Introduction

New regionalism is the point of consensus among development strategies of conflicting economic doctrines. In the dominant neo-liberal paradigm, regionalism is assumed as a crucial phase in the transformation of the international economic system to globalization. The alternative paradigm also considers regionalism as the a point of departure for developing countries to alleviate the hegemonic pressures of capitalism and increase their bargaining power and as an opportunity for the formation and evolution of an alternative economic system. In the meantime, the organization of Islamic conference (OIC) has also viewed the regionalism as a suitable strategy for economic development by proposing the formation of an Islamic common market.

In line with this strategy, this paper deals with the feasibility study of regionalism in the MENA region, the region where Abrahamic religions and the glorious Islamic civilization, -the superior and an unparalleled civilization of the middle ages- were born and grown. However, the industrial revolution and the post Renaissance environment brought about by the West’s scientific boom and boost on the one hand, and the incompetence of retrogressive Muslim rulers of the time on the other hand created a deep gap between the region and the Europe – and later on- the US. The geo-strategic importance of the region and the need for the revival and reconstruction of the Islamic civilization were the main reason behind the selection of the regions member countries as the statistical population.

In this paper, the Linder theory forms the underlying part of the feasibility study of establishing a regional free trade area – and an Islamic common market, in the later stages- as the process of globalization is going ahead. In contrast with the classical trade theories (such as the Heckscher- Ohlin Theory), the Linder theory focuses on the demand side and explains the trade patterns on the basis of the similarities in the demand structures. As the similarities of demand structures largely depends on the convergence of per capita income, the following two hypotheses are simultaneously tested:

Globalization brings about convergence of per capita income trends in the MENA region, while the North- South per capita incomes demonstrate divergent trends. Theoretically, this hypothesis is indebted to Matsuyama’s (1996) symmetric breaking and Deardroff’s (1998) multi cone theories and, in general, suggests the presence of global divergence and hemispherical convergence. The hypothesis test is performed using the difference in differences method.

The Linder theory adequately explains the trade behaviors of the MENA countries with the Muslim countries and the rest of world. That is to say; with global divergence and hemispherical convergence, the MENA’s intra-regional trade will increase and the inter-regional will decrease.

Hence, the feasibility study of regionalization, and the formation of a common market will be carried out in the three steps:

i. The impact of globalization on the convergence or the divergence of per capita incomes across MENA’s countries;

ii. The impact of globalization on the convergence or the divergence of per capita incomes across Northern and southern countries, and

iii. An Empirical Test of Linder’s hypothesis to examine the trade patterns of regional countries before and after the trade liberalization.

Together with Slaughter (2001) studies, that support the divergence of the per capita income trends due to globalization across the world, this paper asserts that globalization paves the way for regionalization and the formation of an Islamic common market in the MENA region.

The table below lists the results of all possible scenarios for the above two hypothesis tests. The first column corresponds to the first hypothesis; the second, to Slaughter’s studies; the third and fourth, to the results of the second hypothesis test.

If the global divergence and the regional convergence were confirmed and the Linder theory could explain the trade behaviors of the region with the Muslim world and other countries, we can expect that globalization will help make the establishment of a common market more feasible. The third (second) scenario illustrates the best (worst) situation in the subject under discussion.

AS Elahi and Nahanvdian (2005) have confirmed the convergence of per capita income trends in the region using DID method and Slaughter (2001) has verified the North-South divergence using the same method, there is no need to test the first hypothesis. It means that the two first stages –global divergence and the hemispherical convergence- of feasibility study of regionalization and the establishment of an Islamic common market have been accomplished. However, to finalize the feasibility project we just need to test the second hypothesis.

The paper consists of two parts: the fist part explains Linder’s theory and the second part tests the relevance of this theory in the trade behaviors of the Middle East and North African countries. In conclusion, considering Slaughter (2001) and Elahi Nahavandian (2005) studies on the one hand and the results of tests of Linder’s theory on the other hand, the feasibility of regionalization in countries under study is examined

Table 1- the possible Scenarios

Scenario Per Capita Income The Test of Linder’s Hypothesis Result

MENA North and South Muslim Countries The Rest of World Intra-regional Trade Inter-regional Trade Feasibility Study

1 Convergent Convergent Verification Verification Increase Increase No Preference

2 Divergent Convergent Verification Verification Decrease Increase Impossibility

3 Convergent Divergent Verification Verification Increase Decrease Highly Possibility

4 Divergent Divergent Verification Verification Decrease Decrease No Preference

5-6 Convergent/ Divergent Convergent Rejection Verification No Judgment Increase –

7-8 Convergent/ Divergent Divergent Rejection Verification No Judgment Decrease –

9-10 Convergent Convergent/ Divergent Verification Rejection Increase No Judgment Likely possibility

11-12 Divergent Convergent/ Divergent Verification Rejection Decrease No Judgment –

13-16 Convergent/ Divergent Convergent/ Divergent Rejection Rejection No Judgment No Judgment –

I. Linder’s Theory

1-1. The Background

Prior to 1960, trade theories were based on supply side and the Heckscher–Ohlin as the most popular theory placed emphasis on the factor endowment as the determinant of trade model and the relative advantage. According this theory labor abundant country should specialize in production and export of labor-intensive goods. They should import the needed capital-intensive commodities from countries with higher per capita income.

The theory was based on a number of assumptions including the similarity in consumption patterns and technology, the existence of constant return to scale and a competitive condition and the irreversibility of factor intensity in the countries involved. In addition to providing a fair explanation of relative advantage -on the basis of factor endowment – this theory ensures the absolute factor price equality drawn upon Samuelson’s contribution. Leontief’s tests (1954 and 1956) on American exports and imports revealed that the US imports was surprisingly capital-intensive goods, whilst, US per capita capital was greater than any other country. Baldwin (1971) also states that this paradox continues to exist. He found that US imports are 27% more capital intensive.

