the economic and political determinants of imf and world

World Development Vol. 34, No. 2, pp. 247–270, 2006� 2005 Elsevier Ltd. All rights reserved

Printed in Great Britain
0305-750X/$ - see front matter
doi:10.1016/j.worlddev.2005.07.016www.elsevier.com/locate/worlddev

The Economic and Political Determinants

of IMF and World Bank Lending in the

Middle East and North Africa

JANE HARRIGANSchool of Oriental and African Studies, University of London, UK

CHENGANG WANGUniversity of Bradford, UK

School of Oriental and African Studies, University of London, UK

and

HAMED EL-SAID *

Manchester Metropolitan University, UKSchool of Oriental and African Studies, University of London, UK

Summary.— This paper assesses the economic and political determinants of IMF and World Bankprogram loans to the Middle East and North Africa. First we assess what is already known aboutthe geo-political influences on aid flows to the Middle East and North Africa (MENA) region andthe potential for this to operate via the IMF and World Bank. From this we conclude that there isscope for IMF and World Bank lending in the region to respond to the political interests of theirmajor shareholders, particularly the United States. We support these arguments with both a qual-itative and a quantitative analysis of the determinants of World Bank and IMF program lending tothe region, focusing on both economic need in the MENA countries and the politics of donor inter-est before concluding.
� 2005 Elsevier Ltd. All rights reserved.
Key words — IMF, World Bank, Middle East and North Africa, aid allocation

* The UK Department for International Development

(DFID) supports policies, programs, and projects to

promote international development. DFID provided

funds for this study as part of that objective but the

views and opinions expressed are those of the authors

alone. Thanks to Bernard Walters for comments on an

earlier draft. Final revision accepted: July 25, 2005.

1. INTRODUCTION

The Middle East and North Africa (MENA)consists of the predominately Islamic culturesof the Gulf Arab countries, the Levant, thecountries of North Africa, plus Iran, and themore industrialized country of Israel. 1 MENAassumes both political and economic signifi-cance. Politically, it is arguably the epicenterof world crisis, chronically war-prone, and thesite of the worlds most protracted conflicts(Hinnebusch, 2003, p. 1); economically, it ownsthe bulk of the world’s oil reserves, driving inparticular the USA economic engine. In lightof the region’s geo-politically and economically

247

strategic position in the world economy, it isclear that economic and political factors areinextricably linked when it comes to the man-ner in which the West, particularly the UnitedStates, responds to the region’s needs.

248 WORLD DEVELOPMENT

There is a long and rich theoretical andempirical literature on the determinants of thegeographical allocation of foreign aid. 2 It isgenerally accepted that this allocation is influ-enced by both recipient need and donor interestand that multilateral aid is less susceptible todonor interest than bilateral aid (Maizels &Nissanke, 1984; Rodrik, 1995). In the pastdonor interest has often reflected the geopoli-tics of the Cold War, with pro-western regimes,regardless of economic need and their recordon human rights, being large recipients of wes-tern aid. 3

Even before the collapse of Communism inthe late 1980s and early 1990s, a new theorywas emerging to the effect that ‘‘Islam [is] thenew Communism and [hence represents] agrave threat to Western civilization’’ (Niva,1998, p. 27). Consequently, ‘‘rogue states’’ wereisolated while pro-western regimes, particularlyif they were threatened by Islamists, were re-warded for serving Western interests (Hubbell,1998, p. 9). Hence, the end of the Cold War re-placed the old dichotomy in the Arab Worldbetween conservative pro-Western and socialistpro-Communist Arab regimes with a new andless covert formula based on ‘‘friends or allies,or good or bad’’ regimes (Perthes, 1998, p. 30).It is possible that past aid allocations to

MENA have been influenced by United Statesinterests in the region, and that the IMF andWorld Bank are not immune from such influ-ences. It is often argued, particularly by theanti-globalization movement, that the twoWashington-based multilaterals are stronglyinfluenced by the economic and political needsof their major western shareholders, especiallythe United States. This influence can take twoforms—determining the geographical flow offunds, that is, who gets what from the IMFand the World Bank; and influencing the condi-tionality attached to such funds, that is, pro-gram loan recipients are expected to undertakeeconomic liberalization programs, which helpto open up their economies to the global eco-nomy and Western economic penetration. Inaddition, we can speculate that if there is evi-dence that past financial flows into pro-westernMENA countries have responded to donorinterest rather than recipient need, then, giventhe post 9/11 foreign policy concerns of thewest, this may well intensify in the future. 4

In light of the above, this paper attempts toassess whether donor interest, particularly thepolitical interests of the United States, haveaffected the flow of funds from the IMF and

