A Global Integration Strategy for Mediterranean Countries Open Trade and Market Reforms

Oleh Havrylyshyn
Published Date:
September 1997
  • ShareShare
Show Summary Details

The economies of the Mediterranean region countries—which in the present study include Algeria, Egypt, Jordan, Lebanon, Morocco, Syria, Tunisia, as well as Israel and Turkey—experienced a period of strong and dynamic economic development in the late 1970s and early 1980s, with positive income growth, export expansion, and increasing manufacturing activity. Together with a stable financial environment of low inflation, this permitted considerable progress in output expansion and structural change, with consequent creation of employment opportunities and increases in the standard of living of the population. But since the late 1980s, with few exceptions, these economies have experienced a much less dynamic evolution. Export expansion has slowed considerably, contributing—together with the termination of the oil windfall—to a slow-down in output growth. As a result, per capita in-come growth in most countries has stagnated, or at best fallen behind the pace of other more dynamic regions. While Israel and Turkey have done somewhat better, with continued export expansion and output growth, they have both recently experienced a sharp deterioration of internal finances and worsening inflation pressures, particularly severe in Turkey.

Against this background of slowed economic progress and deteriorating employment prospects, policy discussions have focused on a new strategy combining, first, an opening up of the economy to international integration forces, and second, domestic restructuring and liberalization. An overarching theme of the new strategy is a retreat of the state from direct and indirect involvement in the market and a redefinition of a new, reduced but not disinterested, role for the state.1 In some of the countries, a conscious choice has already been made not only to pursue this strategy, but also to employ, as an important tactical vehicle, Free Trade Agreements with the European Union (EU), part of the Euro-Mediterranean Partnership Initiative (MEDA). At first view, such an approach would appear to have merit, and fits well with the perceptions of “New Growth Theory” that growth is stimulated not only by mobilizing resource inputs, but also, and sometimes more so, by improvements in technology, efficiency, and productivity, which in turn are often supported very effectively by external trade and investment flows; this is particularly so for small countries.2 But important questions remain about the nature of these Agreements, their timing, implementation, effect on trade opportunities with other parties, including possible trade diversion, and relation to other reform and structural adjustments. An important issue here is the concern raised by many analysts that regional free trade arrangements are not as good a vehicle for obtaining the benefits of open trade as is multilateral liberalization—a position that is, however, countered by others who see this as a more realistic “stepping stone” approach to global liberalization. It is the purpose of this paper to inquire, in the framework of a broad trade-strategy perspective for the Mediterranean countries, how these Agreements can best be used to ensure the maximum benefit for these economies, given the key goal of reversing the region’s recent tendencies toward stagnation.

The study focuses primarily upon Algeria, Morocco, and Tunisia in the Maghreb, plus Egypt, Jordan, Lebanon, and Syria in the Eastern Mediterranean. Israel and Turkey are, however, also analyzed, both as part of the geographical region and as “competitors and comparators” that have in general performed better than the other countries, especially on export expansion. The study presents a brief background of main developments in macroeconomic performance and trade dynamics, and then reviews some relevant theory on growth and trade linkages and on the concept of regional free trade blocs as potential stepping stones to multilateral liberalization. The paper reviews in detail various aspects of the MEDA Agreements: effects on growth and efficiency, possible trade diversion, dynamic efficiency gains, including investment benefits, and the potential credibility effects of tying other market reforms to the vehicle of the Agreements. Next, the potential for increased intraregional trade and how it is affected by the various options related to the MEDA vehicle are considered. Finally, some conclusions are drawn, returning to the goal of maximizing the benefits from the MEDA Agreements within a broader context of the optimal overall trade strategy for Mediterranean countries.

Review of Growth and Trade Performance

The main theme of slowed or even stagnating economic development is that, “despite important achievements, the MENA region (Middle East and North Africa) is yet to exploit fully its considerable economic potential.”3 While this citation refers to a broader group of countries, it applies also to the Mediterranean countries. Consider first the past achievements of these economies, most of them quite rich in natural resources, and all rich in human resources. They have succeeded in maintaining a low inflation environment both during and after the high-growth period; for the Mediterranean group, if one excludes Turkey, the average for 1993–95 of 15.3 percent is well below the developing country average of 36.1 percent. Several—Jordan, Morocco, Tunisia, and, more recently, Egypt—have inflation rates similar to the 6.2 percent in East Asia. These economies are quite open to trade both as measured by the simple share of trade to GDP (ratios ranging from about 45 percent to 125 percent, compared to 96 percent in East Asia) and, as discussed below, by econometric estimates that normalize for size and development level of an economy. Five of them—Israel, Lebanon, Morocco, Tunisia, and Turkey—have accepted Article VIII status. External debt levels and debtservice obligations are, with some exceptions, comfortable to manageable, especially after some recent reschedulings. However, total public debt inclusive of domestic debt is well over 60 percent for many of these countries.

Accomplishments in the period up to 1985 must not be overlooked, as they were significant and were not completely canceled out by the recent period of slower growth. From 1960 to 1985 per capita income grew strongly, at rates above the developing country average. Consequently, poverty in the region, at 5.6 percent of the population with incomes below $1 a day, is far below that of other regions. Finally, as El-Erian and others (1996) describe, many of these countries have initiated recent adjustment efforts that are already showing results in the form of stabilization and recovery. These initiatives include financial stabilization, continued external liberalization, privatization, reduction of price subsidies, and financial sector reforms—the last described in Bisat (1996).

On the other side of the ledger, the recent slowing of progress has meant that per capita income growth has turned negative, foreign investors have grown even more cautious about involvement in this region, and the process of employment absorption has stopped or turned around. Unemployment remains higher than in other regions, as do savings rates—only 11 percent in the non-oil producing countries, compared with 25 percent in developing countries. As described below, intraregional trade, at well below 10 percent, is much lower than elsewhere and, worse, trade integration outside the region has ceased or reversed, reflecting for most countries a stagnation of the earlier export expansion buoyancy and an inability to maintain competitiveness with other regions. Those countries that have reacted most quickly to the stagnation tendency by initiating adjustment efforts—Jordan, Morocco, and Tunisia—have avoided the worst of the deterioration, such as per capita income declines, but only Tunisia and Turkey have consistently maintained the earlier export expansion trend.

Consider in somewhat more detail the recent evolution of trade, illustrated in Tables 1 through 5. Most of the countries, despite continuing levels of protection on the high side, tend to have a medium to high propensity to trade. This is reflected, first, in the ratio of trade to GDP as set out in Table 1, ranging from about 45–55 per-cent in the larger countries like Turkey and Egypt to well over 100 percent in the smaller ones. It is more accurately indicated in a multicountry regression analysis that normalizes this ratio for the size and development level of the economy; such analysis finds that only Algeria, Israel, and Syria trade less than would be expected for their size and per capita income. In comparison, Central European economies are slightly below the expected levels, and East Asian ones only marginally above, and certainly not more than the Mediterranean group analyzed here.4 However, the evolution of the Mediterranean countries’ trade has been rather uneven over time, and generally slower than in the world as a whole (see Table 1, columns 2-4), resulting in a tendency toward declining world share of exports for most of the countries (see Table 1, columns 5–8). Israel and particularly Turkey are exceptions throughout the period since 1980, while Tunisian export growth has exceeded world levels since 1985. Thus, following the analysis in World Bank (1995b), if we define integration into world trade as the difference between trade growth and output growth of a country, it is clear from Table 2 that the Mediterranean countries other than Israel and Turkey experienced positive integration only in the period 1986–90, and even then at a pace slower than all countries on average.

Table 1.Overview of Trade Developments for Mediterranean Countries
Trade to
GDP RatioExport Growth (annual average)Percent of World Exports
Jordan125.16.2-5.511.40.090 120.050.06
Syrian Arab Republic256.8-
Developing countries88.7-2.711.511.2
East Asia96.38.318.815.8
Source: IMF, World Economic Outlook.

Source for Egypt trade ratio: IMF staff; fiscal year 1994/95

Source for Syrian Arab republic trade ratio: IMF staff.

Source: IMF, World Economic Outlook.

Source for Egypt trade ratio: IMF staff; fiscal year 1994/95

Source for Syrian Arab republic trade ratio: IMF staff.

Table 2.Share of Manufactures in Non-Oil Exports(In percent)
Syrian Arab Republic36.241.734.0
East Asia61.572.575.7
Central Europe59.861.468.6
High-Income Countries67.370.969 8
Latin America38.748.560.5
South Asia55.463 869.8
Sub-Saharan Africa36.528.931.6
Source: IMF, Trade Analysis and Reporting System.
Source: IMF, Trade Analysis and Reporting System.

Despite this weak integration experience, an important change in the structure of trade has occurred in almost all the countries, the share of manufactures in non-oil exports increasing markedly and continuously since 1985 (Table 3). Jordan is the only exception, with a relatively low and stagnant ratio just over 20 percent, while several countries—Egypt, Israel, Tunisia, and Turkey—had by 1994 reached a proportion in the same range as the East Asian and Central European competitors (about 70 percent).

Table 3.Trends in World Trade Integration(In percent)
World trade growth14.96.36.4
World output growth3.03.82.7
World speed of integration21.22.63.7
Mediterranean trade growth1,
Mediterranean output growth5.30.34.4
Mediterranean speed of integration0.02.1-2.2
Source: IMF, World Economic.

Growth rate of merchandise export and import volumes.