Tatemoto and Ichimura’s studies (1959) revealed that this paradox exists in Japan too, -though in another form. Japan is a labor abundant country compared to advanced countries but a capital abundant country relative to other nations. Yet Japan’s export was found to be capital intensive commodities and her imports were labor intensive goods. Whal (1961) also studied the Canadian trade pattern and observed that although the trade with US accounted for most of Canada’s trade relation and US was a capital abundant nation Canada’s export was relatively capital intensive . Bharawaj (1962) also discovered the irrelevance of HO theory in the Indo-US trade relations

In general, one can argue that the Leontief paradox was a turning point in the development of new trade theories. Leontief and others attempted to present explanations to justify the inconsistency of the Heckscher Ohlin theory with the trade patterns of the countries under their study, by raising the difference in US labor productivity, the factor intensity reversal, demand biasness, the abundance of natural resources in US, transportation and tariff costs, and negligence of Heckscher–Ohlin to Human capital. However, Daniel Trefler (1993) stresses that in view of the excessively dispersed factor prices in different countries it is naïve to talk of factor price equalization. Moreover, the existence the North-North trade shows that HO is incapable of explaining the trade behavior and the findings of empirical studies have repeatedly confirmed the inconsistency of HO theory with these findings.

In search for a solution to the irrelevance with HO theory with empirical studies, several efforts have been made to reform and reformulate the theory in such a way that the core idea in the theory is saved. Hence, different versions such as HOV and HOC versions of this theory have been presented to save the essence of the theory even if it results to violation of some assumptions. However, the achievement seems to be inconsiderable.

1-2. Linder’s Theory and Its Elements

Linder has examined the trade behavior from quiet a different approached. His approach is usually identified with the demand side. With this approach, it seems that he has succeeded to develop a theory with grater consistency with the statistical truths obtained by Leontief tests and similar tests. However, as Leamer and Levinsohn (1995: 1383) have written, ‘while Linder did not have a formal model, he had a compelling story‘. According Fillat-Castejon and Serrano-Sanz, this approach has caused his theory to enjoy a high degree of success. In their view, Linder’s theory has done better in explaining the trade behavior „Linder considers potential trade to be explained by the so-called ‘trade-creating forces‘, whilst certain ‘brakes‘ will deviate real trade away from its potential, with the pattern of trade and the trading partners of each country being determined by this conjunction of trade creating and braking forces.“

Linder (1961) challenged some beliefs on the theory of international trade at that time, particularly the HO theorem. According to HO, relative endowments of productive factors theory provided an explanation for the model, which placed emphasis on the differenced goods in terms of factor intensity and countries in terms of factor endowment. Therefore, trade of capital-intensive goods with labor-intensive goods between capital abundant and labor abundant countries should be established. While trade exchanges were largely established in goods of similar characteristics and between countries with comparable levels of development enjoying a very high rate of growth. In fact, Linder’s theory tried to provide an answer to these two aspects, namely the pattern of trade and the trading partners.

Another novel point in Linder’s theory is the emphasis it places on the dynamic aspects of the relationships between trade and development. The growth experienced by a country modifies its demand structure and, thereby, the range of both potential and real exports, explaining how the pattern of trade changes over time. Within this scope of potential trade, actual trade is determined from a set of factors that tend to strengthen it, the so-called trade creating forces, and others which tend to limit it, the so-called trade braking forces. All these offer an underlying theoretical basis for prediction of trade behaviors. Those in greater demand within the country will be exported—the so-called expansion thesis—while those in less demand will be imported. This approach involves a form of a priori reasoning which illustrates the intra-industry trade and Linder’s approach. Gray (1988) considers Linder’s approach as a key element in the intra-industry trade a paradigm‘.

1-2-1. The Trade-Creating Forces

As stated above and unlike supply side theories of trade, Linder turned his attention towards demand when seeking to explain trade. According to his thinking, the demand characteristics of two countries would act as decisive factors in explaining potential trade, and it is Linder’s core idea that has developed a significant part of the subsequent literature. In this theory, monopolistic competition is considered to be as a possible factor in the growth of intra-industry trade.

The relationship between demand and international trade can be established in two ways, that is to say, through the complementarity of the demand structures of two countries and through the degree of representativity of the demand for common products. In the first way, emphasis is placed on the trade for the satisfaction of necessities. The second approach emphasizes on the characteristic of demand.

In both methods, the production substantially occurs for the satisfaction of internal necessities of a country and – in view of the production possibilities – the surplus to internal needs are exported to countries with similar demand structures.

Linder rightly argues that the determinants of the Demand structure are the modal or median per capita income and the average per capita income is not a good determinant of demand structure – particularly for countries with a income dispersion. However, its too difficult to get the per capita income distribution of different countries hence, income average is used to determine the demand structure (Linder, 1961:94). To justify the role of per capita income in a country’s demand structure, Linder draws on the concept of the income elasticity of demand (Linder, 1961: 94).

As it is deduced from Engel’s law, by increasing per capita income, higher quality and luxury goods are regarded as necessity and former necessity fall in inferior goods basket resulting in an increase in the demand for luxury goods. With respect to the degree of representativity of demand, when a country’s production potential is greater, the probability of exporting will be higher otherwise; these demands will be satisfied with imports.

An original point in this analysis is that the trading countries enjoy a similar level of demand structure and hence similar per capita income distribution.

