economies to improve their internal condi-
tions and environment before they can gain from open-
ness and globalization. A good performance in internal
factors is essential to growth and development. There are
a number of economies (Argentina, Botswana and so on)
that have achieved the median in IFI, but show low per-
formance in EFI. The 0.61 to 0.80 range of the IFI seems
to be the critical range, as virtually all industrialized ad-
vanced economies achieved an IFI score above 0.61.
Table 2 shows that a number of economies in the
second lowest (0.21 – 0.40) range of IFI experience a
relative high growth rate in the sample period. For ex-
ample, China has a growth rate of 6.749 percent and the
Russian Federation had 6.381 percent and so on. This
suggests that these economies have to improve their IFI
before further reaping the gain from economic openness
and globalization. Among the developing economies,
African economies (e.g. Uganda, Kenya and Senegal) are
the weakest performers in both the EFI and IFI, while
the middle-ranking economies are the few Asian (e.g.
Thailand and Malaysia) and Latin American (e.g. Pana-
ma and Chile) economies. Other Asian economies (e.g.
India, Indonesia, Philippines and Sri Lanka) have per-
formed poorly in both EFI and IFI. The group of devel-
oping economies that have reached the range of 0.61 –
0.80 in the IFI are mostly Eastern European economies
(e.g. Hungary, Slovenia and Czech Republic), which will
probably be the next group of countries that would bene-
fit from globalization. The lesson is that sound perfor-
mance in the various internal factors will facilitate good
performance of external factors. In short, advancement in
the performance of internal factors will help promoting
economic openness.
5. Conclusions
The empirical results in the paper clarify the importance
of internal factors in growth and globalization. The rea-
sonable large number of factors used in the construction
of the two indices provides a comprehensive picture on
the performance of different economies. The regression
result that internal factors are important in promoting an
economy’s growth has led to further investigation and
analysis in the relationship of the two types of factors.
Given a different level of performance in the economy’s
external factors, a higher performance in the internal
factors will produce a higher growth rate. When the per-
formance of an economy’s internal factors is extremely
low, it would be appropriate for that economy first to
improve its internal factors.
In short, performance in the internal factors is the
more fundamental condition than performance in the
external factors. Before the “optimal” level of internal
8The marginal effect can be summarized as follows when a differentia-
ble equation is used: ln ln
ln ln
ii
Below MedianAbove Median
I
FI IFIIFI IFI
YY
EFI EFI



9Measured in purchasing power parity constant 2000 price, Ireland’s
GDP per capita is highest among the 62 world economies.
K.-W. LI
Copyright © 2011 SciRes. ME
188
Figure 3. Effect of a change in EFI.
Figure 4. Relationships between growth, external and internal factors.
factors is reached, economies will experience a rise in
their per capita GDP as their performance of internal
factors improve. The empirical results in this paper lend
further support on the importance of a sound perfor-
mance in domestic factors.[31-32] Economies with
strong performance in external factors and globalization
have sound performance in their internal factors. For
those world economies that are ranked low in the Internal
Factor Index, appropriate economic policies should be
conducted to improve the performance of internal factors.
The conclusion that the internal or institutional factors
are more fundamental to growth than external factors
adds to the debate on the difference between the two
types of factors, or the contribution of single factors.[24,
28]
Despite the useful empirical findings and the policy
lessons economies can draw on, there can be a number of
possible drawbacks in this paper. One is the selection of
factors in the two categories. It is possible that different
factors selection would produce different empirical re-
sults, and the use of 34 external and internal factors can
provide sufficient representation. The problem of possi-
K.-W. LI
Copyright © 2011 SciRes. ME
189
ble overlap among factors can partly be alleviated by the
advantages of the principal component analysis. [33-34]
6. Acknowledgements
The author is indebted to comments from Barbara Stal-
lings, Neantro Saavedra-Rivano, Gianluca F. Grimalda,
Peter J. Newell, Eden Yu, participants in the 2006 meet-
ing of the APEC Study Center Consortium and col-
leagues in the brown-bag seminar of the Department of
Economics and Finance, City University of Hong Kong.
An earlier version of this paper has been published as a
Working Paper in the Center for the Study of Globalisa-
tion and Regionalism, University of Warwick. The two
research assistants, Iris Pang and Michael Ng, have pro-
vided excellent research support. The funding support
from the City University of Hong Kong under the Stra-
tegic Research Grant (Number 7002433) is gratefully
acknowledged. The author is solely responsible for the
remaining errors.