Growth rate of trade minus growth rate of output.

Mediterranean countries include Algeria, Egypt, Jordan, Lebanon, Libya, Morocco, Syrian Arab Republic, and Tunisia.

Source: IMF, World Economic.

Growth rate of merchandise export and import volumes.

Growth rate of trade minus growth rate of output.

Mediterranean countries include Algeria, Egypt, Jordan, Lebanon, Libya, Morocco, Syrian Arab Republic, and Tunisia.

As many observers have noted, trade among the countries of the Mediterranean as a share of their total trade appears to be on the low side—though some analysts contend otherwise.5 The issue is addressed further in the penultimate section of this paper, but it is certainly clear in Table 4, which shows the direction of exports, that the region is by far its own smallest client, accounting for well under 10 percent of individual country’s exports. Only for Syria, whose exports to the region are 33.1 percent of its total exports, do we see an exception. The almost complete lack of integration into the region for Israel is also evident in Table 4. In contrast to the low intraregional trade, we see the very strong dependence of most countries on the EU export market, which receives from nearly half to four-fifths of many countries’ exports. This dependence is particularly high for the Maghreb countries, which is perhaps not surprising given the proximity to and historical affinity for Europe and the longer-standing EU policies for preferred access enjoyed by exporters from these countries. The outstanding exception is Jordan, which exported only about 4 percent of total exports to the EU, even less than Syria, at 13.5 percent.

Table 4.Direction of Exports for Non-Oil Commodities

(Average, 1992-94, except as indicated)1

Syrian Arab Republic43.43.126.733.
Tunisia3.41.25.910. 60.876.61.710.3
Source: IMF Direction of Trade Statistics.

Trade data for some countries were not available for all years. In these cases averages were taken for available years AMU=Arab Maghreb Union, includes Algeria, Tunisia, and Morocco; GCC=Gulf Cooperation Council, includes Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the U.A.E.; EU=European Union; OIC=other industrial countries; ODC=other developing countries; TI=Turkey and Israel; EMED=Egypt, Jordan, Lebanon, Libya, Syrian Arab Republic; AMED=Algeria, Morocco, Tunisia, Libya, Egypt, Jordan, Lebanon, Syrian Arab Republic, Israel (and West Bank and Gaza Strip), and Turkey.


Average for 1993-94.


Source: IMF Direction of Trade Statistics.

Trade data for some countries were not available for all years. In these cases averages were taken for available years AMU=Arab Maghreb Union, includes Algeria, Tunisia, and Morocco; GCC=Gulf Cooperation Council, includes Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the U.A.E.; EU=European Union; OIC=other industrial countries; ODC=other developing countries; TI=Turkey and Israel; EMED=Egypt, Jordan, Lebanon, Libya, Syrian Arab Republic; AMED=Algeria, Morocco, Tunisia, Libya, Egypt, Jordan, Lebanon, Syrian Arab Republic, Israel (and West Bank and Gaza Strip), and Turkey.


Average for 1993-94.


The penetration of EU markets by Mediterranean countries has slowed since 1990 (as seen in Table 5), with the most dramatic decline experienced by Algeria, and in oil exports. Excluding Algeria, as well as Israel and Turkey, which have respectively lost some share and gained some share, the share of EU markets has increased very marginally from 2.08 percent to 2.14 percent in 1994, compared with the sharp increase by Central European countries with new Association Agreements from 2.79 percent to 4.83 percent.6 The experience varies by country, however; Syria and Tunisia increased their penetration, while others remained about the same or declined. Apart from this evolution over time, which confirms the general theme of a slowing in economic development, it is noteworthy how low the absolute numbers are. For the EU, non-oil imports from these countries are a mere 2.14 percent excluding Israel and Turkey, and about 4.46 percent including them. The sluggish performance in penetrating EU markets recently, especially in comparison with new and future competitors from Central Europe, bespeaks the difficulties that need addressing in the Mediterranean economies themselves. The very low levels of penetration speak to the potential to be achieved if the challenges can be wisely met.

Table 5.Share in EU Markets(Imports from country/total EU imports excluding intra-EU trade)
Syrian Arab Republic0.3530.2870.2870.299
Source: IMF, Direction of Trade Statistics.
Source: IMF, Direction of Trade Statistics.

Conceptual Framework

Recent reconsiderations of neoclassical growth theory have focused upon the endogenous determinants of productivity increases, including human capital improvements and technological progress—without denying a role for increased mobilization of resources. This approach was pioneered by Romer (1986) and is elaborated with empirical tests in the thorough book by Barro and Sala-i-Martin (1995). Despite its early state of development, the new theory already yields several important implications. First, government policies to support such goals as stable exchange rates, reduced price distortions, and lower protection do make a difference to overall growth performance. Second, in addition to stability and nondistortive market policies, trade openness is also a very important cause of better growth performance. In-deed, Sachs and Warner (1995, p. 2), in their empirical analysis relating growth to economic reform, argue that “the international opening of the economy is the sine qua non of the overall reform process.” Third, some authors, such as Rodrik (1994), taking a more skeptical view of policy-making, suggest that reforms are most likely to come in reaction to a profound macroeconomic crisis. For the purposes of the present paper, the most important aspect of the new growth theory is the nexus between trade liberalization and growth. We focus on this, elaborating the mechanisms behind this nexus.

Sachs and Warner (1995) admit that the opening up of trade is in their analysis also a proxy for the other accompanying measures: macroeconomic stability, reduced state action, and market deregulation. Nevertheless, they attributed a special role to trade opening: first, because it effectively forces actions in other parts of the program and, second, because it directly stimulates efficiency improvements in reaction to new competitive pressures. They did not elaborate on these two “mechanisms,” but for the first they presumably have in mind the need to allow factor markets to be more flexible in order to permit this reaction; the need to remove government subsidies and other support for noncompetitive firms or risk ballooning of budget deficits; and the need to privatize state firms to allow for the more rapid re-structuring that is forced upon private owners. As for the second mechanism, we can derive from the new growth theories several concrete mechanisms—all of which we will return to later in the paper as concepts of primary importance in discussions of the MEDA approach.

Figure 1 illustrates these mechanisms schematically, focusing on a comparative advantage sector:

  • 1. Opening up trade stimulates the simplest kind of efficiency gain, the elimination of waste in utilization of resources, or technically a movement from the inefficient point QA1 to the current technology isoquant QE1.
  • 2. Resource reallocation improvements come both from removing distortions on traded goods prices as production shifts toward comparative advantage equilibrium7 and from factor market liberalization (shown as the shift to Q*1). Both of these effects al-ready result in output expansion.
  • 3. Introduction of new technology—either through domestic research and development or, more likely in small countries, initially through imported goods and technology—results in an inward shift of production isoquants to Q**1 reflecting a lower output with less factor inputs.
  • 4. Expansion of output (for export or home market) of comparative advantage goods comes when new investments (domestic or foreign) create new capacity and hence growth of output. The capacity expansion permits absorption of underemployed labor, which has the dual benefit of adding new resources for growth and addressing the goal of poverty reduction.

Figure 1.Schema of Dynamic Efficiency Gains From Trade Liberalization

One-unit production isoquants: equilibria after shortand long-run trade liberalization adjustments

In the trade liberalization literature and debates in the past, quantitative estimates of the gains from freer trade were generally positive but small in magnitude, hardly a compelling case for liberalization. Frequent references were made to the fact that such estimates covered static gains only (resource reallocation, or one part only of item (2) above), and that dynamic gains, which could not be calculated, were no doubt much higher. The inferences from new growth theory as described above provide a theoretical background to the increasing accumulation of evidence on the large size of dynamic gains—some of which evidence preceded the theory. Havrylyshyn (1990) reviewed earlier studies linking trade liberalization and productivity growth, concluding that while the evidence for a positive link was not strong, no studies found statistically negative links, while a few, especially micro-panel data studies, showed significant positive relations. One such study (Haddad, 1993) showed the firm-level productivity gains that came in Morocco after the first, limited opening up of the 1980s. Extended applications of computable general equilibrium (CGE) models to cases of trade liberalization—including several for the Mediterranean region, discussed below—also suggest that the dynamic gains from liberalized trade are much larger than the static gains (magnitudes of 4–5 percent of GDP and more, compared with 1 percent or less) (see Rutherford and Tarr, forthcoming). Thus, the arguments of such analysts of regional trade as El-Naggar (1992) and Zarrouk (1992) that further opening up of these economies would generate large but not easily measured dynamic gains may not have been so unrealistic.

Regional Free Trade: A Stepping Stone to Global Free Trade?

The preceding subsection described recent writings on the positive link between liberal trade and growth. Since the MEDA is conceptually similar to a regional free trade zone, one should ask if the positive link remains. Traditional analysis of free trade zones turned on the concept set forth by Jacob Viner that both trade creation and trade diversion would occur and that their effects were on balance positive if the former dominated but negative if the latter did. Further, empirical evidence until the 1980s firmly indicated that, while trade creation did indeed dominate in the European Community, trade diversion dominated in the various preferential agreements among groups of developing countries, so that the agreements did not work well and many lapsed.8 Nevertheless, the issue has recently become of renewed interest as a result of two factual developments. First, as Bhagwati (1993) notes, U.S. leadership in the multilateral liberalization of the postwar period has been reoriented toward achieving free trade within blocs. The creation of NAFTA and discussion of extending it southward is a reflection of this, as is talk of a broader ASEAN arrangement (see DeRosa, 1996), and the MEDA, which directly concerns this paper. Second—and no doubt partly as a consequence of the first development—new efforts at regional integration in developing countries have been started: witness MERCOSUR and ASEAN. Thus the question is again reopened as to whether or not a partial free trade area can be beneficial. However, the arguments are not the same as before (see Krugman, 1993), and include new economic analysis—the dynamic gains discussed above—as well as the political economy argument that regional blocs may be a more realistic way to attain global free trade than multilateral GATT/WTO negotiation rounds, and that thereby they can act as stepping stones to global liberalization.