So to speak, one can claim that the closer the per capita income average the higher the possibility of occurrence of trade. Theoretical developments in the analysis of demand using models inspired by Linder have concentrated on three topics. (Fillat-Castejón and Serrano-Sanz, 2004: 326-7) the association between the level of income and the demand for quality, on the basis of consumer preferences expressed in terms of the characteristics of the goods and not just in terms of quantities. This approach allows us to explain why economic growth leads to a higher horizontal differentiation of products and to an increase in the average quality or sophistication that is demanded. There are non-homothetic preferences and the growth in income affects the demand for different goods in different ways, giving rise to structural changes. Markusen (1986) tested Linder’s observation that people with similar per capita incomes consume similar sets of goods. Non-homothetic preferences, which in Markusen’s analysis have taken the form of an assumption, are formalized and empirically tested in Hunter and Markusen (1988). According to this logic, the change in the structure of demand will have implications over the composition of trade, in that the greater the non-homothetic nature of demand, the more intense will be the trade between two countries with similar per capita incomes. The distribution of income and preferences within countries is an essential point when considering the possible overlapping of demands and defining the varieties or qualities of a good to be traded. The usual models of international Trade neglect the details, but Linder’s ideas allow us to be more exact. Hence, in countries with an even income distribution and with a similar level of per capita income, we expect that the overlapping of demand increases.

However, with an even per capita income and an uneven distribution around an average, a range of qualities will be demanded for each type of product and a form of vertical differentiation, will emerge. When the income is concentrated at higher level better quality products will be traded otherwise lower quality product will gain greater significance. Henceforth, dispersion of income distribution will exert an influence over trade in an aggregate form, and by way of the range of varieties that are susceptible to trade. However, as mentioned above access to dispersion of income of countries and fitting them in model is beyond the scope and space of this paper.

According to Linder (1961: 110), small sized countries establish greater trade with larger ones rather than smaller countries. Therefore, product differentiation is another trade-creating force, although this aspect was hardly developed in his work. However, the volume of trade is positively associated with the size of economy and market. It has subsequently received a great deal of attention, above all in relation to the size of the market. Moreover, this is the reason of greater trade exchanges between small sized -large sized countries compared to small- small countries. Other studies carried out on the basis of Linder ‘s theory suggest that size also conditions the possibilities of diversification and manifests itself in volume and specialization; that is to say, it has not only a quantitative but also a qualitative influence.

For example, Keesing (1968) demonstrated how the larger size of a country led to a higher exports and lower imports, whilst both depended positively on income. Balassa Balassa, B. and Bauwens, L. (1988) confirmed the need for large internal markets for the export of manufactures, due to scale economies, consequently, large countries find themselves in an advantageous situation. Fillat-Castejón and Serrano-Sanz (2004) given the higher income-elasticity of industrial goods, the exports of large countries at any level of per capita income will be systematically biased towards industry in comparison with the average of small countries. , Perkins and Syrquin (1989) observed that large countries present exports which specialize in manufactures, whilst the exports of small countries are specialized in minerals.

These studies along with an assessment of the possible influence of size in specialization leads us to propose the hypothesis that, with respect to each level of per capita income, size causes the country of reference to export standardized goods to and import differentiated goods from its small sized trading partners, and to export differentiated goods to and imports Standardized goods from its large sized trading partners. Hence, as the size of the trading partner grows, exports are stimulated and imports inhibited in the differentiated product sectors of the reference county, while the opposite occurs with standardized products. Thus, the greater the difference in size of economy the greater the potential for trade. Specifically, product differentiation is not relevant.

1-2-2. The Trade-Braking Forces

Brakes on trade are considered as those factors that cause real trade to deviate from potential trade. The three factors explicitly recognized by Linder are the use of scarce factors in the demanded goods, distance and human-made trade obstacles. The use of scarce factors is the main connecting point with the Heckscher–Ohlin theorem; Linder argued that the intensive use of a scarce factor in a variety included within the overlapping of demand causes the efficiency to decrease and the trade to deviate and imposes deadweight loss. This loss arises from implying the expensive domestic factor of production instead of cheap foreign factors of production. Thus, the main source of difference between Linder’s and HO theory is the latter’s emphasis on the endowment of productive. In other words, this theory is developed with a supply side approach whilst, in Linder’s demand side approach overlapping of demand shapes the pattern of trade. However, the use of scarce use of a scarce factor in a variety included within the overlapping of demand is recognized as a trade breaking force; that is to say, that what is considered as an opportunity in HO theory its absence is thought to be an obstacle to trade in Linder’s theory. As far as distance is concerned, firms cannot extend their trading horizons without costs, given that they have to face transport and organization costs because of diminishing rate of return. Tariffs and the other obstacles imposed by people are also considered as the third trade-braking force.

Subsequent developments of Linder’s work have similarly placed emphasis on the role played by information flows. Vahlne and Wiedersheim-Paul (1977) have attempted to reflect this with the concept of „psychological distance“, which takes the form of differences in the level of development. The different levels of technical education are considered to be another trade breaking force. A final brake on the potential trade of a country could take the form of its economic isolation, the result of a divergence in its growth path from that of its neighboring countries. A country that finds itself isolated for this reason will have limited trading horizons. Hufbauer (1970) was the first to declare the divergence in growth path among the trade-braking forces.

2. Empirical Tests

Several empirical studies suggest that Linder’s theory provides a good explanation of trade behaviors of countries and that the countries with similar demand structure experience a higher trade flows. In early years of the introduction of Linder’s theory Hufbauer (1970), Fortune (1979), Sailors et al (1973), Hirsch (1973), and Kohlhagen (1977) provided reliable evidences that supported the model. However, when the role of geographical distance was recognized as another determinant of trad patterns; Hoftyzer (1975), Greytak and McHugh (1977), Kennedy and McHugh (1983), Qureshi et al (1980), showed that Linder’s theory needs serious reforms. In the recent years a number of studies, with different approaches have tried to test lender’s theory. Some studies have followed the gravity model. Bergstrand (1989, 1990), played a significant role in Linder’s theory and gravity model and studies by Thursby and Thursby (1987), Hanink (1988), Greytak. and Tuchinda (1990), McPherson et al. on Linder’s theory are noteworthy. Schott (2004), Hummels, and Klenow (2002), Hallak (2003), and Fillat-Castejón and Serrano-Sanz (2004) carefully tested the structure and the concept of similarity of demand patterns with more and found that it could appropriately explain a country’s trade behavior.

