Working Paper Series
Working Paper No. 06-02
February 2006
How Competitive are Arab Economies?
By
Jay Squalli*
Kenneth Wilson*
Sarah Hugo*
* EPRU, Zayed University
How Competitive are Arab Economies?
Jay Squalli∗, Kenneth Wilson†and Sarah Hugo‡
Working Paper No. 06-02
Abstract
The recent publication of the Global Competitiveness Report 2005-2006 and the Arab World
Competitiveness Report 2005 by the World Economic Forum (WEF) has focused attention upon
the ability of Arab countries to compete in world markets. In the first report, Arab countries are
not well ranked compared to the rest of the world. The second Report highlights the gap that
exists between the more competitive Arab countries such as the UAE and Qatar, compared
to other Middle East and North Africa (MENA) countries. This paper provides an analysis
and evaluation of the approach taken by the WEF in constructing its Growth Competitiveness
Index (GCI). In particular, the paper has identified three areas where the GCI is vulnerable to
criticism. First, the treatment of outliers for hard data items. Second, the treatment of survey
data compared to hard data. Third, the arbitrariness of weight allocation used to construct
the GCI and its various indexes and sub-indexes. The paper suggests an alternative approach,
based upon Structural Equation Modeling, should be used for the determination of weights in
the index calculation process.
Keywords: World Economic Forum; Growth Competitiveness Index; Structural Equation Modeling;
Arab Economies.
JEL Classification: F10; F15.
∗
Economic & Policy Research Unit, Zayed University, P.O. Box 19282, Dubai, UAE, Phone: +971 4 208 2465,
Fax: +971 4 264 0394, E-mail: [email protected]
†
Corresponding Author: Economic & Policy Research Unit, Zayed University, P.O. Box 19282, Dubai, UAE,
Phone: +971 4 208 2470, Fax: +971 4 264 0394, E-mail: [email protected]
‡
Economic & Policy Research Unit, Zayed University, P.O. Box 19282, Dubai, UAE, Phone: +971 4 208 2591,
Fax: +971 4 264 0394, E-mail: [email protected]
1
1
Introduction
The recent publication of the Global Competitiveness Report 2005-2006 and the Arab World Competitiveness Report 2005 by the World Economic Forum (WEF) has focused attention upon the
ability of Arab countries to compete in world markets. These two reports are quoted as authoritative sources on global and regional competitiveness and provide rankings of countries according to
various competitiveness indexes created by WEF. In the first report, Arab countries are not well
ranked compared to the rest of the world. The second Report highlights the gap that exists between
the more competitive Arab countries such as the UAE and Qatar, compared to other Middle East
and North Africa (MENA) countries.
Interest in the economic performance of countries of the Arab world occurs for several reasons.
First, there is the strategic importance of many countries from the region, particularly because
of their role in the provision of hydrocarbons. Second, there is the problem of unemployment
experienced by many Arab countries and the challenges of creating millions of extra jobs in the
region in the coming years. Third, there is the historical importance of the region as a thriving
source of civilization, knowledge and innovation. Fourth, there is the prospect for greater regional
integration via the Gulf Cooperation Council (GCC) and the implications for economic growth
from such integration.
The aim of this paper is to look more closely at the question of Arab country competitiveness.
In particular, the paper aims to examine closely the analysis of Arab country competitiveness
undertaken by the WEF. The WEF assesses country competitiveness in terms of a series of indexes it
creates. These indexes cover such areas as a country’s macroeconomic environment, its technological
readiness and its public institutions. The WEF then uses these indexes to rank countries. Just
how competitive are Arab countries according to these indexes? Are the methodologies used by
WEF to determine country rankings reasonable? What are the policy implications that flow from
these Reports? These and other questions are addressed in this paper. This paper uses the original
WEF data to re-evaluate the calculation of the WEF’s competitiveness indexes. That is, the paper
assesses carefully the methods used by the WEF to construct the various indexes. The paper
2
is structured as follows. Section 2 addresses some important definitional issues, including the
definition and measurement of the Growth Competitiveness Index (GCI), the cornerstone measure
of country competitiveness developed by the WEF. Section 3 takes a closer look at how well the
various Arab countries did when measured by the GCI and provides a detailed assessment of the
measurement of the GCI and its component indexes. Section 4 raises questions concerning the
robustness of the GCI, particularly with respect to the treatment of outliers and the arbitrariness
of weight selection in calculating indexes. In this section, an alternative, less arbitrary method
for measuring the GCI based on structural equation modeling is introduced and some results and
analysis generated by the structural equation modeling are provided. Section 5 provides concluding
comments.