7. References
[1] P. M. Romer, “Increasing Returns and Long Run
Growth,” Journal of Political Economy, Vol. 94, No. 5,
1986, pp. 1002-1037. doi:10.1086/261420
[2] P. M. Romer, “Endogenous Technological Change,” Jou-
rnal of Political Economy, Vol. 98, No. 5, 1990, pp. S71-
S102. doi:10.1086/261725
[3] R. E. Lucas Jr., “On the Mechanics of Economic Devel-
opment,” Journal of Monetary Economics, Vol. 22, No. 2,
1988, pp. 3-42. doi:10.1016/0304-3932(88)90168-7
[4] L. Winters, “Trade Policies for Poverty Alleviation,” In:
B. Hoekman, A. Mattoo and P. English, Eds., Develop-
ment, Trade and the WTO: A Handbook, The World Bank,
Washington D.C., 2002, pp. 28-38.
[5] A. Deardorff and R. Stern, “What You Should Know
about Globalization and the World Trade Organization,”
Review of International Economics, Vol. 10, No. 3, 2002,
pp. 404-423. doi:10.1111/1467-9396.00340
[6] J. Bhagwati, “In Defense of Globalization,” Oxford Uni-
versity Press, New York, 2004.
[7] E. Aisbett, “Why are the Critics so Convinced that Glo-
balization is Bad for the Poor?” Working Paper No.
11066, National Bureau of Economic Research, 2005.
[8] J. Frankel, “Globalization of the Economy,” Working
Paper 7858, National Bureau of Economic Research,
2000.
[9] R. Falvery and E. Kreickemeier, “Globalization and Fac-
tor Returns in Competitive Markets,” Journal of Interna-
tional Economics, Vol. 66, No. 1, 2005, pp. 233-248.
doi:10.1016/j.jinteco.2004.04.004
[10] E. Edmonds and N. Pavcnik, “Does Globalization In-
crease Child Labor? Evidence from Vietnam,” Working
Paper 8760, National Bureau of Economic Research,
2002.
[11] M. Feldstein, “Aspects of Global Economic Integration:
Outlook for the Future,” Working Paper 7899, National
Bureau of Economic Research, 2000
[12] L. Wallach and P. Woodall, “Whose Trade Organiza-
tion,” New Press, New York, 2004.
[13] J. Stiglitz, “Globalization and Its Discontent,” Allen Lane,
London, 2002.
[14] S. Fischer, “Globalization and Its Challenge,” Ely Lec-
ture, American Economic Association Meeting, Wash-
ington D.C., January 2003.
[15] A. T. Kearney, “Measuring Globalization: Economic
Reversals, Forward Momentum,” Foreign Policy, Vol.
Vol. 141, 2004, pp. 54-69.
[16] T. Andersen and T. Herbertsson, “Quantifying Globaliza-
tion,” Applied Economics, Vol. 37, No. 10, 2005, pp.
1089-1098. doi:10.1080/00036840500118200
[17] A. Heshmati, “Measurement of a Multidimensional Index
of Globalization,” Global Economy Journal, Vol. 6, No.
2, 2006, pp. 1-28. doi:10.2202/1524-5861.1117
[18] A. Dreher, “Does Globalization Affect Growth? Evidence
from a New Index of Globalization,” Applied Economics,
Vol. 38, No. 10, 2006, pp. 1091-1110.
doi:10.1080/00036840500392078
[19] S. N. Durlauf, P. A. Johnson and J. R. W. Temple,
“Growth Econometrics,” In: P. Aghion and S. N. Durlauf,
Eds., Handbook of Economic Growth, Vol. 1, Elsevier,
Amsterdam, 2005, pp. 555-680.
doi:10.1016/S1574-0684(05)01008-7
[20] D. Rodrik, “Growth Strategies,” In: P. Aghion and S. N.
Daulauf, Eds., Handbook of Economic Growth, Vol. 1,
Elsevier, Amsterdam, 2005, pp. 967-1014.
doi:10.1016/S1574-0684(05)01014-2
[21] R. Hausmann, L. Pritchett and D. Rodrik, “Growth Ac-
celeration,” Journal of Economic Growth, Vol. 10, No. 4,
December 2005, pp. 303-329.
doi:10.1007/s10887-005-4712-0
[22] A. Harrison, “Openness and Growth: A Times-Series,
Cross-Country Analysis for Developing Countries,” Jou-
rnal of Development Economics, Vol. 48, 1996, pp. 419-
447. doi:10.1016/0304-3878(95)00042-9
[23] D. Rodrik, A. Subramanian and F. Trebbi “Institutions
Rule: The Primacy of Institutions over Geography and
Integration in Economic Development,” Journal of Eco-
nomic Growth, Vol. 9, No. 2, 2004, pp. 131-165.
doi:10.1023/B:JOEG.0000031425.72248.85
[24] B. Axel, M. Busse and S. Neuhaus, “Institutional Quality
and the Gains from Trade,” KYKLOS, Vol. 59, No. 3,
2006, pp. 45-368.