2-1. Methodology of Research

Considering that in this study the trade behaviors of Muslim countries in MENA with the other Muslim and Non-Muslim countries pre and post economic reforms are examined, the panel Data econometric technique was used in the framework of Difference-In-Differences (DID) method. With The DID technique, we may classify the results in two -pre and post reform- periods. The Penal Data econometric technique enjoys a number of advantages compared to the cross section or time series data . Some of these advantages are us under:

i. It allows for a larger number of accessible data as it uses both cross section and time series data;

ii. Unlike the method that uses one-dimensional data, Penal Data enables the researcher to test dynamic and behavioral hypotheses with a higher level of certainty;

iii. Penal Data provides a better instrument to analysis the nature of disturbance, unobserved and latent terms (Nerlove, 2002: 3-4)

As data sets in panel data are greater than time series and cross sectional data sets and on the other hand explanatory variables usually vary –both with time and from one individual to another- panel data estimators are more efficient that other estimators (Verbeek, 2004: 343)

The general form used in Penal Data models has the following expression:

(1)

where Yit is dependant variable i in time t, Xit represents a k-dimensional vector independent variable and ?it is assumed to be related to individuals and time ? represents fixed effects ?i stands for cross effects and ? Yt represents special time effects.

In Panel Data, coefficients are estimated in two different ways: fixed effects model and random effects model. In the fixed effects model the intercept of liner regression vary from individual to another. Whilst in the second model intercept are the same but there is random error for all individuals in this model ?i + ?it acts as the error term, which consists of two parts: individual specific part and the common part. It assumed that the ?i ? ?it is independent of each other and of Xit. Although this suggests that OLS estimators are still unbiased and consistent but due to the compound structure of error terms ?i + ?it these estimators will not be too efficient if ?2 ? =0. On the contrary, GLS estimators are more efficient-despite being consistent and unbiased. Therefore, in the fixed effects Models GLS estimators are used See: (Verbeek, 2004: 345-51).

2-1-1. The Specification of An Optimal model

A highly important decision in using Penal Data technique is selection of an optimal model. The appropriate interpretation is that the fixed effects approach is conditional upon the values for ?i. That is, it essentially considers the distribution of Yit given ?i, where the ?i s can be estimated. This makes sense intuitively if the individuals in the sample are ‘one of a kind‘, and cannot be viewed as a random draw from some underlying population. This interpretation is probably most appropriate when i denotes countries, (large) companies or industries, and predictions we want to make are for a particular country, company or industry. Inferences are thus with respect to the effects that are in the sample.

In contrast, the random effects approach is not conditional upon the individual ?i s, but ‘integrates them out‘. In this case, we are usually not interested in the particular value of some person’s ?i; we just focus on arbitrary individuals that have certain characteristics. The random effects approach allows one to make inference with respect to the population characteristics. One way to formalize this is noting that the random effects model states that

(2)

While the fixed effects model estimates

(3)

The estimated coefficients are equal only when the Holds. Accordingly, given the time period if there is a small number of individuals and the identification of each individual has a particular significance that fixed effect model is used otherwise random effects model also may be an appropriate approach. However, even when random effects framework are appropriate because of relative frequency the fixed effects might be more appropriate; as there may be a correlation between ?i ? Xit but this correlation is neglected in random effects and this may lead to inconsistency of estimators. . (Verbeek, 2004: 351-2)

The best way to specify the optimal model is Hausman test. The null Hypothesis in this test states that there is no correlation between ?i and Xit. Hence the fixed effects (GLS estimators of ( )is not efficient –although it is consistent. In the alternative hypothesis (H1) correlation between ?i and Xit is accepted. As a result fixed effects estimators ( ) is consistent and efficient but random effects estimator ( ) is inconsistent.

Thus under the null hypothesis there is no systematic difference between the two estimators. So we can rearrange the hypothesis as

(4)

(5)

Hausman’s Statistic test (HT) is as below:

(6)

This statistic has asymptotically distribution with K degree of freedom where K is the elements of vector ? i.e.

(7)

(See: Baltagi (2001), Green (2003) and Verbeek (2004)

2-1-2. Modeling

To formulate an empirical model to test Linder’s theory in a DID criteria emphasis is placed on two explanatory variables:

i. the difference in the per capita income of a particular country and the per capita income of the region under study (Muslim world and ROW). This variable, which we shall call it Linder variable is obtained from the square of the deviation per capita income of the reference country from the average of per capita income of the region Lin=(PCGi-PCGj )2. PCG refers to per capita income and the subscript i stands for country i in the MENA region and subscript j shows Muslim countries and ROW.

ii. The size of a country’s economy as a ratio of the whole region’s GDPT. This variable is computed as:

(8)

In this model the volume of trade of the country to the region under study i.e. XTij is the dependent variable. (all figures in constant 1996 US dollars).

Moreover, to model the DID method, dichotomous variables should be used. To examine the effects of globalization on both the slope and the intercept the three dichotomous variables are introduced:

i. G as a proxy of globalization, a dichotomous variable for which two values are assigned: zero for period prior to 1991 and 1 for periods after 1991. This variable gives us the intercept;

ii. Product of G into Linij which gives G_Lin¬ij;

iii. Product of G multiplied by GDPT¬ij which yields G_GDPT¬ij¬.