2
Definitions
Any analysis of Arab economies must confront the question: what is an Arab country? The International Monetary Fund (IMF) identifies 20 Arab states in the MENA region including: Algeria,
Bahrain, Djibouti, Egypt, Iraq, Jordan, Kuwait, Lebanon, Libya, Mauritania, Morocco, Oman,
Qatar, Saudi Arabia, Somalia, Sudan, Syria, Tunisia, the UAE, and Yemen. The IMF, in addition,
includes Afghanistan, Iran, Pakistan, and the West Bank and Gaza in MENA to bring the total to
24. Israel and Turkey are not included.
In this study, because of data limitations we are restricted to the following eight countries:
Bahrain, Egypt, Jordan, Kuwait, Morocco, Qatar, Tunisia, and the UAE. Unfortunately, these are
the only countries for which the WEF provide index data in the Global Competitiveness Report
2005-2006. Since this paper is assessing the competitiveness of Arab countries using the WEF data
we are therefore restricted to these eight countries.
The WEF produces a GCI for 117 countries in total including each of these eight Arab countries.
1
According to the WEF, its sample of 117 countries account for 97% of world output. The GCI is
the embodiment of the WEF’s view of what it means to be competitive. The WEF’s interpretation
of competitiveness may be summarized as follows (Lopez-Claros et al, 2005 p. 3): “We think of
3
competitiveness as that collection of factors, policies, and institutions which determine the level of
productivity of a country and that, therefore, determine the level of prosperity that can be attained
by an economy.” Therefore, according to the WEF view a more competitive economy should be
able to achieve greater economic growth in the medium to longer term. Hence the GCI is designed
to capture the underlying drivers of productivity that ensure sufficient and rising prosperity for a
country’s citizens. According to the WEF (Lopez-Claros et al, 2005 p. xiv):
The GCI brings together a number of complementary concepts aimed at providing
a quantified framework for measuring competitiveness. In formulating the range of
factors that go into explaining the evolution of growth in a country, it identifies “three
pillars”: the quality of the macroeconomic environment, the state of the country’s public
institutions, and, given the importance of technology and innovation, the level of its
technological readiness.
The GCI is constructed using a combination of hard data obtained from a range of independent
sources and survey data drawn from the WEF’s annual Executive Opinion Survey. Examples
of hard data include inflation rates and university enrollment figures. In contrast the Executive
Opinion Survey is able to help capture concepts for which hard data are usually unavailable, but
are typically important drivers of the economic growth process. Examples of information derived
from the Executive Opinion Survey cover such things as judicial independence and the extent of
inefficient government activities.
The GCI is broken down into three constituent indexes each representing one of the “three
pillars”: the Macroeconomic Environment Index (MEI); the Public Institutions Index (PII); and
Technology Index (TI). The WEF invoke several important assumptions in constructing these
indexes. First, they separate the countries into two categories: core innovators and non-core
innovators. Core innovators are the more technologically advanced countries. According to WEF,
technological innovation is more important to the economic growth of countries at or close to
the technological frontier. Therefore they classify countries as core innovators if technological
innovation is more critical for growth. To separate core innovators from non-core innovators they
4
count the number of US utility patents (patents for innovation) each country has per capita, for
the most recent year. Countries with more than 15 patents per million people are classified as core
innovators, while all others are classified as non-core. Therefore, to reflect this difference between
the core and non-core economies, the WEF uses a different formula to construct the GCI for each
group as follows:
1
1
1
Core-innovators GCI = T I + P II + M EI
2
4
4
1
1
1
Non-core innovators GCI = T I + P II + M EI
3
3
3
(1)
(2)
All eight Arab economies are classified as non-core innovators and hence the GCI for each Arab
country is constructed using the second formula. A second important assumption involves the construction of the TI. Whereas innovation is more important in core economies, the ability to absorb
and adopt foreign technologies will be more important to non-core economies. Hence technology
transfer will be more important in Arab countries than in core-innovators. This means that the
WEF constructs the TI differently for each of the core and non-core economies. For core economies
the TI is comprised of only the innovation sub-index (ISI) and the Information and communication
technology sub-index (ICTSI). For non-core economies, a third sub-index is included, the technology transfer sub-index (TTSI). Compare Figures 1 and 2. The WEF therefore uses a different
formula for each group as follows:
1
Core-innovators TI = ISI +
2
1
3
Non-core innovators TI = ISI + T T SI +
8
8
1
ICT SI
2
1
ICT SI
2
(3)
(4)
Since all eight Arab countries are classified as non-core, the second formula is used to calculate the
TI for each.