[25] K.-W. Li and X. Zhou, “Openness, Domestic Perfor-
mance and Growth,” Economic Letters, Vol. 107, No. 1,
2010, pp. 13-16. doi:10.1016/j.econlet.2009.12.012
[26] X. B. Zhou and K.-W. Li, “The Impact of Openness and
Domestic Performance of Growth using Non-parametric
Estimations,” Applied Economics Letters, Vol. 18, No. 11,
2011, pp. 1065-1070.
K.-W. LI
Copyright © 2011 SciRes. ME
190
[27] D. Rodrik, “Has Globalization Gone too Far?” Institute
for International Economics, Washington D.C., 1997.
[28] B. Sambit, S. Dowrick and J. Holley, “Institutions and
Trade: Competitors or Complements in Economic De-
velopment?” Manuscript, The Australian National Uni-
versity, 2008.
[29] B. Lockwood, “How Robust Is the Foreign Policy-Kea-
rney Globalisation Index?” The World Economy, Vol. 27,
2004, pp. 507-523.
doi:10.1111/j.0378-5920.2004.00611.x
[30] A. Rencher, “Methods of Multivariate Analysis,” 2nd
Edition, Wiley-Interscience, New York, 2002.
doi:10.1002/0471271357
[31] G. James and H. Kum, “Inequality and Economic Growth
– Data Comparison and Econometrics Test,” University
of Texas Inequality Project, Working Paper Number 21,
2002.
[32] D. Dollar and A. Kraay “Institutions, Trade and Growth,”
Journal of Monetary Economics, Vol. 50, No. 1, 2003, pp.
133-162. doi:10.1016/S0304-3932(02)00206-4
[33] C. Fernandez, L. Eduardo and M. F. J. Steel, “Model
Uncertainty in Cross-Country Growth Regressions,” Jou-
rnal of Applied Econometrics, Vol. 16, No. 5, 2001, pp.
563-576. doi:10.1002/jae.623
[34] X. Sala-i-Martin, “I Just Ran 2 Million Regressions,”
American Economic Review, Vol. 87, No. 2, 1997, pp.
178-183.
[35] United Nations, “United Nations Comtrade,” 1998-2002,
New York.
[36] B. Balassa, “Traded Liberalization and ‘Revealed’ Com-
parative Advantage,” The Manchester School of Eco-
nomic and Social Studies, Vol. 33, No. 2, 1965, pp.
99-123. doi:10.1111/j.1467-9957.1965.tb00050.x
[37] B. Balassa, “‘Revealed’ Comparative Advantage Revi-
sited: An Analysis of Relative Export Shares of the In-
dustrial Countries, 1953-1971,” The Manchester School
of Economic and Social Studies, Vol. 45, 1977, pp.
327-344. doi:10.1111/j.1467-9957.1977.tb00701.x
[38] B. Balassa, “The Changing Pattern of Comparative Ad-
vantage in Manufactured Goods,” Review of Economics
and Statistics, Vol. 61, No. 2, 1979, pp. 259-266.
doi:10.2307/1924594
[39] B. Balassa, “Comparative Advantage in Manufactured
Goods: A Reappraisal,” Review of Economics and Statis-
tics, Vol. 68, No. 2, 1986, pp. 315-319.
i:10.2307/1925512
[40] International Monetary Fund, “International Financial
Statistics,” International Monetary Fund, Washington
D.C., 1998-2002.
[41] United Nations, “Balance of Payments Statistics,” New
York, 1998-2002.
[42] International Telecommunication Union, “International
Telecommunication Union Database,” International Tel-
ecommunication Union, Geneva,1998-2002.
[43] International Telecommunication Union, “Netcraft Se-
cure Server Surveys,” International Telecommunication
Union, Geneva, 1998-2002.
[44] World Tourism Organization, “Compendium of Tourism
Statistics,” New York, 1998-2002.
[45] Central Intelligent Agency, “The World Factbook,” The
Central Intelligent Agency, Washington D.C., 1998-2002.