Thus, the empirical (Penal Data) model classified by the trade relation of a country with the concerned regions (Muslim world and rest of the word) may be written as:

(9)

For years after 1991 that the most of countries in the region have launched economics reforms, G=1 and for years prior to that G=0. Hence for the years before 1991 we will have

(10)

and for the years after 1991 we will have

(11)

Now, if Hausman test rejected the first hypothesis fixed effects model should be employed as a result ?j=0, otherwise according to random effects model ?j will take a constant value in the country’s trade relations.

According to Linder’s theory it is expected that ?20.

2-1-3. Data and Sample

As stated, this paper is mainly aimed at examining the effects of globalization on the trade relation of MENA region. As a result of statistical problems, some country had to be dropped from the sample. The remaining countries are Jordan, Algeria, UAE, Iran, Bahrain, and Tunisia, Iraq, Saudi Arabia Oman, Qatar, Kuwait, Lebanon, Libya, Morocco, Egypt and Yemen.

These countries‘ trading partners are categorized as Muslim world (all members of OIC) and the rest of the world. Penn table and SESRTCIC website were the sources for the raw data, which were used in this research after certain processes and calculations of needed indexes and ratios were performed. From 1975- 2002 was chosen as the period under study. To test the stationary of variables unit root test was used. Given that, our null hypothesis was the non-stationarity i.e.

H0: | ? | = 1 (12)

H1: | ? | < 1 (13)

The result (table 2 and 3) shows that null hypothesis is rejected. Hence, the stationary of variables and their convergence with the passage of time are confirmed.

2-1-4. Hypothesis Testing process

After introducing several Penal Data Models and examining the significance level (t-statistic) and the regression (f- statistic), significant variable were identified. Hausman test shows that the null hypothesis cannot be rejected in both tests (trade relation with Muslim countries and the rest). It means that GLS estimator is of random effects ( ) is consistent, unbiased and efficient. See table 6, and 7.The results of verified regressions are depicted in table 4 and 5.

2-2. Interpretations of the results

Considering that one coefficient of dichotomous variable (G) leaked the due significance it was omitted from the model. According to the results included in table 4, the trade behavior of countries under study with Muslim world is explained as below:

(14)

(3.47) – (33.99) (-7.01) (5.113) (2.20)

Subscript i in this equation and equation 15 and 16 refers to Islamic countries (IC).

As we observe all coefficients enjoy a high level of significance. Based on these results trade behavior of these countries prior to economic reforms (1991) can be expressed by

(15)

The Negative sign Linder coefficient suggests that with a decline in the per capita income gap across MENA and Muslim world for the period prior to globalization promoted trade relations. Moreover, a positive sign of the variable of size of economy shows that a faster growth of countries in region compared to the whole Muslim world will lead to an increase in the export of these countries. That is to say the greater the economic diversity the greater the economic relations. All these theories are fully consistent with Linder’s theory.

To drive the equation of region’s trade behavior with Muslim world and for the period after economic reforms we have to obtain ?1+ ?3 and ?2+ ?4 from regression 9 i. e.

(16)

According to coefficients obtained it becomes clear that on the one hand, globalization intensifies the effects of size of the economy, but on the other hand, it decreases the effects of similarity of demand patterns on trade trends. Yet, as the per capita income gap across Muslim countries declines extension of trade relations may be expected.

Consequently, we can argue that even globalization has not eliminated the relevance of Linder theory. Moreover, in view of the results of Elahi and Nahavandian (2005) and Slaughter (2001) studies on the effects of globalization on the convergence in regional incomes and divergence in global incomes in process of globalization the potential for regionalism will be strengthened paving the way for establishment of FTA and – the in the later stage- Islamic Common Market.

On the other hand, to explain the trade behavior of MENA region with other countries two dichotomous variables (G and G_GDPTit) were omitted from regression No. 9 as they lacked the due significance.

As we see in table No 5 the trade behaviors of countries of the region with the rest of world may be written as:

(17)

(2.78) – (-3.76) (2.90) (3.08)

The coefficients of this regression equation also enjoy a high level of significance. Considering the G_Lin dichotomous variable, we can drive the following equations for the two periods of the pre and post economic reforms

(18)

(19)

These two equations explicitly reveal that Linder theory can explain trade behaviors of MENA countries with the rest of the world. Considering Slaughter (2001) studies which supports global divergence in globalization process we can conclude that as the income gap increases the volume of trade in MENA countries will fall. The decline in region’s trade relations with rest of the world and the extension of volume of trade in Islamic world sets the stage for the formation of a Free Trade Area.

Conclusions:

In view of the new wave of regionalism and establishment of a common market recently raised by OIC, this paper focused on feasibility of regionalism in the Middle East and North Africa in the age of globalization. Using a new and novel method it was shown that globalization not only acts as a hindrance to the creation of an FTA in MENA rather it prepares the ground for the realization of this goal. The rational for feasibility study could be summarized as under:

i. According to Slaughter (2001) studies the per capita income trends of advanced and developing countries are diverging;

ii. According to Elahi and Nahavandian (2005), there is a converging trend in the per capita incomes of MENA countries;

iii. According to Linder’s theory – tested in this paper- , the more convergence (divergence) in per capita income the greater (smaller) the trade volume.

Thus, with the convergence in the region’s per capita income and the confirmation of Linder’s theory it is expected that intra regional trade with Islamic countries will rise and with the non-Muslim countries will decline.