Bhagwati (1993) attributes the basic vision of the stepping stones approach to William Brock, former U.S. Trade Representative, and while arguing against it he admits the concept has a possible theoretical logic and appeal. Many analysts have since debated the issue9 and the proponents’ main arguments are as follows. First and most important, a sort of realpolitik on global trade considers that multilateral liberalization is not going to go much further, or at least not soon. Second, it is argued that the trade diversion costs are not likely to be as large as previously, for the following reasons:

  • much more intra-industry trade will occur in the same way it did for Europe earlier, minimizing trade diversion;
  • the agreements are more comprehensive than, and not selective as, earlier ones;
  • the agreements come upon the heels of a general reform in the economy; and
  • the agreements are related to some longer-term commitment or “anchor” for reform toward an outward orientation—such as integrating with the EU, formally in the case of Central Europe, or less formally in the case of the Mediterranean countries.

There are arguments on the other side, of course, many of which are put by the same analysts who outline the potential idealized path of regional stepping stones. Perhaps the most powerful argument for now is one that only time will settle: will broader global liberalization indeed follow, or will partial free trade become the terminal point? The debate is in fact better seen not as a theoretical one, but as an empirical set of second-best judgments. In the conceptual debate, this view is best captured in the work of Krugman, who on the one hand (1993) argues there may be more prospect in practice of going the stepping stone route, but on the other hand demonstrates theoretically that blocs, especially a small number of large blocs, are clearly a second-best welfare result compared with global free trade.

Indeed, those who list the benefits of the regional approach invariably provide a corresponding list of conditions to be met and policy steps to be taken for the benefits to materialize: parallel reforms must take place; liberalization in the rest of the world cannot be delayed too long; and vested protectionist interests must be faced with a clear calendar of commitments and follow-up implementation. In this vein, even one of the strongest critics of the approach, Bhagwati (1993), proposes a mechanism for ensuring that regionalism comes out as second-best and not third-best or worse: his idea of “programmatic regionalism” is to set a clear time-specific commitment for moving from regional liberalization to global. All of the conditions behind the proponents’ arguments are thereby distilled into one calendar target.10 One cannot help but be reminded of the 19th century writings on infant industry protection and Mill and Bastable’s notion of a “sunset law” for such protection to ensure it was not perpetuated by vested interests; Bhagwati’s proposition might be described as a “sunrise law.” In the next section, we discuss arguments for the MEDA providing an “anchor” for the countries to help pursue the general reform needed for success. If, as Sachs and Warner (1995) argue, the opening up of trade forces other reforms, then a double “sunrise law,” setting a target for both completion of reforms and global liberalization, will strengthen these pressures and will fix the anchor even more solidly.

Maximizing the Benefits of the Euro-Mediterranean Agreements

In the context of the theoretical arguments noted above, we consider next the conditions under which the Euro-Mediterranean Agreements can yield maximum benefits.

Nature of the Agreements

The Euro-Mediterranean Agreements narrowly consist of international agreements to establish a free trade zone, bilaterally between the EU as a whole and individual Mediterranean countries or, in the case of Turkey, as a customs union. For those countries that have already signed, it is in effect an asymmetric liberalization at this stage, because the countries in question already had nearly duty-free entry to EU markets for manufactures, though not agricultural goods. This access is now affirmed by international treaty, and the countries will be opening their markets to EU exports over the coming 12-year period. But the Agreements go beyond a free trade zone, although the other steps are less clearly specified, and their nature needs to be inferred not only from the texts of the Agreements themselves, but also from the important text of the Barcelona Declaration of November 1995 (Euro-Mediterranean Partnership, 1995). Thus, Article 18 of the Agreements for Morocco and Tunisia (European Union, 1995), as well as other language, strongly suggests the aim of eventually liberalizing agricultural trade, with the issue to be formally reviewed in the year 2000. There are vaguer references to gradual liberalization in services, harmonization of regulations in the markets, the rights of establishment, economic cooperation involving technical assistance to achieve uniformity of standards and procedures, and financial cooperation. The last is to be changed significantly from the previous commitments of EU Financial Protocols establishing levels of grant and loans by country. Under the new Euro-Mediterranean Partnership Initiative (MEDA), there is a general commitment of budgetary resources of 4.7 billion ECUs for the region over 1995–99, plus a roughly equal amount from the European Investment Bank to be distributed, as the Barcelona Declaration (1995) states, in a way that promotes the implementation of the Agreements and general reform of the economy, market orientation, and the private sector. This has been understood by those involved in MEDA to mean that the financial support of the EU will depend on the progress of the Agreements’ implementation and of reform in individual countries. The mechanisms for this new procedure within the EU are not yet fully worked out, but it is clear that the Mediterranean partners are in a position to take the initiative with proposals for the use of these resources.

The connection between financial support and the general pace of reforms and implementation of the Agreements is important in that it may provide an additional firmness to the “anchor” of the EU Agreements themselves, with their fixed 12-year calendar for bilateral free trade with the EU. Given the weight of EU trade in most of these countries, it is not surprising that the Agreements are nearly universally perceived as generating pressure for comprehensive reforms in other aspects of the economy, as well as the eventual liberalization of intraregional and global trade. The large weight of EU trade also creates the perception that trade diversion from the rest of the world is unlikely to be very significant. But apart from the existence of a free trade bilateral zone with Europe, none of the expected results is automatic, nor is the trade diversion effect clear. Many steps will need to be taken by the countries themselves, by the EU, by other international organizations, and perhaps by the Mediterranean countries as a group to achieve the hoped-for optimal effects.

The rest of this section of the paper inquires into the various expected effects and how they may best be managed so as to attain the optimal result. It summarizes some empirical estimates of welfare effects using CGE or other models; discusses the three most important potential costs in the short run—fiscal losses, disemployment effects, and trade diversion; and reviews the major benefits in the form of dynamic efficiency improvements, giving particular attention to new investment.

Overview of Expected Net Gains From MEDA

The asymmetric nature of MEDA means that the static gains of lower-cost imports netted for tariffs plus export gains are likely to be one-sided, that is, favoring Europe, but only marginally important for Europe given the magnitudes. In addition, future resource reallocation gains can be captured in CGE model estimates (see Table 6), but these too are limited and the estimates do not cover all of the expected dynamic benefits. While it is generally recognized that the measured gains are small, many still support MEDA because of the potentially large dynamic gains. The arguments in favor of MEDA are generally from the perspective of “deeper integration,” as Lawrence (1996) or Page and Underwood (1995) describe it. The Mediterranean partners would “integrate” with Europe not only via the free trade zone but also through contractual assurance of future market access; harmonizing of statistical standards, customs procedures, market regulations and investment mechanisms; and future liberalization in financial and other services. Such deeper integration would create the environment for traditional resource allocation efficiency as well as the research, development, human capital enhancement, and technology transfer that are central elements of the new growth theory. The vehicle of investment expansion—attracting additional domestic and foreign investment—is crucial to the achievement of the dynamic gains. Lawrence compares the connection through the MEDA between the Mediterranean countries and Europe to NAFTA, connecting Mexico with the United States, and concludes that MEDA is a form of regionalism that could overcome many of the historical shortcomings of earlier preferential arrangements. Comparing the MEDA to Central European Association Agreements, Hoekman and Djankov (1996a) suggest the latter “have been an important element in anchoring expectations,” and expect the former could do the same for the Mediterranean. It appears that policymakers in those countries that have signed Agreements, as well as many in countries still negotiating, generally share such a bold vision of MEDA’s benefits.

Table 6.Welfare Effects of the Euro-Mediterranean Partnership Initiative(In percent of GDP)
MEDA No Agricultural LiberalizationMEDA + Agricultural LiberalizationMEDA + Liberalization vis-à-vis Rest of WorldMEDA Dynamic Gains
Sources: For Egypt, Galal and Hoekman (1996); for Morocco, Rutherford, Rutstrüm, and Tarr (1993); for Tunisia, Rutherford, Rutstrüm, and Tarr (1995) and Brown, Deardorff, and Stern (1995); and for Turkey, Harrison, Rutherford, and Tarr (1996).
Sources: For Egypt, Galal and Hoekman (1996); for Morocco, Rutherford, Rutstrüm, and Tarr (1993); for Tunisia, Rutherford, Rutstrüm, and Tarr (1995) and Brown, Deardorff, and Stern (1995); and for Turkey, Harrison, Rutherford, and Tarr (1996).

