Summary :

1.            All Mediterranean countries, both North and South, except Turkey, represent today the area of the world that is most affected by the economic crisis, and this crisis in Europe increases significantly the difficulties of the Mediterranean and the Middle East countries. Obviously, the “Arab Spring”, and particularly those revolutions that succeeded in changing the government in the countries we study, namely Egypt and Tunisia, as well as the specific democratic evolution followed in Morocco, have multiple causes; but according to us as economists, the “economic” determinants of these social and political movements have played (and continue to play) a major role.

Nevertheless, the “Barcelona process” of Euro-Mediterranean partnership had formally initiated in November 1995 a cooperation policy with EU promising more growth and development to the twelve countries from the Southern and Eastern shores of the Mediterranean. This policy consisted mainly in the creation of a free trade zone governed by Association Agreements in line with the World Trade Organization (WTO) rules, plus some political cooperation and assistance.

Thus, the basic idea of these Agreements and the Euro-Mediterranean partnership principle relied on the same perspective as the ones at the origin of WTO: structured trade liberalization through progressive removal of tariffs and quotas, as per the liberal international trade theories, which will be advantageous for the countries on both shores of the Mediterranean.

In particular, the Southern and Eastern Mediterranean countries were considered capable of rapidly becoming competitive once their manufacturing firms were “upgraded”, given their natural endowments in labor and attractive wage rates. The hopes lied in the foreign direct investment (FDI) inflows and the capital inflows in the form of bank loans and public aid that were supposed to go towards the most promising activities, and to represent the major pillar of the opening-favorable policy choices.

This opening had to lead to the economic growth stimulation, and its results (through the “trickle down” mechanism) had to be beneficial to all; in particular, the creation of a number of jobs able to absorb the labor force of the young generations arriving in bigger and bigger numbers at the labor market, was the indirect but eagerly expected aim of this liberalization.

2.            So what is the situation today, seventeen years “after Barcelona”, with the effects of these liberalization and economic cooperation policies? Why shall we consider their deviations that lead to the “Arab Spring” revolutions?

To answer these questions, we should first remember the extreme fragility of the liberal international trade theories. In fact, beyond the over-simplified assumptions of the “twoness”, meaning the “2 x 2 x 2” models (two countries, two products, two factors of production) that are the essence of these theories, it is impossible to prove the analytical validity of the four big theorems that compose them (the Hecksher-Ohlin, the Samuelson, the Stolper-Samuelson and the Rybczinsky theorems) and that are considered to scientifically validate the liberal doctrine, according to which the absolute liberalization of trade generates the most wealth and welfare to the (populations of the) trade partners[1].

Thus, it is not surprising that most of the time trade liberalizations don’t lead to the “virtuous cycle” forecasted by the liberal doctrine: trade liberalization → trade balance improvement (exports increase) → employment increases in the exporting sectors → distributed revenues increases → internal demand increase → induced investment increase → general employment increase →… etc.

Hence, it becomes necessary to study case by case the reasons why trade liberalization, when put in practice, does or does not succeed. This is what motivated our study, given the relative failure of the “Barcelona process”; a failure that is obviously part of the (multiple) causes of the “Arab Spring” revolutions.

3.              Since the major problem affecting the three economies we study – Egypt, Morocco and Tunisia – is unemployment, particularly youth unemployment, we focus our research on the effects of trade liberalization on employment during the last two decades. First, we propose three monographs, in which we analyze the consequences of the “Barcelona process” on international trade and the “performances” of the three countries in terms of growth and employment “before” and “after” the implementation of the Association agreements.

If we summarize the big lines, it appears that the results of this first part of our report are that the trade deficits are far from shrinking, but have actually constantly widened in all three countries, and still do today. Therefore, they have not drawn the expected benefits from the application of the Association agreements, regardless of their date of implementation (1998 for Tunisia, 2000 for Morocco and 2004 for Egypt). Thus, growth was not stimulated as expected, which lead to an increase in unemployment, and particularly in youth, female and skilled labor unemployment in all three countries, especially given their incessant demographic growth, even though we have observed some interesting demographic transition phenomena in the second half of the 1990s (lower birth rates in Morocco and Tunisia).

In other terms, growth was not inclusive, neither “before” nor “after Barcelona”, despite the opening. It is thus clear that the liberalization process was lead astray.