It should be noted that for a successful regionalism several conditions should be satisfied, the major condition being the economic complementarity of member countries. However, this issue was beyond the scope of this paper which has focused on examining the impacts of globalization on the volume of trade exchanges between MENA countries and Islamic world and rest of the world

Table 2- unit root test (variables set 1)

Pool unit root test: Summary

Date: 01/28/05 Time: 13:27

Sample: 1980 2003

Series: XTIC_ALG, XTIC_BHR, XTIC_EGY, XTIC_IRN, XTIC_IRQ,

XTIC_JOR, XTIC_KWT, XTIC_LBN, XTIC_LBY, XTIC_MAR,

XTIC_OMN, XTIC_QAT, XTIC_SAU, XTIC_SYR, XTIC_TUN,

XTIC_UAE, XTIC_YEM, GDPTIC_ALG, GDPTIC_BHR,

GDPTIC_EGY, GDPTIC_IRN, GDPTIC_IRQ, GDPTIC_JOR,

GDPTIC_KWT, GDPTIC_LBN, GDPTIC_LBY, GDPTIC_MAR,

GDPTIC_OMN, GDPTIC_QAT, GDPTIC_SAU, GDPTIC_SYR,

GDPTIC_TUN, GDPTIC_UAE, GDPTIC_YEM, LINIC_ALG,

LINIC_BHR, LINIC_EGY, LINIC_IRN, LINIC_IRQ, LINIC_JOR,

LINIC_KWT, LINIC_LBN, LINIC_LBY, LINIC_MAR, LINIC_OMN,

LINIC_QAT, LINIC_SAU, LINIC_SYR, LINIC_TUN, LINIC_UAE,

LINIC_YEM

Exogenous variables: Individual effects, individual linear trends

Automatic selection of maximum lags

Automatic selection of lags based on SIC: 0 to 4

Newey-West bandwidth selection using Bartlett kernel

Cross-

Method Statistic Prob.** sections Obs

Null: Unit root (assumes common unit root process)

Levin, Lin & Chu t* -15983.7 0.0000 51 1012

Breitung t-stat -1.88254 0.0299 51 961

Null: Unit root (assumes individual unit root process)

Im, Pesaran and Shin W-stat -2000.53 0.0000 51 1012

ADF – Fisher Chi-square 997.442 0.0000 51 1012

PP – Fisher Chi-square 987.951 0.0000 51 1031

Null: No unit root (assumes common unit root process)

Hadri Z-stat 18.0027 0.0000 51 1061

** Probabilities for Fisher tests are computed using an asympotic Chi

-square distribution. All other tests assume asymptotic normality.

Table 3- unit root test (variables set 2)

Pool unit root test: Summary

Date: 01/28/05 Time: 13:24

Sample: 1980 2003

Series: XTRW_ALG, XTRW_BHR, XTRW_EGY, XTRW_IRN,

XTRW_IRQ, XTRW_JOR, XTRW_KWT, XTRW_LBN, XTRW_LBY,

XTRW_MAR, XTRW_OMN, XTRW_QAT, XTRW_SAU,

XTRW_SYR, XTRW_TUN, XTRW_UAE, XTRW_YEM,

GDPTRW_ALG, GDPTRW_BHR, GDPTRW_EGY,

GDPTRW_IRN, GDPTRW_IRQ, GDPTRW_JOR,

GDPTRW_KWT, GDPTRW_LBN, GDPTRW_LBY,

GDPTRW_MAR, GDPTRW_OMN, GDPTRW_QAT,

GDPTRW_SAU, GDPTRW_SYR, GDPTRW_TUN,

GDPTRW_UAE, GDPTRW_YEM, LINRW_ALG, LINRW_BHR,

LINRW_EGY, LINRW_IRN, LINRW_IRQ, LINRW_JOR,

LINRW_KWT, LINRW_LBN, LINRW_LBY, LINRW_MAR,

LINRW_OMN, LINRW_QAT, LINRW_SAU, LINRW_SYR,

LINRW_TUN, LINRW_UAE, LINRW_YEM

Exogenous variables: Individual effects, individual linear trends

Automatic selection of maximum lags

Automatic selection of lags based on SIC: 0 to 4

Newey-West bandwidth selection using Bartlett kernel

Cross-

Method Statistic Prob.** sections Obs

Null: Unit root (assumes common unit root process)

Levin, Lin & Chu t* -22022.1 0.0000 51 998

Breitung t-stat -3.10329 0.0010 51 947

Null: Unit root (assumes individual unit root process)

Im, Pesaran and Shin W-stat -3454.40 0.0000 51 998

ADF – Fisher Chi-square 748.777 0.0000 51 998

PP – Fisher Chi-square 727.799 0.0000 51 1027

Null: No unit root (assumes common unit root process)

Hadri Z-stat 14.9077 0.0000 51 1065

** Probabilities for Fisher tests are computed using an asympotic Chi

-square distribution. All other tests assume asymptotic normality.

Table 4- trade behaviors of the MENA countries with the Muslim countries

Dependent Variable: XTIC?

Method: GLS (Variance Components)

Date: 11/02/04 Time: 10:25

Sample: 1975 2002

Included observations: 28

Number of cross-sections used: 17

Total panel (unbalanced) observations: 364

Variable Coefficient Std. Error t-Statistic Prob.