Given this perception, empirical estimates of potential gains are not easy, though the use of CGE models to estimate the long-run gains at least from resource allocation effects provides a better lower-bound estimate than do partial equilibrium measures of so-called deadweight loss triangles. The few studies doing this for MEDA are summarized in Table 6, showing the annual welfare gains as a percent of GDP. Estimates of other impacts are discussed later. Several points arising from Table 6 merit attention. First, while the magnitudes for resource allocation gains from the present MEDA are generally positive, they are not very large. Second, agricultural liberalization of EU markets adds little benefit, because the effect for Mediterranean countries may be two-sided: opening export markets but raising costs of imported products. Third, far more significant gains are added when liberalization is done vis-è-vis the rest of the world (third column), because the competitive forces of lower-cost imports forcing efficiency reallocation are then at their maximum. Fourth, other dynamic gains may be quite large in magnitude, though only the study of Tunisia by Rutherford, Rutstrüm, and Tarr (1995) attempted to estimate such effects. They found that gains from harmonization of standards and lower costs of trading together account for 2.3 percent, compared to the 1.6 percent for resource allocation effects alone. This effect, which is not estimated for the other countries, explains the much higher value of dynamic gains for Tunisia. However, it covers only some of the potential dynamic gains, so if the others could be included the magnitudes might be greater still.

Short-Run Adjustment Costs

The most immediate impact of MEDA that would entail some cost and require adjustment is the fiscal loss of tariff revenues. We pose three questions in this connection:

  • What might be the fiscal impact of lost tariff revenues resulting from free trade with the EU?
  • What is the flexibility of the fiscal system in compensating for these losses?
  • What is the flexibility of the fiscal and broader public sector in adapting to these losses in other ways?

To answer the first question one can first use the share of customs duties in government revenues, though the actual impacts in a general equilibrium framework may be different and can vary under different policy scenarios.11 However, results from the CGE studies cited suggest that the simple ratio is not far off as an indicator for cross-country comparison. That indicator, as seen in Figure 2, implies clear country groupings. Algeria, Egypt, Israel, and Syria, with customs duties shares around 10 percent of government revenue, would not face serious difficulties in addressing the effect of these revenue losses—though adjustment would still be required. In a middle group, with shares near 20 percent, are Jordan and Morocco, while Tunisia and, far more so, Lebanon, are clearly cases in which the loss of revenue would be major and require high priority attention.

Figure 2.Fiscal Loss Adjustment

Source: Havrylyshyn (1996).

1 Gross tax capture coefficient is intended to determine how close the present system is to an idealized one in which a uniform rate with no exemptions applies with complete compliance.

But it is also important to consider a country’s flexibility to compensate for or adapt to the losses in tariff revenue; in doing so, we see the seriousness of the problem is somewhat lessened, at least for most countries but Lebanon. First, and applicable to all countries, is the observation of Grissa (1996) that the schedule for tariff cuts spreads them out over a long time, with smaller cuts coming earlier, so that there is enough time for most countries to make the necessary adjustments. Second, value-added tax (VAT) systems in most countries can be improved both by greater administrative efficiency and by broadening the base. A rough indicator of these two effects is shown as the gross tax capture coefficient in Figure 2.12

Two countries with moderate or high revenue dependence—Morocco and Tunisia—have considerable room for improving collection of VATs. This is a key conclusion of fiscal studies by Abed and others (1996) and Abdel-Rahman, Nizoux, and Goyou (1996). Tunisia’s recent inclusion of retail trade in the VAT coverage shows the flexibility of the country’s system. The simple ratio of collected VAT to consumption is particularly low in Tunisia, and the gross tax capture coefficient, which reflects both coverage of the VAT system and compliance, is very low for both countries. These ratios are also very low for Syria, reinforcing the assessment that the tariff revenue loss problem is quite manageable in that country.

Perhaps even more viable is an option available to all countries—reducing the size of the public sector. Of all the countries, only Lebanon faces a serious problem of lost revenues with little evident room for compensation or adaptation. Its present tax situation is very fragile, with revenues still below 20 percent of GDP, tax collection seriously affected by both compliance and administrative problems, and little public sector ownership to privatize. The precariousness of the situation in Lebanon is reflected in the fact that no tax coefficient estimates could be done: there is no VAT, and envisaged adoption of a general sales tax (GST) continues to be delayed.

While for many countries the fiscal loss is important, the options for compensating revenues and reducing government size are significant and render the problem manageable. Furthermore, the notion of “tax loss” needs to be put in perspective: the loss of government is a gain to consumers, whose disposable income rises, particularly so if part of the adjustment occurs by reducing government expenditures. Having said this, it must be emphasized that the adaptation will not occur automatically but will require attention and effort. Further, the manageability of the fiscal loss should not be misinterpreted to mean there is not a general risk of budgetary deterioration. There is, therefore, a separate need for macroeconomic stabilization policies.

The estimates for overall welfare gains described above all point to much higher benefits if trade is liberalized globally and not just bilaterally. A surprising corollary result is that the additional fiscal losses are quite small. The Morocco and Tunisia studies compute the necessary increase in VAT revenues to compensate for fiscal losses. The numbers, especially for Tunisia, are not small at about 33 percent, but full global liberalization increases this loss by only a few percentage points to 39 percent. The adjustment is smaller in Morocco, but there again there is little additional impact from global liberalization.

Employment effects are also estimated in these studies, calculated not as disemployment but as the percentage of the labor force that shifts from declining sectors to expanding ones. The Morocco study by Rutherford, Rutströ, and Tarr (1993) estimates shifts in the range of 2.3 percent to 3.3 percent with the current MEDA, or 2.7 percent to 3.6 percent for full global liberalization. The numbers for Tunisia by the same authors (1995) are much higher, comparable calculations giving median values of 7.0 percent and 9.2 percent, respectively. The Tunisia study by Brown, Deardorff, and Stern (1997) gives the much higher number of 22 percent, which given other results one might consider an upper limit. But all of these are changes over a long period of adjustment—12 years under the MEDA—so that even the upper bound number for Tunisia does not seem like an exceptionally large shock. Employment will experience the same effect as for fiscal losses: moving to full global liberalization, while resulting in a much larger additional welfare gain (Table 6), adds very little adjustment cost.

Finally, we consider the potential costs of trade diversion that may occur under the preferential nature of MEDA, though of course in doing so one must take into consideration the Uruguay Round reductions that these countries will be implementing in parallel. The first point to note is that for most of these countries the bindings under the Uruguay Round are at levels near or above actual current most favored nation (MFN) tariffs, so that any further lowering of MFN tariffs during the period of MEDA implementation would not be automatic but, rather, a policy choice. Trade diversion in Viner’s traditional sense may not be very substantial for two reasons: Europe is already the major supplier for many of these countries, and while for some products other sources will clearly be less costly, Europe is not, broadly speaking, a highly cost-inefficient source.13 However, viewed in the broader perspective of dynamic gains the opportunity costs of maintaining significantly higher protection against the rest of the world are large: they are at a minimum the differences in welfare gains shown in Table 6 for MEDA alone (column 1) and those for global liberalization (column 3). If the one-sided liberalization under MEDA is, nevertheless, thought beneficial because of the large dynamic gains that competitive pressures will generate, then the same logic can be extended that even greater gains will come from wider liberalization—in particular as the additional adjustment costs appear to be quite small.

It is significant in this context that the expected effect from the Uruguay Round on these countries is quite limited, as many observers have noted.14

This is because earlier preferences in Europe will be eroded, agricultural prices may go up, and the efficiency effects of services competition will be hampered by limited commitments in services by the Mediterranean countries. This underlines the importance of the opportunity to pursue, in parallel with MEDA, more rapid MFN liberalization than foreseen under the Uruguay Round commitments.

Conditions for Realizing Dynamic Gains

The type of dynamic gains that lie behind the optimistic expectations of MEDA’s benefits cannot be empirically estimated. The one limited exception for Tunisia does confirm that they are larger than the static gains, but this is too thin a piece of evidence to give firm comfort. In this subsection we make no further attempt to address the question of how large these benefits may be, but pose a different question: under what conditions are these dynamic gains most likely to be realized?

In a broad sense the conditions for maximized gains are similar to those set forth by supporters of the new regionalism, who argue that it has better prospects for success than earlier preferential arrangements. The list of conditions given by Lawrence (1996) is typical:15

  • 1. general trade strategy that is outward-oriented rather than import-substituting;
  • 2. market orientation replacing highly interventionist state;
  • 3. private sector development promotion;
  • 4. “deeper integration” (as defined earlier in this paper) and not just trade barrier reductions as part of the trade agreement; and
  • 5. other reforms besides those in the trade sector, including budgetary and fiscal improvements, goods and factor markets reforms, financial sector reforms, and adequate social safety nets.

It may be useful to make these five conditions more concrete, describing them where possible in terms of policy actions that may be taken by the countries themselves, their European partner(s), other countries, or international organizations.

To maintain an outward orientation, liberalization via MEDA alone is not sufficient. As many of MEDA’s proponents argue, more or less parallel global liberalization is necessary to maximize the stimulus for competitive and efficient production. At the same time, any remaining government supports for domestic producers must be reduced substantially and those that remain must not discriminate in favor of domestic markets as was often previously the case. Government efforts to assist the restructuring process will be an important part of the MEDA, and will be partly sup-ported by EU grants or loans. The need for some adjustment mechanism of this sort is not in question, for it serves both the direct purpose of adjustment and the political economy objective of minimizing opposition to the liberalization. In doing so, however, it does risk confusing the message to producers about the government’s intentions, and being captured by interests who do not wish to adjust but want to continue receiving government assistance. To avoid this risk, one might consider a clear target date for two items: a “sunset law” for restructuring programs and mechanisms and a “sunrise law” for the onset of global MFN liberalization. MFN liberalization would have to be much more rapid than the Uruguay Round commitments, which leave a large gap between MFN tariff rates and those under MEDA free trade.