4.              In order to control the credibility of the observations we formulated, based on purely descriptive statistics, and in order to find analytical elements to explain them, we use econometrics to evaluate the effect of trade liberalization and other usually considered variables directly on growth and indirectly on employment, should growth become inclusive. These developments constitute the second part of our report.

We use a two-step econometric approach.

4.1.         First, we studied the effect of trade liberalization (using the “exports variable” together with other determinants) on employment dynamics for a panel of 60 developing economies having liberalized their trade relations; ten MENA countries are comprised in this panel set. The study uses dynamic panel data estimations for the elasticity coefficients of employment (for labor force between 24-65 years) versus several variables judged econometrically significant in explaining employment. These variables are growth (of GDP per capita), investment, human capital (proxied by health expenditures), agricultural productivity and exports; the role of the access to financing (proxied by private credit to GDP) appeared to have an ambiguous effect: its marginal contribution to employment using a quadratic form in the estimations is positive under a threshold (around 21%) and negative above it. When it comes to the elasticity coefficient of employment towards exports (for the entire panel set), it appears relatively small.

4.2.         The most useful result of this first part comes from the econometric estimations of our models using five regional proxies (separating our 60-countries’ panel into five geographic regions): the employment versus GDP elasticity coefficients for the MENA countries seem significantly different than the ones obtained from the global panel. This particularity incited us to proceed with a specific study of that region (10 countries in total), including the three countries that we study[2].

The results from this additional econometric study confirm the specificities of the MENA countries in terms of employment dynamics. Thus, for these countries improving agricultural productivity appears to have an important role in determining employment, but the sign of this elasticity is negative, contrary to the result observed for the global panel. The same applies to financing access: its marginal contribution to employment (tested in linear form) is negative; and if we proceed, as for the global panel, with estimating using the quadratic form, the contribution remains negative below a certain threshold and positive above it, which is exactly the opposite of the result for the complete panel. Finally, we remark that the elasticity coefficient of employment versus the increase in private investment is positive but extremely small (virtually null).

5.              Facing those results confirming the MENA region specificities, we found it crucial to study separately the employment dynamics for each of our three countries over a longer period (more than three decades) covering both the “before” and the “after Barcelona”. To achieve this, we used the most recent time series data econometric techniques to deepen our analysis.

The results from these rigorous estimations are as follows:

Employment elasticities toward the five statistically significant explanatory variables

The increase in GDP per capita, i.e. growth, appears as the major employment determinant for Egypt and Morocco, but this is not the case for Tunisia. The human capital development (measured by the rate of secondary school enrollment) is the second most significant determinant of employment for Tunisia, after FDI, while it is the least significant for Egypt, for which three other determinants (private investment, exports and FDI) seem to have relatively similar influence. However, private investment seems to have virtually no effect on employment in Morocco and Tunisia. And the later country, Tunisia, to the contrary, has to count on FDI (and education, which give an idea about the type of FDI attracted in the country) in order to develop employment opportunities.

Foreign direct investments are in fact usually considered as employment determinants. Our tests have verified this hypothesis: the estimated elasticity coefficients are positive for the three countries. However, the values of the coefficients are very different: employment creation in Tunisia is “five times more elastic” to FDI than in Egypt, and almost “thirty times more elastic” than in Morocco.

We need to go back to private investments though: the fact that this investment appears as a truly efficient employment determinant together with the fact that the panel study on the MENA countries concluded that banking credit appears to have a negative influence on employment development (except above a relatively high threshold) reveals that the financial system, in the three countries, does not respond to the firms’ financing needs. Yet, these firms, mostly SMEs, could be significant creators of employment if they can find easy ways for obtaining financing.

This result highlights a major structural problem: the failure of private investment to create enough employment and thus to contribute to a robust inclusive growth in these three countries.

Finally, export performances, the fifth employment creation determinant, have an interesting role in Morocco, but play only a minor role in Tunisia and a mitigated role in Egypt.

Conclusion 

To put the liberalization process back on the right track, a full set of accompanying economic policy measures is needed. This includes short-term measures and more substantial medium and long-term structural measures to make the development process more “inclusive” in terms of employment.

These measures, under the form of a coherent proposal for both time frames, are spelled out in the last pages of the report; they are also available in the separate “Policy Brief” document joined to our Report.

[1] As soon as we move to three countries, and/or three products, or we consider that there are « specific factors (of production) », none of these « major liberal doctrine-founding » theorems holds true…

[2] Algeria, Egypt, Jordan, Lebanon, Mauritania, Morocco, Syria, Tunisia, Turkey and Yemen.