A third important assumption involves the weighting of hard data as compared to survey data
in the creation of the various sub-indexes. Hard data is always weighted more than survey data.
For instance in the calculation of the ISI the hard data items are weighted at 3/4 compared to the
survey data, which are weighted at 1/4. In the case of the ICTSI, hard data are weighted at 2/3,
compared to survey data at 1/3. In the calculation of the Macroeconomic stability sub-index, hard
5
data are weighted 5/7 compared to the survey data weight of 2/7. In each case the choice of the
weights is arbitrary. We return to the question of the allocation of weights in Section 5.
3
How Well did Arab Countries Perform on the GCI?
Table (1) contains the GCI scores and ranks for the eight Arab countries, as well as for the TI, PII
and MEI. It should be emphasized that the table includes both GCI scores and ranks. A higher
score means better performance in the various constituent components of the GCI. However, a high
score does not guarantee a low rank number. If all countries score highly then good performance
may not translate necessarily into a low rank number.
The UAE is the best GCI-ranked Arab country at 18 of 117, just ahead of Qatar at 19. The
UAE GCI score of 4.99 compares to a score of 5.94 achieved by Finland, the number one ranked
country. The theoretical maximum score is 7. Morocco is the least competitive Arab country with
a GCI score of 3.49 and a rank number of 76.
The UAE and Qatar achieve their relatively high scores and good rankings primarily because of
their macroeconomic environments. Of the 117 countries ranked by the WEF, the UAE and Qatar
are ranked respectively 5th and 6th on the basis of their macroeconomic environment. This means
there is virtually nothing that either of these countries can do to improve their macroeconomic
performance and achieve further improvement in their GCI. By contrast Morocco is ranked 67th
of 117, Egypt 55th and Jordan 52nd. Given that the MEI makes up one third of the GCI, any
improvement in the components of MEI will impact positively on the GCI score. To emphasize this
point, Table 2 provides a breakdown of the composition of the MEI into its constituent sub-indexes:
Macroeconomic Stability sub-index (MSSI); Government Waste measure (GW); and Country Credit
Rating (CCR).
In constructing the MEI, the MSSI is weighted at 50% compared to GW and CCR which are
each weighted at 25%. According to the MSSI the UAE and Qatar are ranked respectively 2nd and
4th of 117 countries. On the GW measure the UAE and Qatar are ranked 5th and 3rd respectively.
Hence there is virtually nothing that the UAE or Qatar can do to improve either their MSSI or
6
GW measure. With respect to CCR, the UAE and Qatar are ranked respectively 29th and 33rd.
However, these rankings have more to do with the regional location of the middle east, deemed to
be a risky region in international country-risk assessments. Again there is virtually nothing either
of these countries can do to improve their CCR.
Turning to the other MENA countries several important lessons can be gleaned from Table (2).
Morocco’s high MEI rank number of 76 is primarily the result of its relatively poor MSSI score
of 4.09 and related rank of 88th position. Likewise Jordan and Egypt have considerable room for
improvement in their MSSI. Only Morocco performs poorly on the GW measure, being ranked
54th. Table (2) is useful in confirming that Morocco, Egypt and Jordan need to further improve
their macroeconomic stability to realize further gains in international competitiveness as measured
by the GCI.
Table (3) provides details for the eight Arab countries on the composition of the TI. The UAE
is the best TI-ranked Arab country at 33rd. Morocco is the worst TI-ranked Arab country at 78th.