In the same spirit, the MEDA Agreements themselves can be reviewed to achieve more complete bilateral liberalization, particularly in agricultural goods. While the estimated effects in Table 6 do not suggest much additional gain from access to EU markets of agricultural goods, these estimates do not capture potential dynamic gains in agriculture, which should be no less than those in manufacturing. Further, the signal of full, open, and outward-oriented free trade conveyed by complete bilateral liberalization should not be undervalued.

A market-oriented economy, meaning liberalization of factor and domestic goods markets, is needed to permit faster and more complete reactions to the competitive pressures that come from opening the economy. The dynamic gains hoped for depend critically on restructuring and defining new opportunities for production, exports, imports, trade, and services. If state regulation affects prices or inhibits new investments, the economy’s reactions to competitive pressures will be slowed significantly. The longer-term gains, in the form of new investment, new technology, and firm-level endogenous efficiency improvements in the spirit of the new growth theory, all depend upon market liberalization even more than do some of the shorter-run reactions, and it is agreed by all concerned that the biggest dynamic gains are to be had from these longer-term phenomena. Attracting new investments is a key aspect of the expectations, and we return later in this sub-section to address it more fully.

Promoting private sector development is supportive of market orientation, and should take place in two ways: by privatizing state enterprises and by deregulating market entry to simplify the procedure for new entrepreneurial activity, including the right to free establishment by foreign as well as domestic firms. The principles stated in the Barcelona document refer explicitly to this, and equally important, it and other EU documents relate the level of financial support from the EU for a country’s program to the progress in these areas. The importance of this “anchor” of private sector development should not be underestimated. The onus is on both partners of MEDA to jointly implement these principles as quickly and effectively as feasible.

The deeper integration foreseen in statistical and procedural matters, in harmonization of product standards, and in reduced trade costs covers a vast array of activities engaged in by both sides of MEDA, and cannot easily be made more concrete or target based. However, there are certainly areas in which targets for “equivalence” with EU or other standards are feasible. Thus, the Association Agreements for Central European countries set target dates for harmonization of statistics, customs procedures, and certain laws. Analogous targets can be set by the Mediterranean countries independently or in conjunction with their partners. Other anchors may be found, such as the IMF’s new set of standards for public dissemination of data.

In addition to the accompanying reforms that are necessary, a climate of fiscal stability is also needed. While most Mediterranean countries have done well on this score, this is not the case for Turkey, and in some others there is a risk of backsliding, which is only partly exacerbated by the fiscal losses noted earlier. To maintain prudent budgetary positions, changes in tax regimes, civil service reform, and expenditure reviews will be needed. There is a role to be played in these reviews by the IMF, as well as the EU with its financial and technical assistance. Financial sector reforms have begun but are not far advanced in many of these economies, as described by Bisat (1996). A lagging of improvements in this sector could result in serious bottlenecking of the hoped-for dynamic gains: even if private investors are freed from burdensome government regulations and price distortions and want to seize the new opportunities of an open economy, if they face an uncompetitive financial market with excessive spreads on interest rates, and/or continued competition for credit from a government needing to finance its deficits, they will not invest and the expected dynamic gains will not be forth-coming. Finally, social safety nets will need to be put in place or, more correctly in most countries, repaired. While the region as a whole compares favorably with others in terms of poverty levels (World Bank, 1995b), the social programs that exist are costly and leaky, draining budgetary re-sources without effectively addressing the needs. Adjustments to an open economy and continued high population growth will exacerbate the problem, so serious efforts are needed here too.

Does MEDA Meet the Conditions for Optimal Gains?

Are these conditions likely to be met in the Mediterranean countries? Hoekman and Djankov (1996a) and Lawrence (1996) express some doubt about how firmly grounded all these different elements are in the MEDA initiative. They note that apart from the EU free trade calendar of 12 years, the other measures of deeper integration are quite vague. In comparison, the Central European Association Agreements have greater specificity on right of free establishment, trade in services, capital movements, and financial sector regulations. While this is correct, and possible tightening in upcoming Agreements or reviews can be envisaged, it may not have been realistic to expect negotiations on an already bold initiative to go even further. But if the anchoring effect of the Agreements is to hold effectively in this area, the general principles agreed to must be filled in with details later. From the EU side, the new form of financial cooperation cannot have been too finely detailed in advance, and will require time to evolve, including on aspects of common interest to the EU and the IMF/World Bank. In this evolution, the end-goal and the fundamental principles outlined above must be kept in mind at all times.

The most important actors in all this are of course the policymakers of the Mediterranean countries themselves. The situation varies among countries, and these differences will determine the relative priorities of policy actions to be taken. Nsouli, Bisat, and Kanaan (1996) and Havrylyshyn (1996) review what may be defined as the state of “readiness” of countries to maximize MEDA benefits. For some, like Lebanon and, to a lesser extent, Tunisia, the fiscal losses may be very large and will require intense efforts. For others, like Algeria, Jordan, and less so, Morocco, external debt is a heavy burden, while in Egypt, Israel, and Lebanon it is the domestic debt that raises the threat of new imbalances. Algeria and Syria, even with their recent liberalization, still have the highest degree of state involvement. All have some degree of pressure on their budgets, with Lebanon and Turkey in a particularly dramatic situation, albeit for different reasons and with different effects on inflation—inflation in Turkey is approaching 100 percent, and that in Lebanon is below 20 percent. What is common to all the countries is that once they have made a commitment to the MEDA approach—as have Israel, Morocco, Tunisia, and Turkey—they will have made a commitment to extract the maximum benefits from that approach and will therefore need to address concretely how to achieve the requisite conditions described above.

Stimulating Increased Investment

The role of the new investment, both foreign and domestic, that one expects will be attracted by the Agreements deserves special emphasis, for without such an increase in investment, the dynamic gains will not be very large. The aspects of the Agreements that are thought to provide such an attraction comprise: new production opportunities as resource reallocation takes place; the possibilities for higher efficiency in existing activities; and, last but not least, the credibility that becoming an informal part of the European space would grant. Page and Underwood (1995), comparing the possibilities for Tunisia and Morocco with experience in other integration schemes, suggest that a major potential benefit of the Agreements may come from substantially increased foreign investment. Mexico received an increase of $30 billion in investment in the run-up to NAFTA, Portugal and Spain experienced an increase in investment after EC integration equivalent to 2 percent of GDP, and, more recently, the leading reformers in Central Europe have enjoyed similar increases.

There are no estimates of investment effects, but it is useful to ask the same question as for dynamic gains: what conditions will assure the greatest attraction of investment? However, one should not focus only on foreign investment, even though it may carry certain advantages that domestic investment does not, including greater inherent technological experience and marketing capacities.16 In an outward-oriented strategy, furthermore, the conditions that will attract foreign investors will not differ much, if at all, from those that attract investment generally; even the increased external credibility that results from an agreement with a large partner should have the same effect of mobilizing investment domestically.

The Agreements will not by themselves be sufficient unless potential investors believe the Agreements are generating the expected benefits in resource reallocation, efficiency, etc. As the actual results will not come early, the stimulus to new investments (which, circularly, are needed to realize the results) will at first have to be based on perceptions and expectations of investors on the credibility of opening up and trade liberalization. To measure this credibility, investors will surely look for more concrete signs than the formality of the Agreements themselves. The signs of an improving investment climate include the sorts of concrete policy actions that we have outlined as the “conditions” for maximizing dynamic gains: steps to ensure the sustain-ability of budgetary discipline, liberalization of factor and goods markets; deregulation of investment and enterprise establishment procedures; liberalization and increased competition in financial markets; clear signals (e.g., target dates) on the phasing out of government support for inefficient firms, be they state or private; and clear signals and schedules for trade liberalization vis-à-vis the rest of the world.

It has been a long-standing view of foreign investment experts that the old strategies of import substitution for attracting such investment—protected home markets, tax privileges, and indirect subsidies for export activity—are not very effective and that the best investment climate is an outward orientation to the market and the private sector, and a regime that treats foreign and domestic investment uniformly. There is a risk here that, as in the past, more inward-oriented policies with granting of special privileges will be seen as succeeding for a while, because they do generate some amount of interest by investors. The difference is that the more open strategy can yield much greater and more sustainable results in terms of attracting investment. Thus, to ensure a new flow of investment under the Agreements, firm and decisive progress has to be made on a comprehensive set of economic policies by the authorities of the Mediterranean countries; the Agreements will not bring this investment automatically.

A special concern has been expressed that the MEDA consists of a parallel series of free trade agreements and that the continuing high barriers to trade among the Mediterranean countries will lead to a hub-and-spoke pattern of investment: investors will set up in Europe to service all the individual Mediterranean country markets, rather than investing in any one smaller market.17 This is a major risk to the process. Hoekman and Djankov (1996a) note that this risk is minimal in the Association Agreements with Central Europe given the establishment of the Central European Free Trade Agreement, and they suggest such a strategy be pursued by the Mediterranean countries. That is certainly an option that would reduce the hub-and-spoke effects, and in any event—as the next section of this paper discusses in more detail—reducing intraregional barriers is beneficial in other ways as well. But this would have to be achieved effectively and soon, in a situation where a complex web of regional agreements has been in place for some time but has not been successfully implemented.

