Integration, Specialization, and Symmetry of MENA Countries

The list contains: Tunisia, Algeria, Libya, Morocco, Egypt, Iran, Jordan, Saudi Arabia, Kuwait, Syria, Turkey, Utd. Arab Em, Lebanon, Qatar, Bahrain.

Cross countries dissimilarities in patterns of trade and production structure as well as responses to specific shocks imply that the co-movement between business cycles symmetry and bilateral trade intensities might vary with the sample of countries involved in the country pair. For that reason, our estimations will consider correlation for pairs of countries involving the following groups: the highest industrialized countries (G7) and MENA developing countries. Hence, we will take into account the following three combinations: pairs involving only members of each group (that is, pairs of only G7 members, MENA countries), pairs involving (G7, MENA) combinations.

To calculate these indicators, we use the GDP at constant price (1990) in national currency taken from the World Bank’s “World Development Indicators”.

This measure has a zero value in the case of two countries having perfectly synchronized cycles.

The measures of Trade Intensity were built using the IMF’s Directions Of Trade Statistics (DOTS) data set.

Calderon et al. (2007) used a larger set of countries, including other developing countries such as those of Latin American, East Asia and the Pacific, Eastern Europe and Central Asia.

This is corroborated by the reported coefficients in Table 1 (the highest MENA-MENA correlation coefficients were registered during 2000–2010) as well as by the Kernel density estimates (Figure 3).

The regressions were done for the entire period 1980–2010, split into three sub-periods 1980–1989; 1990–1999; 2000–2010. For every sub-period, we determine the average of the bilateral trade. For the 15 MENA countries this gives a total number of observations equal to k(k–1)/2 crossed correlations.

FR adopted a larger list by introducing the total population, the real GDP, and a dummy variable for countries being members of General Agreement on Tariffs and Trade. This extension does not change the results. In our work, we mention that actually, inside the MENA region, we are experiencing a relative intensification of commercial trade. But, until recently, no real Free Trade Agreement has been signed within this region. Among the North African countries, a FTA is presently under implementation, and another one has been proposed and is under negotiation among certain Middle Eastern countries. Hence, no dummy variable can be introduced in our regressions to test the effects of FTAs among bilateral trade correlation of MENA nations.

To motivate our choice of instrumentals, we regress the bilateral trade measures on the different instruments. A complete display of the coefficient estimates is available from the author upon request.

Given this problem, and before moving to the estimations, it is indispensable to make sure that instruments are the good ones. The procedure consists in adopting the test of Durbin and Wu-Hausman (not reported here, these results are available from the authors upon request). This test allows to verify if there is a real difference between the OLS and the 2-SLS with IV, and therefore, to test the existence of endogeneity. Once finished with the DWH test, we move to the over-identification test of Sargan which permit to test the validity of the instruments.

We have to underline that another appropriate approach can be conducted. It concerns the use of what is called a “3 stage least squares 3SLS”. This procedure was adopted by Imbs (2003), and consists in constructing a simultaneous equations system. In fact, the GMM and the IV methods can be criticized because interrelationships and indirect dynamics among variables cannot be efficiently analyzed when examining only a single equation. We are planning to conduct this alternative method in a forthcoming paper.

To get these two variables, we use the consumer price index (1990), the real GDP (1990) and the public spending. Two sources of database are used; International Financial Statistics and the World Development Indicators. Other variables could be used such as the average margin between the real interest rates, the standard deviation of the bilateral exchange rates, and the correlation of the money supply growth rates.

We note that for MENA countries, no policy coordination mechanisms are conducted like in Euro-zone. MENA economies do not have a common currency, or, though maintaining different currencies, have not permanently and rigidly fixed exchange rates among themselves and full convertibility of the respective currencies into one another. They tend to follow relatively fixed exchange rate regimes vis-a-vis the principals Euro zone and US dollar currencies.

Statistics are available for the 1986–2009 period, and both indexes of Grubel & Lloyd and the productive structure similarity are calculated for the following countries: Tunisia, Algeria, Turkey, Egypt, Morocco, Libya, Saudi Arabia (MENA) and USA, United Kingdom, Canada, France, Germany, Italy, Japan (G7). We note that several works use levels of industrial classification built by the ISIC (The International Standard Industrial Classification).

A complete display of tables describing the calculated indexes of intra-industry trade and the productive structure similarity for all pairs of countries is available from authors upon request.

Against the specification (1) ′, a criticism can be underlined. It concerns the existence of a possible problem of multi-colinearity. In fact, introducing both measures of inter and intra-industry trade, which can present a strong correlation, could yield inappropriate estimations. To avoid this, Imbs (2003) and Inklaara, Jong-A-Pina, and de Haan (2008) make distinction of three indicators of international specialization introduced simultaneously in the same regression: an indicator of productive specialization, the part of the intra-industry trade, and an indicator of export similarity.

In the three compartments of Table 5, the Hansen test is significant which proves a certain endogeneity of the used instruments. Hence, these instruments could influence the products correlations through other mechanisms than the bilateral trade, and it will be necessary to introduce them. The DWH statistics show high values which means that trade intensity would not be treated as an endogenous variable.

In many studies, the term S describes the importance of the employment in the sector k with regard to the total employment. This measure is rarely adopted because of the difficulty to find reliable data set concerning employment.

The column (1) of Table 6 gives results where only the productive structure measure is included as an explanatory variable. In this column, the used method is the OLS without introduction of instruments.