The TI is comprised of three sub-indexes: Innovation sub-index (ISI); ICT sub-index (ICTSI);
and Technology transfer sub-index (TTSI). Although Arab countries are ranked relatively poorly
on ISI this has a minimal impact upon the TI since ISI is weighted only 1/8 in calculating TI.
ICTSI is much more important to TI being weighted at 50%. Hence, there is much room for
improvement for Morocco which is ranked 83rd on this sub-index with a very low score of 2.07.
Again we see that most Arab countries are ranked relatively poorly on ICTSI and this explains
the subsequent low scores and relatively poor ranks on TI. According to these data, the single
most important thing that Arab countries could do to improve their TI, and hence their GCI, is
to improve their information and communications infrastructure. On technology transfer (TTSI),
several Arab countries do relatively well, particularly Qatar, UAE, Egypt and Bahrain.
Table (4) contains scores and ranks for PI for the eight Arab countries. Only two sub-indexes,
each weighted at 50%, are used to construct the PI: the Contracts and Law sub-index (CLSI) and
the Corruption sub-index (CSI). We see that the relatively good PI for Qatar is due primarily to
its good CLSI score. Whereas the UAE enjoys a relatively good PI because of its relatively low
7
level of corruption as measured by CSI. The low PI score and poor rank of Morocco is due mainly
to its very low CSI score though it also performs worse than all other Arab countries in the sample
on CLSI.
In summary, the evidence from the tables presents a picture of mixed Arab-country performance.
Generally, the UAE and Qatar perform well on most indexes, whereas Morocco performs relatively
poorly. For most Arab countries, there is considerable room for improvement in the drivers of
growth competitiveness.
4
How Robust is the GCI?
It is clear that the WEF is attempting to provide advice to governments, business leaders and others
about the relative economic growth environment of as many countries as they can. Their aim is
to identify those countries with the right macroeconomic environment, technology readiness and
economic-institutions in place that enhance economic growth, while also identifying those countries
that fall short of best practice. To that end the calculation of the GCI and its component indexes
and sub-indexes has merit. However, since these index scores are used to rank countries and create
league tables then a closer look at the technical aspects of exactly how the various indexes are
constructed must be undertaken.
There are three areas that must be considered more closely before we can decide exactly how
robust is the GCI and its related indexes and sub-indexes. The four areas are: (i) the treatment
of outliers in hard data variables; (ii) the treatment of survey data compared to hard data; and
(iii) the arbitrariness of weight allocation used to construct the GCI and its various indexes and
sub-indexes.
4.1
Treatment of Outliers
In the case where indexes and sub-indexes are constructed from survey data only, there is no
potential outlier problem since all variables are constrained to the 1-7 range by the nature of the
Executive Opinion Survey design. However, in the case of some hard data variables, it is possible
for some values to be well outside the standard normal range. In such cases a careful assessment
8
must be made concerning whether these outliers distort the calculation of the individual country
scores. For the ease of calculation of indexes, the WEF convert all hard data items to the 1-7 scale
using the following formula after controlling for outliers:
6×[
(country value − sample minimum)
]+1
(sample maximum − sample minimum)
(5)
The treatment of outliers is important since it will affect the actual scores calculated using equation
(5). For example, the decision to drop countries from the top or bottom end of the distribution
will effectively increase the scores of all other countries when hard data are converted to a 1-7 scale
score using equation (5). Given that these scaled scores are then used to calculate the indexes and
that hard data are weighted more heavily than survey data, the decision to include or exclude an
outlier can have an important impact.
But just how does the WEF control for outliers? Unfortunately, the WEF do not reveal exactly
how they deal with outliers for each of the 14 hard data variables used in the construction of the
indexes. However, through a pains-taking process of backward engineering we have been able to
determine exactly how they dealt with outliers in each of the 14 cases.