There is another option to consider, that is, full multilateral liberalization on a much faster schedule than required by the Uruguay Round, more in keeping with the MEDA calendar of 12 years. This would achieve the effects of regional liberalization to offset hub-and-spoke effects, but would also go much beyond this. Full liberalization need not be fully parallel in magnitude and time to the MEDA, but it should not be far behind. Most important, it must have credibility if positive effects are to follow, and it should be announced in advance as a concrete target. The issue arises whether commitments to existing preferential arrangements, in particular the Arab Maghreb Union of Algeria, Mauritania, Morocco, and Tunisia, would not be breached by a shift to a very fast global liberalization. First, if the global liberalization goal were, for example, average MFN tariffs of 5-10 percent at the end of the 12-year MEDA period, this leaves room for a parallel free-trade arrangement in the region. Second, and more important, the issue of these prior commitments has inevitably been put on the table by the MEDA, and as many policymakers in the region see it, the regional agreements will need to be reviewed and put either on an effective regional free-trade track or on an MFN track.

Each of the options has its advantages and constraints. Regional free trade as a first step has a historical commitment on which to build, while a much faster MFN approach needs to be based on a yet unformulated policy commitment. That commitment can, of course, be developed, for the MFN approach has a compelling advantage as is well demonstrated in the various analyses of MEDA effects: the benefits from the extra step of global liberalization are very large and come at little cost, a cost that was in effect already invested by the authorities when they committed to the MEDA vehicle for implementing the courageous strategy of opening up trade.

Potential for Intraregional Trade

In view of the frequent concerns expressed about too little intraregional trade, it is useful to consider some tentative indicators of its potential.

Is There a Natural Potential for Increasing Intraregional Trade?

Many authors, such as El-Naggar (1992), Zarrouk (1992), Fischer (1993), and El-Erian and Fischer (1996), have remarked on the low levels of trade within the region, accounting for 5-10 percent of exports to the world—depending on the definition of the region or whether or not oil is included. While on the face of it the answer to the above-stated question would seem to follow from the numbers, some authors have rightly put the question to a more solid test, essentially trying to find some benchmark that defines how much trade is “too low,” since the simple share of exports or total trade may not account for size of markets and other determinants of trade. This has been done by World Bank (1993), Deardorff (1995), and Ekholm, Torstensson, and Torstensson (1995), using regression estimates of gravity models that define a “natural” level of trade between two partners given such determinants as size, level of development, distance, and historical or cultural affinities. A predicted value higher than the actual value indicates a potential for more trade. There are differences in the models, so they obtain different results as the summary in Table 7 shows. A second kind of calculation, the trade intensity index (as defined in the table), normalizes for a country’s global involvement in trade, so a propensity greater than 1 means that the proportion of world exports to that country is greater than its proportion of world imports. This index is not so straightforward to interpret as the gravity model results: for example, one would certainly expect that neighbors, given lower trans-port and trade costs, would trade more with each other, and thus have an index above 1—but how much above? There is no natural benchmark, and we have therefore shown in Table 7 the index for other regions in the world, including some for much earlier periods.

Table 7.Mediterranean Intraregional Trade: Estimates of Potential Levels and Regional Comparisons
Ratio, of Predicted

to Actual
Increase as Percent of ExportsAs a Percent of Regional
Gravity Models
Israel4.74Very high3.7InfiniteNear zero base
Syrian Arab Republic7947.5
All others0.37-0.853.94.8
Trade Intensity Indices for Intraregional Trade1
A. Latin America Southern ConeE. East Mediterranean
Paraguay11.24-31.32F. Mediterranean by country—1995
B. Australia and New ZealandIsrael0.45
C. EEC2Syrian Arab

D. Japan and East Asia2
Sources: Ekhoim, Torstensson, and Torstensson, 1995; Deardorff, 1995; World Bank, 1993; Primo Braga, Safadi, and Yeats, 1994; Kojima, 1964; Drysdale, 1969; Awartani and Kleinman, 1995; and calculations using the IMF’s Trade Analysis and Reporting System.

Trade intensity is defined as the share of one country’s exports to a partner divided by the share of world exports to that partner. That is TI (ij) = [X(ij) / X (it) I WX (ij) / WX (it) ], where X (ij) and WX (ij) are the value of country is and world exports to region j, respectively and X (it) and WX (it) are the value of total exports of country i and the world, respectively.

1990 values were estimated using 1960 values and multiplying by the incerease of the simple intraregional shares shown in Frankel, Stein, and Wei (1995).

Sources: Ekhoim, Torstensson, and Torstensson, 1995; Deardorff, 1995; World Bank, 1993; Primo Braga, Safadi, and Yeats, 1994; Kojima, 1964; Drysdale, 1969; Awartani and Kleinman, 1995; and calculations using the IMF’s Trade Analysis and Reporting System.

Trade intensity is defined as the share of one country’s exports to a partner divided by the share of world exports to that partner. That is TI (ij) = [X(ij) / X (it) I WX (ij) / WX (it) ], where X (ij) and WX (ij) are the value of country is and world exports to region j, respectively and X (it) and WX (it) are the value of total exports of country i and the world, respectively.

1990 values were estimated using 1960 values and multiplying by the incerease of the simple intraregional shares shown in Frankel, Stein, and Wei (1995).

While there is room for interpretation, and indeed some analysts do conclude that the level of intraregional trade in the Mediterranean is not “too low” except for that between Israel and Egypt (Ekholm, Torstensson, and Torstensson, 1995; Yeats, 1996; and Hoekman and Djankov, 1996a, citing the previous two), the evidence of Table 7 seems to support the conclusion that it is indeed too low. Only the gravity model estimate of Ekholm, Torstensson, and Torstensson (which uses a smaller control sample) strongly suggests otherwise. Yeats’s (1996) conclusion that trade intensities are much greater than 1 for Middle East countries has to be set against the values for other regions. In Latin America for 1990, when the old unsuccessful arrangements were effectively nonfunctional, intraregional trade of most countries was far higher than in the Mediterranean group, and only Chile had levels as low as some of the higher ones in the Mediterranean group. The values for East Asia and the EEC in the 1960s were similar to those of the Mediterranean region, but once we adjust this for the increased intra-European trade since then (those simple shares within the EEC rose from about 35 percent to nearly 50 percent), the values are higher. For the East Asian and European regions, the high end of the range is probably a more appropriate comparison as the very large size of the larger European economies (which have the lower intensity ratios) makes them an unlikely control group for the Mediterranean countries.

There is certainly no doubt by any of the analysts concerning Israel’s very low integration into Mediterranean trade, but even for the other economies the balance of the evidence appears to indicate that a more “natural” level of trade would be quite a bit higher than at present. Given the various differences of approach and the lack of conceptual firmness in defining the natural level of intraregional trade, we see little purpose to a summary prediction indicating the order of magnitude by which trade might increase. Rather, it may be more useful to consider, first, what factors have in the past inhibited the full potential from being realized, and second, whether the implementation of the MEDA would contribute to removing those inhibitors.

Achieving the Full Potential for Intraregional Trade

Of the possible explanations for too little intraregional trade, the literature indicates three as being most important:

In addition, there is the very specific situation of Israel and the West Bank and Gaza Strip for which the major explanation of lower intra-regional trade, in the consensus view, is the political situation. There is little this study can add to the substantial writings on the issue of integrating Israel except for the quantitative information given throughout the analysis, which includes Israel either separately or as part of the totals for the region.

Perhaps it is best to deal first with the issue of similar comparative advantages. While the relatively low to middle level of industrial development of most of these economies would make them appear similar to each other at first glance, with comparative advantage in the same labor intensive goods, the quantitative indicators summarized in Table 8-Table 10 do not bear out such a perception. Table 8 shows calculations of a complementarity index, which matches up at a detailed product level the exports of a given country with the imports of a receiving country: the index is 0 when no good exported by one country is imported by the other, and 1 when the export shares of one match exactly the import shares of the other. This index was developed in World Bank (1995a) to address the “new regionalism” question, in the sense that the higher this index—as for historically successful arrangements—the better the prospects for regional liberalization being trade creating rather than trade diverting. The index for all the Mediterranean countries including Israel and Turkey is in fact quite high in comparison to the others, at 0.45 approaching the level of NAFTA and the early EEC. Excluding these two countries, the index is much lower, though still more like that of the South American and East Asian regions, and much above those for old “unsuccessful arrangements” and the present values for sub-Saharan Africa. It is imperative to interpret these numbers cautiously, for they certainly are not predictors of success for regional preferential arrangements. Rather, they simply answer the question: how similar are the products that countries of the region export to those that the countries of the region import? The answer, relative to other parts of the world, is fairly similar. Markets for each other’s exports can therefore surely be found, even without the likely spillover that MEDA would provide. That is, new products and more efficient production result from the resource allocation forced by MEDA and create potential intraregional trade, since there is good reason to believe that each country would find different products to export.

Table 8.Trade Complementarity Indices for Selected Trade Arrangements1
Earlier ArrangementsIndexRecent ArrangementsIndexPotential ArrangementsIndex
Successful arrangementsNAFTA0.56Americas “AFTA”
EEC0.53MERCOSUR0.29(NAFTA+5 countries)0.31
Canada-U.S. Free Trade0.64Asia-Pacific “APEC”
Arrangement(17 countries)0.35
Unsuccessful arrangementsSub-Saharan Africa
LAFTA0.22(20 countries)0.09
Andean Pact0.07All Mediterranean20.45
excluding Israel
and Turkey0.27
Source: For Mediterranean, computations using the IMF’s Trade Analysis and Reporting System; for others, World Bank, 1995b.