C 771.8152 351.2558 2.197302 0.0286

GDPTIC? 172.0235 33.64398 5.113055 0.0000

LINIC? -3.12E-09 4.45E-10 -7.006903 0.0000

G_GDPTIC? 278.5233 33.99053 8.194144 0.0000

G_LINIC? 1.75E-09 5.04E-07 3.471607 0.0006

Random Effects

_ALG–C -816.3480

_BHR–C -170.2156

_EGY–C -932.7171

_IRN–C -1182.359

_IRQ–C 448.6195

_JOR–C -314.7449

_KWT–C 249.3942

_LBN–C -456.1962

_LBY–C -348.9692

_MAR–C -678.7201

_OMN–C -82.74419

_QAT–C -826.9708

_SAU–C 3883.275

_SYR–C 300.9173

_TUN–C -509.9924

_UAE–C 1491.185

_YEM–C -732.7498

GLS Transformed Regression

R-squared 0.806204 Mean dependent var 1260.885

Adjusted R-squared 0.804045 S.D. dependent var 1740.713

S.E. of regression 770.5584 Sum squared resid 2.13E+08

Durbin-Watson stat 1.938733

Unweighted Statistics including Random Effects

R-squared 0.813467 Mean dependent var 1260.885

Adjusted R-squared 0.811389 S.D. dependent var 1740.713

S.E. of regression 755.9813 Sum squared resid 2.05E+08

Durbin-Watson stat 1.959710

Table 5- trade behaviors of the MENA countries with ROW

Dependent Variable: XTRW?

Method: GLS (Variance Components)

Date: 11/02/04 Time: 10:41

Sample: 1975 2002

Included observations: 28

Number of cross-sections used: 17

Total panel (unbalanced) observations: 349

Variable Coefficient Std. Error t-Statistic Prob.

C 7740.308 2512.330 3.080929 0.0022

GDPTRW? 116986.6 40382.56 2.896959 0.0040

LINRW? -2.78E-08 7.39E-09 -3.763694 0.0002

G_LINRW? 1.93E-08 6.92E-07 2.784526 0.0057

Random Effects

_ALG–C 1791.944

_BHR–C -5145.395

_EGY–C -7095.146

_IRN–C 2628.343

_IRQ–C -413.8024

_JOR–C -8215.257

_KWT–C 827.2209

_LBN–C -9016.838

_LBY–C 131.9288

_MAR–C -5479.285

_OMN–C -4193.818

_QAT–C -3752.049

_SAU–C 33539.94

_SYR–C -6665.189

_TUN–C -5458.853

_UAE–C 10743.85

_YEM–C -8078.223

GLS Transformed Regression

R-squared 0.738121 Mean dependent var 9530.519

Adjusted R-squared 0.735844 S.D. dependent var 12827.13

S.E. of regression 6592.647 Sum squared resid 1.50E+10

Durbin-Watson stat 1.885858

Unweighted Statistics including Random Effects

R-squared 0.751595 Mean dependent var 9530.519

Adjusted R-squared 0.749435 S.D. dependent var 12827.13

S.E. of regression 6420.810 Sum squared resid 1.42E+10

Durbin-Watson stat 1.890042

References

? Balassa, B. and Bauwens, L. (1988), „The determinants of intra-European trade in manufactured goods“, European Economic Review, 32(7), pp. 1421–1438

? Baldwin, Robert (1971), „Determinants of the Commodity Structure of U.S. Trade,“ American Economic Review 61

? Baltagi, Badi H. (2001), „Econometric Analysis of Panel Data“, Second Edition, West Sussex, England: John Wiley and Sons

? Bergstrand, J. H. (1989), „The Generalized Gre vity Equation, Monopolistic Competition, and the Factor Proportions Theory in International Trade,“ Review of Economics and Statistics, February, 143-53.

? Bergstrand, J. H. (1990), „The Heckscher–Ohlin–Samuelson Model, the Linder Hypothesis and the Determinants of Bilateral Intra-Industry Trade“, The Economic Journal, 100, pp. 1216–1229.

? Bharawaj, R. (1962), „Factor Proportions and the Structure of India-U.S. Trade“ Indian Economic Journal, October pp.57-84

? Casas, François and E. Kwan Choi (1984), „Trade Imbalance and the Leontief Paradox,“ Manchester School Vol. 52, pp. 391-401

? Casas, François and E. Kwan Choi (1985),“The Leontief Paradox: Continued or Resolved?“ Journal of Political Economy Vol. 93, No. 3 Jun, 610-615.

? Deardorff, Alan V. (1998), „Does Growth Encourage Factor Price Equalization?“, The University of Michigan, Research Seminar In International Economics, Discussion Paper No. 431, available at: http://www.spp.umich.edu/rsie/workingpapers/wp.html

? Elahi, Nasser and Mohammad Nahavandian (2005), „The Impact of Globalization on the Per Capita Income Convergence Across the Islamic Countries of the Middle East and North Africa (MENA), Nameh-ye-Mofid [Persian], 45(10), pp. 81-110

? Fillat-Castejón, Carmen and José Ma Serrano-Sanz (2004), „Linder Revisited: Trade and Development in the Spanish Economy“ International Review of Applied Economics, July Vol. 18, No. 3, 323–348

? Fortune, J. N. (1972), „Income distribution as a determinant of imports of manufactured consumer commodities“, Canadian Journal of Economics, 5, pp. 257–267

? Fortune, J. N. (1979), „Income distribution and Linder’s thesis“, Southern Economic Journal, 46, pp. 158–167

? Gray, H. P. (1988) „Intra-Industry Trade: An „Untidy“ Phenomenon“, Weltwirtschaftliches Archiv, 124, 211-29

? Greene, William (2003), „Econometric Analysis“, Fifth Edition. Prentice Hall

? Greytak, D. and McHugh, R. (1977), „Linder’s Trade Thesis: An Empirical Examination,“ Southern Economic Journal, January, 1386-89.

? Greytak, D. and Tuchinda, U. (1990), „The Composition of Consumption and Trade Intensities: An Alternative Test of the Linder Hypothesis,“ Weltwirtschaftliches-Archiv 126, , 50 – 57.

? Hallak,J. C. (2003), „The Effect of Cross-Country Differences in Product Quality on the Direction of International Trade“, February, University of Michigan

? Hanink, D. M. (1988), „An Extended Linder Model of International Trade,“ Economic Geography, October, 322-34.

? Hanink, D. M. (1990), „Linder, Again,“ Weltwirtschaftliches-Archiv 126, , 257-67.