In constructing the indexes, the WEF appears not to have been consistent in the way they
treat outlier issues related to hard data items. For any researcher, the treatment of outliers is not
a straight-forward matter and there is no single universally agreed method to adopt. Rather, it is
more a case of common sense and reasonable practice. A popular method, beyond simply eye-balling
the data, is to apply Grubbs’ test. However, the definitive use of Grubbs’ test requires that the
data be normally distributed, which is certainly not the case for several of the hard data variables
used by the WEF. Of the 14 hard data items in only six cases did we not detect outliers and find
no adjustment by the WEF, namely: telephone lines, tertiary enrollments, personal computers,
internet users, cellular phones, and country credit rating. Although we did not detect an outlier in
the case of real effective exchange rate, we observed that the WEF dropped Zimbabwe. In several
cases, the WEF made unusual adjustments, such as in the case of government surplus/deficit and
utility patents.
There are four cases where an alternative approach to that adopted by the WEF could have
9
been used. In the cases of government surplus/deficit, inflation, utility patents, and internet users,
a different outlier treatment could have been used. If this alternative approach had been used in
each of these cases, then the resulting index scores would have been different.
4.2
An Alternative Approach to Determining Index Weights
One of the most surprising features of the calculation of the various indexes is the ad hoc and
unjustified way the various weights are chosen and assigned to variables given their importance to
the calculation of each index and sub-index and also the fact that the WEF consistently weight
hard data items more heavily than survey data, except in one case.
2
Because the created indexes are latent variables which are indirectly represented by observed
variables, it becomes important to select variables that are theoretically and intuitively appealing.
The assignment of ad hoc weights coupled with an ad hoc selection of observed variables cast
doubt on the reliability of the GCI and corresponding sub-indexes. In what follows, we present an
alternative approach of assigning weights to the latent sub-indexes in representing the GCI index.
The procedure is completed via structural equation modeling (SEM). SEM exhibits properties
similar to multiple regression modeling, but more robustly takes into account relationships between
(a group or sub-group of) latent variables, including correlations, covariances, nonlinearities, and
error terms (correlated and uncorrelated). SEM also allows for a set of observed variables to
indirectly represent a composite latent variable in a relatively more objective fashion.
While SEM can play many roles, in this paper it is strictly confirmatory. This is particularly
important as the structural path model proposed by the WEF can be confirmed (or dis-confirmed)
by determining whether variances and covariances in the data exhibit patterns that are consistent
with the WEF’s assumptions and findings. If the WEF is dis-confirmed, then a new set of weights
can be derived which ultimately lead to new (more robust) rankings.
3
SEM produces regression estimates which measure the degree of causal importance in explaining a particular latent variable, the higher the coefficient estimate the more important a particular
observed or latent independent variable is at explaining a corresponding latent dependent variable.
These regression coefficients can also be used in a more compelling way to determine the different
10
weights that are assigned to independent variables in explaining a dependent latent variable. This
can be done by simply summing up the different λs in a system of equations with the same independent variable, and assigning the ratios as weights.
4
For instance, for non-core innovators, SEM
estimations, as shown in Figure (1), have assigned the following values: λ1 = 0.64, λ2 = 0.77, and
λ3 = 0.37. The weights are therefore calculated as:
λ1
w 1 = P3
0.64
= 0.36
1.78
i=1 λi
λ2
0.77
w 2 = P3
=
= 0.43
1.78
i=1 λi
0.37
λ3
=
= 0.21
w 3 = P3
1.78
i=1 λi
=
(6)
where w1 , w1 , and w3 represent the weights assigned respectively to TI, PII, and MEI in the
composition of the GCI. Table (5) provides a comparison of the arbitrary weights used by the
WEF for core and non-core innovators against the weights generated by SEM. There are substantial
differences in every case and some are quite large.
For non-core innovators, the WEF weights each of TI, PII, and MEI at 0.33, SEM generates
respective weights of 0.36, 0.43, and 0.21. In the case of non-core innovators, the WEF identify
three drivers of technology represented by the ISI, TTSI, and ICTSI. However, because TTSI was
perfectly correlated with the survey data items of ISI, these variables had to be dropped from the
SEM. This means that the separation of the countries into core and non-core innovators and the
exclusion of technology transfers variables from the calculation of the core innovators’ GCI is an
unnecessary contrivance. The perfect correlation detected by SEM emphasizes that technology
transfer impacts are captured perfectly by the survey items used in the construction of the ISI.