The complimentarity index is defined as C(jk) = 1 - [sum (IM(ik) - X(ij)1)/2], where X(ij) is the share of good i in total exports of country j, and M(ik) is the share of good i in total imports of country k.

Includes Algeria, Morocco, Tunisia, Egypt, Israel, Jordan, Syrian Arab Republic, and Turkey.

Source: For Mediterranean, computations using the IMF’s Trade Analysis and Reporting System; for others, World Bank, 1995b.

The complimentarity index is defined as C(jk) = 1 - [sum (IM(ik) - X(ij)1)/2], where X(ij) is the share of good i in total exports of country j, and M(ik) is the share of good i in total imports of country k.

Includes Algeria, Morocco, Tunisia, Egypt, Israel, Jordan, Syrian Arab Republic, and Turkey.

This last point is supported by the comparisons of the revealed comparative advantage index (RCA) shown in Table 9 and 10. A value greater than 1 for a given product means the country exports relatively more of that good than the world imports. We have calculated these indices for each country at a 3-digit SITC (Standard International Trade Classification) level of detail (128 product groups) and correlated the vector of indices between each pair of countries in the region. The results in Table 9 suggest that for the most part the structure of revealed comparative advantage is very different in the various countries of the region. The same conclusion is reached in Hoekman and Djankov (1996a), who calculated the RCA index for exports to Europe only. High correlations are seen only for Tunisia and Turkey. Medium correlations of 0.3 to 0.5 are observable in four cases: Algeria-Syria, Tunisia-Jordan, Syria-Egypt, and Israel-Jordan. But in most of these cases the correlation is lower for the years 1991–94 than it was for the preceding five-year period, and the results of Hoekman and Djankov (1996a) for exports only to the EU also show lower correlations in the second period. The remaining 31 paired correlations are in the very low range, below 0.3.

Table 9.Global RCA Correlation Coefficients, 1991–941
AlgeriaMoroccoTunisiaEgyptJordanLebanonSyrian RepublicIsraelTurkey
Egypt0 1780.0700.0501
Syrian Arab Republic0.3070.0010.0730 3040.025-0.0101
Source: Computations using IMF Trade Analysis and Reporting System data.

The revealed comparative advantage index (RCA) is defined as the share of a country’s exports of product j divided by the share of world exports of product j.

Source: Computations using IMF Trade Analysis and Reporting System data.

The revealed comparative advantage index (RCA) is defined as the share of a country’s exports of product j divided by the share of world exports of product j.

A further indication of the dissimilarities, and hence flexibility, of intraregional trade opportunities comes from correlating the RCA index for the mid-1980s with that for the early 1990s (see Table 10). For global exports, there appears to have been a continuity of the structure, with Algeria and Egypt showing modest change but more than the others.18 But the change in RCA over time for intraregional exports is substantial, even during a period when the level of such trade was probably inhibited to below potential. The implementation of MEDA’s free trade with Europe may create some amount of trade diversion away from the region, but the broader objectives of the strategy, including reforms of domestic markets and parallel MFN liberalization, would mitigate any diversion. Indeed, the stronger these accompanying measures, the more likely that the MEDA effect would switch from a small amount of trade diversion to complementary additional regional trade.19

Table 10.Correlation Coefficients of RCA in 1991–94 to RCA in 1984–87
Global ExportsRegional Exports
Syrian Arab Repubic0.9340.333

The possible effects of high protection against intraregional trade are not easy to identify because in many cases the MFN tariff levels, which are shown in Figure 3, do not in fact apply de jure in the various existing arrangements such as the Arab Maghreb Union (AMU), but the implementation of these, as for example noted in Zarrouk (1996), is uncertain. Taking the values of Table 8 for global comparison, we see that, apart from Turkey and Israel, protection remains moderately high. If it is not substantially reduced soon, the gap vis-à-vis the EU will grow, and of course the trade diversion effects will be higher. It is correctly said by many observers that pressures by local importers, who will bear the cost of trade diversion, as well as the exporters, who will lose some markets, will inevitably lead to policy decisions to accelerate rest-of-world liberalization as well, be it MFN all at once, or regional free trade followed by further MFN cuts. The size of the European partner will make this effect important, but it will not be automatic or come quickly, and for two reasons it may be useful for the authorities to consider preempting these pressures by moving ahead in advance. First, the diversion costs will apply more to consumer goods than industrial inputs or equipment because in the tariff structures of these countries the latter goods have low and sometimes zero tariffs. Second, third parties feeling losses from the trade diversion effects of MEDA may not wish to raise the issue, either informally or in WTO procedures, because for many of them—the U.S., Japanese, and East Asian economies—the size of these markets will remain of limited importance.

Figure 3.Weighted Average Tariff, 19961

(In percent)

1 Except for Algeria, 1992.

Figure 4.Intra-lndustry Trade: Manufacturing Sector

(Weighted average, 1991–94)1

1 Except for Central Europe, 1993, and East Asia, 1978.

We finish this argument on intraregional trade prospects by relating it to the important hub-and-spoke problem noted earlier. The wider the gap grows between EU and regional tariff levels, the more likely the hub-and-spoke phenomenon is to come about. Conversely, the sooner regional and global protection approaches the MEDA levels, the more likely that European and other foreign investors will seek to invest, not in Europe to service several isolated markets but in one of the Mediterranean countries to service the entire market: EU, the country in which it is located, and the region. The expansion of intraregional trade that would follow is self-evident, and would begin to fulfill the potential for intraregional trade that is indicated by most quantitative and qualitative analyses.

The possibility that trade costs are particularly biased against regional trade has been noted by many, but these are very difficult to measure. The World Bank study Claiming the Future (1995b) describes the burdensome process of port and customs procedures in Egypt, which is characteristic of many other situations in the region. That the region’s road, telecommunication, and even sea connections are poor is not in question, but one cannot expect substantial infrastructure investments in the region before the other factors that have inhibited intraregional trade are ameliorated, in particular the remaining protective barriers, including unimplemented elements of preexisting regional agreements. If, in contrast, the openness strategy includes rapid structural reforms internally, and regional-plus-global liberalization externally to avert the hub-and-spoke effects, production investments—both foreign and domestic—are more likely to be sited in Mediterranean countries. This will be the signal for governments, private investors, and financial markets that infrastructure investments increasing the regional connections are also a good investment.

In the next, concluding section we return to a point already raised: how to make the elements of the strategy not only optimal but also credible to markets, so that the various investments will begin as soon as possible.

Options for a Trade Policy Strategy in Mediterranean Countries

Instead of attempting to summarize the preceding discussion, which encompasses a wide range of possibilities and a comprehensive list of policy measures, it may be more useful to conclude by focusing on a narrower set of potential strategy paths. To simplify the consideration of policy choices available to Mediterranean country governments, we define four “scenarios,” as set out in Table 11.

Table 11.Schematic for Trade policy Strategy
EUFTA (Reforms as feasible)(EUFTA plus schedule of domestic reforms)(MEDA 1 plus schedule of regional liberalization)(MEDA 2 plus schedule of global liberalization)
CredibilityModest only, since reforms and rest-of-world liberalization uncertainModerate but increasingly stronger as reforms progressStronger, especially if global liberalization expected to followVery strong, as long as implementation steady
CostsFiscal loss strongFiscal loss lessened by tax reforms, higher revenue baseFiscal loss marginally higher than MEDA1Fiscal loss disemployment restructuring perhaps marginally above MEDIA1 and MEDA2, but greater credibility may give quicker and advance reaction by markets
DisemploymentDisemployment, but shorter periodDisemployment, perhaps marginally above MEDA 1
Prolonged restructuring costsRestructuring cost but shorter periodRestructuring cost as in MEDA 1
Trade diversionTrade diversionLess trade diversionNo trade diversion
BenefitsSome long-run efficiency of reallocationGreater long-run reallocation gainsVarious dynamic gains as in MEDA 1, but a little strongerVarious dynamic gains as in MEDA 1 and MEDA 2, but stronger and earlier
Possibly limited new investmentSome endogenous efficiency and productivity improvementsOvercomes much of hub-and-spoke effects; therefore more investment, foreign and domesticFully overcome hub-and-spoke effects and obtain greatest investment increase
Some “harmonization” gains
Increased investment, including some foreign
RisksStrong risk of fiscal and external imbalancesReduced risk of fiscal and external imbalanceReduced risk of fiscal and external imbalancesMinimized risk of fiscal and external imbalances, except for unexpected shocks
Producers continue or return to protectionist mentalityReduced risk of protectionist mentalityMuch reduced risk of protectionist mentalityVirtually overcome risk of protectionist mentality
Very high risk of hub-and-spoke effectsHigh risk of hub-and-spoke effectsLow risk of hub-and-spoke effectsVirtually overcome risk of hub-and-spoke effects
  • 1. An EU free trade arrangement (EUFTA) would involve a minimalist application of the Agreements with Europe, implementing the free trade in 12 years, but with domestic structural reforms implemented only on a “wait and see” basis, as circumstances permitted.
  • 2. MEDA1 would include a strong commitment to the broader aspects of the Euro-Mediterranean initiative as described in the Barcelona Declaration (Euro-Mediterranean Partnership, 1995), making more concrete both the schedule and targets for their implementation; this scenario assumes appropriate supportive action of the EU and others for comprehensive domestic structural reforms.
  • 3. MEDA2 would go beyond MEDA1 and would not only strengthen the domestic reforms and accompanying measures, but would include a clear target-dated commitment to regional liberalization roughly parallel to the EUFTA.
  • 4. MED A3 adds to MEDA2 a commitment to, and a clear schedule for, global liberalization, in effect augmenting and accelerating the present Uruguay Round commitments.