[46] United Nations, “United Nations Development Program
Indicators,” New York, 1998-2002.
[47] World Bank, “World Development Indicators,” The
World Bank, Washington D.C., 1998-2002.
[48] Transparency House, “Corruption Index,” Transparency
House, Washington D.C., 1996-2003.
[49] World Bank, “Aggregate Governance Indicators,” The
World Bank, Washington D.C., 1998-2002.
[50] Heritage Foundation, “Index of Economic Freedom,” The
Heritage Foundation, Washington D.C., 1998-2002.
K.-W. LI
Copyright © 2011 SciRes. ME
191
Appendix
Data
All data are obtained from established international
sources. The per capita GDP data are obtained from the
World Development Indicators, The World Bank. The
inter-industry trade index and the intra-industry trade
index are compiled using the UN Comtrade Database,
SITC Rev.3, for all the 62 economies with all commodi-
ties up to two-digit level.[35] The performance of in-
ter-industry trade is estimated from an economy’s re-
vealed comparative advantage (RCA) index.[36-39] An
economy’s RCA index can be calculated by:



,it gigwgiwt
RCAXXX X, (A1)
where ig
X
denotes economy i’s export of commodity g,
wg
X
is world export of commodity g, i
X
is economy
i’s total export and w
X
is total world exports, where
1,iN, 1,,tT and 1,,
g
G
. When the val-
ue of ,it g
RCA exceeds unity, economy i is said to have a
revealed comparative advantage in good g at time t. The
total number of export industries of individual economies
with revealed comparative advantage greater than unity
is selected and normalized (NRCA) to form an indicator
for the economy’s inter-industry trade performance
(it
TRCA):

iti it
TRCANRCA MAXNRCA. (A2)
The intra-industry trade index (IIT) can be calculated as:


,,
1,,
,,
1,,
1 *100
1*100
j
j
nijgijg
g
jij gij g
g
it
nijgijg
g
ijij gij g
gt
XM
XM
IIT XM
MAX XM












 


















,(A3)
where Xij,g is the export value of good g from country i to
country j, Mij,g is the import value of good g to country i
from country j, and
j
n= total number of economy i’s
trading partners. Equation (A3) shows the weighted av-
erage of individual industry indices, where the weights
are the shares of industries in total trade.
The sources of data for the 17 factors in each of the
Internal Factors Index and the External Factor Index
shown in Appendix Table 1 are:
Table 1. The classification of external factors and internal factors.
External Factors Data SourceInternal Factors Data Source
Economic integration (% GDP):
1) Total trade flows
2) Foreign direct investment
3) Portfolio capital flows
4) Investment income
Inter-industry trade (SITC, 2-digit):
5) Revealed comparative advantage
Intra-industry trade (SITC, 2-digit):
6) Export and import: same product
Technology connectivity:
7) Internet users (% population)
8) Internet hosts (per capita)
9) Secure servers (per capita)
Personal contact:
10) International travel & tourism
(% population)
11) International telephone traffic
(minutes per capita)
12) Remittances (% GDP)
13) Personal transfers (% GDP)
International engagement:
14) Membership in international
organizations
15) Government transfer (% GDP)
16) International treaties ratified
17) Personnel and financial
contribution to United Nations
Security Council missions
(% population)
IFS
IFS
IFS
BOPS
UN
UN
ITU
ITU
Net
SSCT
ITU
BOPS
BOPS
WFB
BOPS
OFW
UNDPI
Institutional establishment:
1) Patent applications
2) Corruption Perception
Index
3) Voice and accountability
4) Political stability
5) Government effectiveness
6) Regulatory quality
8) Rule of law
8) Control of corruption
9) Property right protection
10) Regulatory scores
Education and health:
11) Public spending on
education (% of GDP)
12) Primary school
pupil-teacher ratio
13) Total health expenditure
(% of GDP)
14) Physicians per thousand
people
15) Primary school enrolment
(% gross)
Quality of labor force:
16) Youth unemployment
(% of labor force ages
15-24)
17) Labor force, children
10-14 (% of age group)
WDI
CI
AGI
AGI
AGI
AGI
AGI
AGI
IEF
IEF
WDI
WDI
WDI
WDI
WDI
WDI
WDI
K.-W. LI
Copyright © 2011 SciRes. ME
192
Table 2. External factors and internal factors indices: 1998-2002 average.