As a consequence, we produce a third set of SEM weights whereby we combine all countries into
a single group and generate weights using the model structure for core innovators. These weights
are contained in the final column of table (5). The weights for TI, PII, and MEI are, respectively,
0.44, 0.34, and 0.22.
The use of SEM enables the calculation of a new set of alternative GCI scores and ranks. Table
(6) presents three sets of GCI scores and ranks for the eight selected Arab countries. First, the
11
WEF scores and ranks are reproduced. Second, scores and ranks generated by SEM after separating
countries into core and non-core innovators categories are included. The third set of scores and rank
orders are generated by SEM when all countries are combined into a single data set and treated
equally. It is this third set of results that we believe represents the most accurate set of GCI scores
and rank numbers. According to this approach, the UAE’s GCI score falls from 4.99 to 4.42 and
its rank order falls from 18th to 29th. Qatar falls from 4.97 to 4.27 and from 19th to 35th. All
Arab countries fall although some only marginally. The worst performing country, Morocco, falls
from 3.49 to 2.94 but holds its position at 77th. There is still considerable diversity amongst Arab
countries and relativities between them remain unchanged by the use of SEM to calculate weights.
4.3
Latent Variable Reliability Analysis
Because of the inevitable heteroskedasticity in the data and potential presence of outliers, it is
important to determine the extent to which the selected variables are related to each other. Since
causal relationships between observed variables and a single latent variable cannot be established a
priori, the validity of parameter estimates derived by structural equation modeling may be biased
because of a potential poor fit between the variables used. Thus, in order to derive latent variables
that are based on a plausible choice of observed variables, it becomes crucial to test for internal
consistency of the measurement scale and the identification of unsuitable items. Internal consistency
can be tested using Cronbach’s alpha which is expressed as:
α=
V c̄
1 + (N − 1)c̄
(7)
where V represents the number of variables and c̄ represents the average inter-variable correlation.
According to the social science literature, internal consistency is supported when α > 0.80. It is
evident that as V rises, α rises as well. Furthermore, when c̄ is low, then α is also low.
As summarized in Table (7), internal consistency is established for the survey data variables
used in deriving GCI sub-indexes for non-core innovators. This may be due to the nature of the
data and the strong correlation that exists among the different variables. In fact, the survey data
variables explaining the MSSI fit marginally well with α = 0.67. Hard data, on the other hand, are
12
problematic as in five out of the times they are used, internal consistency is violated. In fact, the
hard data variables only fit well together for the ICTSI with α = 0.90. In the other cases, α values
range between 0.04 and 0.53. Moreover, when survey and hard data variables are used together,
then Cronbach’s alpha is driven down by the poor fit of hard data variables.
The arbitrary selection of variables in the composition of the different growth competitiveness
indexes and sub-indexes is clearly put into question. The poor fit of hard data variables indicates
that the variables selected in representing the ISI and MSSI for non-core innovators cannot be
reasonably combined. Rather, they have to be treated independently of each other, with each hard
data variable representing its own sub-index, thus making the task of determining weights, scores,
and ranks more challenging. It is also interesting to note that the WEF weights the hard data
variables more heavily than survey data. Although this approach is intuitively appealing due to
the higher objectivity of hard data, it fails to meet internal consistency requirements. This casts
further doubt on the contrived procedures adopted by the WEF in selecting data variables and
determining the proper weights assigned to each sub-index and corresponding observed variables.
5
Concluding Remarks
The WEF is an influential body that produces its Global Competitiveness Report yearly. In addition, it recently published a companion volume focusing directly on Arab countries. The purpose
of these volumes is to determine the extent to which Arab countries are competitive with the
rest of the world. To analyze and evaluate international country competitiveness, the WEF produces a comprehensive GCI which embodies three pillars of international growth competitiveness:
technology, macroeconomic environment, and public institutions.
The purpose of this study has been to evaluate carefully and comprehensively just how well
Arab countries perform according to the GCI and its various indexes and sub-indexes. Generally
speaking, except for the UAE and Qatar, Arab countries, particularly those of North Africa, are
not internationally growth competitive.