? Hirsch, Z. and Lev, B., „Trade and Per Capita Income Differentials: A Test of the Burenstam- Linder Hypothesis,“ World Development, September 1973, 11-17.

? Hoftyzer, J. (1975), „Empirical Verification of Linder’s Trade Thesis: Comment,“ Southern Economic Journal, April, 694-98.

? Hoftyzer, J. (1984), „A Further Analysis of the Linder Trade Thesis,“ Quarterly Review of Economics and Business, Summer, 57-70.

? Houthaker, H.S. (1957), „Can Speculators forecast prices?“, Review of Economicsand Statistics, 39, 73-87

? Hufbauer, G. C. (1970), „The impact of national characteristics and technology on the commodity composition of trade in manufactured goods“, in R. Vernon (Ed.) „The Technology Factor in International Trade“, New York: National Bureau of Economic Research

? Hummels, D. and Levinsohn, J. (1995), „Monopolistic competition and international trade: reconsidering the evidence“, Quarterly Journal of Economics, 110(3), pp. 799–836

? Hummels, D. and P. Klenow (2002), „The Variety and Quality of a Nation’s Exports,“ available at: www.klenow.com/HummelsandKlenow.pdf

? Hunter, L. C. and Markusen, J. R. (1988), „Per capita income as a determinant of trade“, in R.C. Feenstra (Ed.) „Empirical Methods for International Trade“, Cambridge, MA:MIT Press, pp. 89–109

? Jones, R., and P. Kenen (eds.) (1984), „Handbook of International Economics“, Vol. 1, North-Holland

? Keesing, D. B. (1968), „Population and industrial development: some evidence from trade patterns“, American Economic Review, 58, pp. 448–455

? Kennedy, T. E. and McHugh, R. (1983), „Taste Similarity and Trade Intensity: A Test of the Linder Hypothesis for U.S. Exports,“ Weltwirtschaftliches-Archiv 119, 84-96.

? Kohlhagen, S. W. (1977), „Income Distribution and ‘Representative Demand‘ in International Trade Flows—An Empirical Test of Linder’s Hypothesis,“ Southern Economic Journal, July, 167-72.

? Leamer, E. E. and Levinsohn, J. (1995), „International trade theory: the evidence“, in G. Grossman and K. Rogoff (Ed.) „Handbook of International Economics“, Vol. III, Amsterdam: North-Holland

? Leontief, W. W. (1953), „Domestic Production and Foreign Trade: The American Capital Position Re- Examined,“ in J. Bhagwati, ed., „International Trade: Selected Readings“, Middlesex, England: Penguin Books

? Linder, S. B. (1961), „An Essay on Trade and Transformation“ New York: Wiley and Sons,

? Markusen, J. R. (1986), „Explaining the volume of trade: an eclectic approach“, American Economic Review, 76(5), pp. 1002–1011

? Matsuyama, Kiminiori (1995), „Economic Development as Coordination Problems, “ 1123, Northwestern University, Center for Mathematical Studies in Economics and Management Science

? Matsuyama, Kiminori (1996), „Why Are There Rich and Poor Countries?: Symmetry-Breaking In The World Economy“ , NBER Working Paper 5697

? McPherson, M., M. Redfearn, and M. Tieslau (2001), „International Trade and Developing Countries: An Empirical Investigation of the Linder Hypothesis,“ Applied Economics, 33, 649-657.

? Nerlove, M. (2002), „Essays in Panel Data Econometrics.“ New York, NY: Cambridge University Press,

? Perkins, D. H. and Syrquin, M. (1989) Large countries: the influence of size, in H. Chenery and T. N.Srinivasan (Eds) „Handbook of Development Economics“, Vol. 2, (Amsterdam: Elsevier Science Publishing) pp. 1691–1753

? Qureshi, U. A., French, G. L. and Sailors, J. W. (1980), „Linder’s Trade Thesis: A Further Examination,“ Southern Economic Journal, January, 933-36

? Sailors, J. W., Qureshi, U. A. and Cross, E. M. (1973), „Empirical Verification of Linder’s Trade Thesis,“ Southern Economic Journal, October, pp.262-68

? Schott, P. (2004), „Across-Product versus Within-Product Specialization in International Trade,“ The Quarterly Journal of Economics, Vol. 119 pp. 646-677

? Slaughter, Matthew J. (2001), „Trade Liberalization and Per Capita Income Convergence: A Difference-In-Differences Analysis“ , Journal of International Economics, Volume 55, Issue 1, October, 203-228

? Stolper, Wolfgang F. and Karl Roskamp (1961), „Input-Output Table for East ermany, with Applications to Foreign Trade“ Bulletin of the Oxford Institute of Statistics, November Vol. 23, pp. 379-392.

? Tatemoto, M. and Ichimura, S. (1959) „Factor proportions and foreign trade: the case of Japan“, Review of Economics and Statistics 41, 442–46

? Thursby, J. G. and Thursby, M. C. (1987), „Bilateral Trade Flows, the Linder Hypothesis, and Exchange Risk,“ Review of Economics and Statistics, August, 488-95.

? Trefler, Daniel (1993),“International Factor Price Differences: Leontief Was Right!,“ Journal of Political Economy, Vol. 101 pp. 961-87.

? Vahlne, J.-E. and Wiedersheim-Paul, F. (1977), „Psychic distance—an inhibiting factor in international trade“, Vol. 2, Working Paper, Department of Business Administration, University of Uppsala.

? Verbeek, Marho (2004), „A Guide to Modern Econometrics“ , Washington: Wiley

? Wahl, D. F. (1961) , „Capital and Labor Requirements for Canada’s Foreign Trade,“ Canadian Journal of Economics and Political Science, August