In Table 11 we summarize the implications of the analysis in this paper by noting for each of these four scenarios the implications for credibility of the government strategy, what benefits each scenario may yield, what costs will be faced, and finally, the risks that each scenario entails. The preceding discussion has described the appropriate nature of structural domestic reforms and what concrete policy measures are required, so we need not repeat this here in the form of recommendations. Using Table 11, which is largely self-explanatory, we address here seven policy-related questions that are most often asked.

1. What is the optimal trade policy strategy for Mediterranean countries to maximize the gains from the Agreements with the European Union?

The most risky and least appealing option in Table 11 is the first, EUFTA, because too little structural reforms and other accompanying measures would mean low credibility, limited benefits, high costs, and high risks of unsustainability. Virtually all analysis concurs that the benefits come in the form of future dynamic gains whose magnitude and timing depend crucially upon the improved flexibility of markets to react to the new opportunities, as well as on the credibility of the strategy. A strategy with only the asymmetrical opening up to competition from the EU has little to recommend it. This conclusion is not new, of course, and is fully recognized in the broader approach to the initiative implicit in the Agreements and the Barcelona Declaration. This broader view we have been designating by the shorthand MEDA.

At a minimum, however, countries would want to adopt MEDAl or even one of the options that adds accelerated liberalization toward third countries. MEDAl already has a great deal more to recommend it as long as the accompanying structural reforms are implemented in good time. The costs are probably less than with the first option, as the likelihood of larger and earlier gains, with higher growth, raises the revenue base and probably accelerates the pace of endogenous restructuring. But one very important risk remains that makes the MEDA2 and MEDA3 options appealing: the hub-and-spoke effect on investments.

For this reason, a more clearly stated schedule for parallel external liberalization should be given serious consideration by governments. The costs in the short run may be somewhat higher, though this is not inevitable—especially with MEDA3, which would virtually eliminate any trade diversion costs. But even if these costs are higher, they will be only marginally so, with dynamic benefits much increased because of a stronger competitive stimulus, greater credibility and hence earlier efficiency gains, and virtual offsetting of the very high risk of hub-and-spoke effects. Thus, while the benefit-cost ratio under MEDAl is doubtless positive, which justifies opting for such a strategy—as signatories have in principle already done—it is likely that the benefit-cost ratio of MEDA2 and certainly that of MEDA3 may be higher still. Of added importance, MEDA3 most fully captures the spirit of the courageous choice already made by the signatories toward an open economy.20 The choices are not of course to be made on such considerations alone, but a clear statement of the implications from each option will be helpful in making those choices.

2. What are the key factors determining the extent of potential benefits?

Perhaps the most important factors determining the extent of benefits are financial stability, structural reforms, and outward orientation. Structural reforms include improvements in the social safety net to help deal with short-run adjustments, freeing up market forces to enable the players in the economy to react freely and swiftly to the stimulus of the EUFTA, thereby realizing the reallocation gains that are the first major benefit expected. Delays in these reforms may cause bottlenecks in supply reactions, and risk slowing or reversing the process. The financial “omnibus” that will carry most of the needed restructuring in production will be investment. Outward orientation has increased considerably in Mediterranean countries, with Tunisia and Morocco among the more advanced. Nevertheless, to avoid trade diversion as well as hub-and-spoke investment patterns, a timely move to regional and multilateral liberalization is essential.

One might consider another factor, which is not a policy measure but a policy quality: credibility. Credibility can be enhanced by concrete steps, such as a clear explanation of long-term policy goals and of the strategy in general, consistent implementation of needed structural reforms, and announcement and maintenance of a schedule of reforms. This is what the numerous writers who mention the usefulness of MEDA as an “anchor” have in mind—a tying-in that lends credibility to the governments’ policies.

How ready are the Mediterranean countries to achieve the potential benefits, or to put it differently, how close are they to the necessary climate of financial stability, market orientation, outward orientation, and policy credibility? While “readiness” is an ambiguous concept, the analyses of Nsouli, Bisat, and Kanaan (1996) and Havrylyshyn (1996) give some broad indications: all Mediterranean countries must keep watch over financial stability; all, despite some early efforts, have a considerable way to go on structural and market-oriented reforms; and, finally, all retain medium to high levels of multilateral protection. But this should not be interpreted to mean that they are not ready for initiatives such as MEDA. Rather, the task undone defines the challenges ahead in drawing greater gains from MEDA.

3. What elements are most important in future reviews and implementation of the Agreements, as well as for Agreements not yet completed?

The Agreements are incomplete in several ways, allowing flexibility for improvements as experience is gained. Some are still in discussion and formulation; Article 18 of the existing Agreements explicitly calls for a review in the year 2000 on agricultural liberalization; and many aspects are only broadly specified and can be made more concrete through the continued dialogue between the parties. Some of the most important elements in future dialogue are: (1) continuing the process of agricultural liberalization even if empirical estimates suggest net gains may be small, because one should not underestimate the possible “pleasant surprises” that trade liberalization has always brought historically, and because such a step would greatly enhance the credibility of the Agreements; (2) formulating a more concrete set of goals for several of the “deeper integration elements,” such as standards, harmonization of regulations, and liberalization of trade in services; and (3) ensuring that financial support levels are related to the pace of implementation of reforms and supportive of orientation to the market, an enhanced private sector, and an open economy.

4. What are the risks of adjustment and restructuring programs aimed at easing the process of resource reallocation?

Special programs to assist affected firms in their restructuring process are in existence, and more are planned. It is well recognized by the governments that these programs risk being mistaken as merely another form of the earlier state-support and protectionist strategies. The size of these programs has so far been limited, bearing in mind this risk. To further minimize the risk of such misperceptions and of capture by vested interests, a termination date should be set to send a clear signal to those involved. The extent and progress of structural reforms in general will also determine the extent of this risk—yet another reason to make sure the strategy includes a schedule of implementation for reforms.

5. How serious are the problems of trade diversion and hub-and-spoke effects, and how might they be offset?

If the hub-and-spoke risks are not overcome, investments may be too low to achieve the needed momentum of restructuring and growth. Taking the additional step of third-party liberalization is perhaps the most effective way to offset this risk, though faster, deeper, and more credible structural reforms will also help. It should be noted that the concerns expressed by most analysts on this matter are not predictions, and if structural reforms and the consequent reallocation of resources take place quickly and are substantial, the hub-and-spoke effects may be much reduced. But the consensus appears to be that there is a very high risk, so a wiser policy course would be to consider measures with a greater chance of success, namely, an acceleration of liberalization to third countries—at a minimum, those in the region. The timing of this ancillary liberalization is not unimportant. The present vagueness on how much and when countries would extend liberalization risks postponing liberalization until too late, which diminishes the credibility of the openness strategy and results in a strong hub-and-spoke effect. Third-party liberalization should not be too far out of step with the 12-year calendar for the EUFTA.

6. What is the role of the EU in promoting the strategy and implementation path that would maximize the benefits for Mediterranean countries?

The EU, appropriately, would not wish to limit the policy options of Mediterranean governments as they proceed to implement the Agreements. But without any risk of doing so, the EU can still help in several ways to promote the maximum benefits. First, it should continue to show flexibility in the regular review and dialogue process. Second, it should, in cooperation with Mediterranean partners, make progressively more concrete the principles of “deeper integration,” to avoid this being delayed too long, which would in turn reduce the dynamic gains. Third, in dialogue with the partners, the EU should move as quickly as possible to formulate and institutionalize more clearly the mechanism for financial and economic cooperation, in particular to ensure the mechanism’s harmonization with the progress in structural reforms. In this last, both of the parties would wish to maintain effective collaboration with multilateral institutions and other bilateral donors that could contribute to the effective implementation of the opening-up strategy.21

7. What policies would best promote the full potential for intraregional trade?

Intraregional trade is certainly below potential levels, not only in Israel and the West Bank and Gaza Strip, but also in all other countries in the region. The long-standing goal of promoting such trade and other parallel forms of cooperation is not at all threatened by the MEDA initiative but, if the appropriate accompanying measures are followed, the initiative can, as a corollary, enhance such trade. The Agreements themselves clearly specify this as an objective, but there may not be enough concrete mechanisms to ensure that it happens. A moderate degree of success in attaining the expected reallocation and dynamic efficiency gains would likely have some positive spillover effect on intraregional trade, offsetting any trade diversion, because more efficient production could become exportable to the region. However, such an effect is likely to be very minimal. More certain gains can come from implementing an accelerated opening up within the region and to the rest of the world, on an MFN basis. The benefits would probably be greatest if this were done as one step encompassing both the region and all other countries at the same time, but if a first step of regional liberalization were to be associated with a clear commitment and target date for the second step to be implemented globally, a considerable proportion of the benefits could still be expected. Even more important on the timing, as noted already, the calendar for regional and global liberalization and its magnitude cannot differ very much from the 12-year calendar for free trade with the EU.

    Other Resources Citing This Publication