This paper also provides an analysis and evaluation of the approach taken by the WEF in
13
constructing its GCI. In particular, the paper has identified three areas where the GCI is vulnerable
to criticism. First, the treatment of outliers for hard data items. In several cases, we have identified
alternative methods for dealing with outliers. Second, the crucial role of the variable utility patents
in the calculation of the GCI is questioned and serious doubts concerning the use of this variable
are raised. Third, the paper suggests an alternative approach, based upon SEM, should be used for
the determination of weights in the index calculation process. SEM is a robust statistical approach
and overcomes the arbitrariness of the weight selection method used by the WEF.
Although there is merit in the approach of the WEF to assessing growth competitiveness of
individual countries, this paper shows that considerable care needs to be taken when undertaking
such analyses, particularly given the political sensitivity associated with creating league tables of
country rank order. Nevertheless, the calculation of the GCI and related index and sub-index scores
provides considerable information on the constraints to growth competitiveness that policy makes
in the various Arab countries in this survey could gain benefit from.
Notes
1. The list of the 117 countries is contained in Squalli et al. (2006) Table A1.
2. That one case involves the equal treatment given to CCR (hard data) and GW (survey data)
each weighted equally at 1/4 in the calculation of MEI.
3. For details on the structure of SEM used to calculate the weights for the WEF’s GCI path
dependency process and a more complete discussion of the material in this section, see Squalli et
al. (2006).
4. In the calculation of weights using SEM, we have used the data generated by WEF. That is we
have used their outlier adjusted method and we have used U.S. utility patents to emphasize the
point that irrespective of data issues, the WEF choice of weights is totally arbitrary.
14
GCI
O1
0.64
O2
TI
O11
0.67
ISI
O3
0.77
0.37
PII
O12 1.00 O21
ICTSI
MEI
O22 1.00
0.71
CLSI
CSI
O31 0.92
O32
MSI
O33 1.00
3.23
CCR
Figure 1: Structural Equation Modeling for Non-core Innovators
15
GW
Table 1: Growth Competitiveness Index of Arab Economies
Country
Bahrain
Egypt
Jordan
Kuwait
Morocco
Qatar
Tunisia
UAE
GCI
Rank Score
37
4.48
53
3.96
45
4.28
33
4.58
76
3.49
19
4.97
40
4.32
18
4.99
Rank
41
58
52
48
78
40
60
33
TI
Score
3.73
3.36
3.46
3.56
2.96
3.76
3.35
4.04
PII
Rank Score
38
5.10
53
4.46
31
5.28
37
5.11
85
3.69
19
5.75
40
5.02
24
5.52
MEI
Rank Score
32
4.62
55
4.07
52
4.10
21
5.09
67
3.82
6
5.40
34
4.59
5
5.43
GCI represents the Growth Competitiveness Index, TI represents the Technology Index, PII represents the Public
Institutions Index, and MEI represents the Macroeconomic Environment Index.
Table 2: Macroeconomic Environment Index of Arab Economies
Country
Bahrain
Egypt
Jordan
Kuwait
Morocco
Qatar
Tunisia
UAE
MEI
Rank Score
32
4.62
55
4.07
52
4.10
21
5.09
67
3.82
6
5.40
34
4.59
5
5.43
MSSI
Rank Score
20
5.11
59
4.47
57
4.50
1
5.72
88
4.09
4
5.66
43
4.65
2
5.70
GW
Rank Score
36
3.68
34
3.75
23
4.03
38
3.63
54
3.23
3
5.13
7
4.86
5
5.00
CCR
Rank Score
47
4.58
64
3.57
70
3.39
31
5.26
58
3.87
33
5.17
52
4.20
29
5.32
MSSI represents the Macroeconomic Stability Sub-Index, GW represents Government Waste, and CCR represents
Country Credit Rating.
16
Table 3: Technology Index of Arab Economies
Country
Bahrain
Egypt
Jordan
Kuwait
Morocco
Qatar
Tunisia
UAE
Rank
41
58
52
48
78
40
60
33
TI
Score
3.73
3.36
3.46
3.56
2.96
3.76
3.35
4.04
ISI
Rank Score
52
2.45
64
2.36
47
2.57
69
2.23
93
1.77
70
2.22
57
2.41
44
2.67
ICTSI
Rank Score
42
3.08
68
2.33
53
2.66
45
3.01
83
2.07
44
3.02
56
2.61
34
3.53
TTSI
Rank Score
19
5.01
14
5.07
31
4.82
33
4.73
38
4.54
7
5.25
34
4.64
10
5.16
ISI represents the Innovation Sub-Index, ICT represents the ICT Sub-Index, and TTSI represents the Technology
Transfer Sub-Index.