Ranking
External Factors Index Internal Factors Index
Economies Index Economies Index
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
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
Ireland
United States
Netherlands
Switzerland
Sweden
Finland
Singapore
Denmark
Austria
United Kingdom
Canada
New Zealand
Australia
Norway
Germany
France
Hong Kong
Portugal
Spain
Italy
Czech Republic
Israel
Slovenia
Hungary
Slovak Republic
Japan
Malaysia
Panama
Greece
Poland
Korea
Croatia
Argentina
Chile
Philippine
Brazil
Russian
Thailand
Mexico
China
Turkey
Romania
South Africa
Indonesia
Ukraine
Botswana
India
Tunisia
Colombia
Peru
Senegal
Venezuela
Nigeria
Egypt
Kenya
Morocco
Pakistan
Sri Lanka
Uganda
Saudi Arabic
Iran
Bangladesh
1.00
0.70
0.72
0.65
0.65
0.62
0.64
0.61
0.60
0.60
0.60
0.56
0.50
0.48
0.49
0.48
0.47
0.40
0.38
0.37
0.35
0.32
0.30
0.27
0.28
0.27
0.26
0.25
0.24
0.23
0.23
0.20
0.19
0.17
0.16
0.15
0.15
0.15
0.14
0.14
0.13
0.13
0.14
0.12
0.12
0.10
0.11
0.11
0.10
0.08
0.08
0.07
0.07
0.07
0.06
0.05
0.05
0.04
0.04
0.03
0.03
0.01
Sweden
Switzerland
Finland
Denmark
United States
Norway
Canada
Germany
Singapore
Netherlands
New Zealand
Austria
United Kingdom
Australia
Ireland
Spain
France
Japan
Portugal
Hong Kong
Slovenia
Italy
Israel
Czech Republic
Hungary
Malaysia
Chile
Greece
Poland
Saudi Arabic
Tunisia
Korea
Panama
Slovak Republic
Argentina
Morocco
Botswana
Brazil
Thailand
Romania
Egypt
South Africa
Croatia
Sri Lanka
Turkey
Peru
Mexico
Venezuela
Colombia
Russian
Philippine
India
Iran
China
Indonesia
Ukraine
Senegal
Kenya
Pakistan
Uganda
Bangladesh
Nigeria
0.93
0.91
0.90
0.93
0.89
0.87
0.88
0.88
0.86
0.84
0.83
0.86
0.84
0.85
0.80
0.74
0.73
0.73
0.72
0.71
0.71
0.70
0.66
0.63
0.63
0.53
0.60
0.59
0.56
0.52
0.48
0.48
0.47
0.47
0.44
0.41
0.43
0.39
0.40
0.37
0.36
0.38
0.37
0.34
0.32
0.32
0.30
0.30
0.30
0.29
0.28
0.26
0.21
0.22
0.16
0.21
0.19
0.13
0.12
0.10
0.03
0.00
K.-W. LI
Copyright © 2011 SciRes. ME
193
IFS = International Financial Statistics, International
Monetary Fund;[40]
BOPS = Balance of Payment Statistics, United Na-
tions;[41]
UN = United Nations Co mtr ade, United Nations;[35]
ITU = International Telecommunication Union Data-
base, International Telecommunication Union;[42]
Net = Netcraft Secure, International Telecommunica-
tion Union;[43]
SSCT = Server Surveys Compendium of Tourism Sta-
tistics, World Tourism Organization;[44]
WFB = The World Factbook, Central Intelligence
Agency;[45]
OFW = Official websites of selected basket of treaties;
UNDPI = United National Development Program In-
dicators, United Nations;[46]
WDI = World Development Indicators, World
Bank;[47]
CI = Corruption Index 1996-2002, Transparency House;
[48]
AGI = Aggregating Governance Indicators 1996-2004,
World Bank;[49]
IEF = Index of Economic Freedom, Heritage Founda-
tion.[50]
There are few exceptions. For example, Hong Kong
has probably little international engagement in govern-
ment transfer and does not engage in financial contribu-
tion to the United Nations Security Council missions.
The few missing values in the country series are replaced
by the average of the immediate past and future years. In
the EFI, the maximum number of missing economies in
the 1998-2002 sample periods is 4, and their percentage
ranged between 5.9% and 11.8%. For the IFI, the cor-
responding figures for the maximum number of missing
economies are 40, and the percentage ranged between
5.9% and 35.3%. A complete set of data is obtained for
the three years of 1998-2001, while some data in 2002
are either provisional or unavailable. In the case of IFI,
the few provisional data of 2002 are replaced by the cor-
responding figures in 2001.