Table 4: Public Institutions Index of Arab Economies
PI
Country
Bahrain
Egypt
Jordan
Kuwait
Morocco
Qatar
Tunisia
UAE
Rank
38
53
31
37
85
19
40
24
Score
5.10
4.46
5.28
5.11
3.69
5.75
5.02
5.52
CLSI
Rank Score
39
4.54
45
4.37
26
5.05
31
4.91
60
3.76
15
5.52
28
4.99
38
4.78
CSI
Rank Score
34
5.65
67
4.55
40
5.51
45
5.31
98
3.63
27
5.98
51
5.04
16
6.25
CLSI represents the Contracts & Law Sub-Index and CSI represents the Corruption Sub-Index.
17
Table 5: Weights of GCI Index and Sub-Indexes
Path
GCI → TI
GCI → PII
GCI → MEI
TI → ISI
TI → ICTSI
PII → CLSI
PII → CSI
MEI → MSSI
MEI → CCR
MEI → GW
Core
WEF SEM
0.50
0.34
0.25
0.17
0.25
0.49
0.50
0.43
0.50
0.57
0.50
0.61
0.50
0.39
0.50
0.27
0.25 0.10*
0.25
0.63
Non-Core
WEF SEM
0.33
0.36
0.33
0.43
0.33
0.21
0.125 0.40
0.50
0.60
0.50
0.41
0.50
0.59
0.50
0.18
0.25
0.63
0.25
0.19
Core & Non-Core
SEM
0.44
0.34
0.22
0.47
0.53
0.48
0.52
0.15
0.62
0.23
Because of perfect correlation between the TTSI and ISI, the former is dropped from the estimations of the weights
for non-core innovators. An asterisk * denotes a weight that is derived from a statistically insignificant coefficient
estimate of λ32 .
Table 6: Comparative Rankings of Arab Economies
Country
Bahrain
Egypt
Jordan
Kuwait
Morocco
Qatar
Tunisia
UAE
WEF
Score Ranking
4.48
37
3.96
53
4.28
45
4.58
33
3.49
76
4.97
19
4.32
40
4.99
18
SEM with different weights
Score
Ranking
4.20
40
3.56
59
4.02
44
4.24
37
3.09
78
4.56
30
4.00
45
4.69
27
SEM with same weights
Score
Ranking
3.94
42
3.37
57
3.77
46
3.99
41
2.94
76
4.27
35
3.79
45
4.42
29
The scores and rankings under ’SEM with different weights’ are calculated using the weights derived from SEM and
in which the weights used for core innovators are different from those for non-core innovators. However, the scores
and rankings under ’SEM with same weights’ represent those derived through SEM when the GCIs for core and
non-core innovators are assumed to be the same.
Table 7: Cronbach’s Alpha for Non-Core Innovators
Latent Variable
Innovation Sub-Index
Technology Transfer Sub-Index
ICT Sub-Index
Contracts & Laws Sub-Index
Corruption Sub-Index
Macroeconomic Stability Index
Soft Data
0.88
0.79
0.84
0.88
0.95
0.67
18
Hard Data
0.04
Soft and Hard Data
0.75
0.90
0.90
0.39
0.53
References
[1] Fortune, J. and D. R. Utley (2005) “Team Progress Checksheet: A Study Continuation.”
Engineering Management Journal 17, (2005): 21-27.
[2] Lopez-Claros, A., M. E. Porter and K. Schwab (Eds.). The Global Competitiveness Report
2005-2006 (World Economic Forum, 2005).
[3] Lopez-Claros, A. and K. Schwab (Eds.). The Arab World Competitiveness Report (World
Economic Forum, 2005).
[4] Squalli, J., K. Wilson & S. Hugo. An Analysis of Growth Competitiveness, Economic & Policy
Research Unit Working Paper No. 06-01, Zayed University, 2006.
19