World Bank to the MENA region. Althoughthere is a large body of literature on the multi-ple determinants of aid allocation, much of theempirical work does not disaggregate aid bydonors and when it does it tends to focus onbilateral donors. In addition, the more recentempirical work tends to employ panel datawhich aggregates recipients. Although there isa small but growing literature on the influenceof the political preferences of the IMF andWorld Bank’s principal shareholders on lend-ing decisions (Barro & Lee, 2001; Bird & Row-lands, 2001; Fleck & Kilby, 2001; Killick, 1995;Rowlands, 1995; Thacker, 1999) to the best ofour knowledge, this is the first paper which at-tempts to specifically analysis this phenomenonin the geopolitically strategic MENA region. Inaddition, the fact that we concentrate on spe-cific MENA countries enables us to capturepolitical influence in a manner somewhat differ-ent from the existing studies which tend to lookat IMF lending in aggregate. Such an analysisis timely given the new foreign policy interestof Western powers in MENA.The remainder of this paper is divided up as

follows. In the next section, we assess what isalready known about the geo-political influ-ences on aid flows to the MENA region andthe potential for this to operate via the IMFand World Bank. From this we conclude thatthere is scope for IMF and World Bank lendingin the region to respond to the political interestsof their major shareholders. We support thesearguments with both a qualitative and a quan-titative analysis of the determinants of WorldBank and IMF program lending to the region,focusing on both economic need in the MENAcountries and the politics of donor interest be-fore concluding.

2. POLITICS, AID, AND MENA—WHATDO WE ALREADY KNOW?

Three facts are already established in the lit-erature—bilateral aid flows are influenced bydonor political interest; flows into MENA,most notably Egypt and Israel, are partly polit-ically determined; and the United States wieldsconsiderable power and influence in the IMFand World Bank.

(a) Trends in aid flows to MENA

The MENA region has been the second larg-est regional recipient of aid in the period since

ECONOMIC AND POLITICAL DETERMINANTS 249

1960. From 1960 to 2001, the MENA region re-ceived nearly $329 billion of aid (in 2000prices), which only its poor neighbor sub-Saha-ran Africa exceeded by a large margin. 5 Interms of the importance of different donors tothe region, the United States championedMENA in the 1960s, 1980s, and 1990s, whilethe GCC was the largest donor in the 1970s(see Figure 1). The multilateral donors havebeen less important to the region than the bilat-eral donors and although the World Bank is thelargest multilateral donor its role pales intoinsignificance compared to that of the UnitedStates, as shown in Table 1. The same is trueof the IMF (Appendices A and B provide de-tails of all IMF and World Bank program loansto MENA countries).

(b) Donor interest as a determinant ofaid flows to MENA

Donor interest seems to play a significantrole in aid allocation to MENA. Many aidallocation studies based on models which incor-porate variables representing both donor inter-est 6 and/or recipient need have reached theconclusion that donor interest is an important

-1000

0

1000

2000

3000

4000

5000

6000

7000

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

Figure 1. The source of net ODA disbursement (US$ m

Note: Non-DAC is taken

determinant of the geographic allocation ofaid, especially on the part of bilateral donors(Berthelmy & Tichit, 2002; Feeny & McGilliv-ray, 2002; Frank, 1969; Hayter, 1971, 1981;Hensman, 1971; Jalee, 1968; Maizels & Nis-sanke, 1984; McGillvray, 2003; McKinlay &Little, 1977, 1978, 1979). Our own recent study,based on a mathematical model of the aid allo-cation process and employing a fixed effectsmodel with panel data produced a similar result(Harrigan & Wang, 2004). 7

A number of the aid allocation studies intro-duce dummies to reflect specific strategic linksbetween donors and certain recipients. This ismost common in the context of MENA, wheredummies are often introduced for Egypt and Is-rael when the database includes these two coun-tries. Most of these studies find these dummiesto be positive and significant, for example, Ale-sina and Dollar (2002), Berthelmy and Tichit(2002), Feeny and McGillivray (2002) amongmany others. 8 The Egypt and Israel dummiesreflect that fact that these two countries arekey strategic allies to the West, especially theUnited States, such that donor interest is likelyto have a positive influence on aid allocations.According to Feeny and McGillivray (2002,

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

Mutilaterals

DAC Countries

Non-DAC Countries

illion-2000 price). Data source: DAC on-line database.

as proxy for GCC aid.

Table 1. US aid and World Bank aid to MENA (US$ million)

United States US share (%) WB WB share (%) Total

Africa—North of Sahara

60–69 9995.95 39 �49.3 25,61070–79 7545.67 16 1051.4 2 48,46680–89 16303.53 42 1191.7 3 39,08490–99 13785.49 34 308.2 1 41,1052000 629.13 29 22.8 1 2,1842001 586.09 24 �1.0 2,403

Middle East

60–69 5056.72 42 �76.4 11,96970–79 11554.3 22 582.0 1 51,54080–89 19787.32 33 691.3 1 60,86090–99 17554.76 44 606.7 2 39,7392000 1216.42 39 47.3 2 3,1412001 500.46 19 57.9 2 2,674

Data source: DAC online database.


p. 14) ‘‘Israel’s relationship with the UnitedStates is arguably one of the most intense be-tween a donor and a recipient.’’Our own work (Harrigan & Wang, 2004) has

enabled us to go beyond simple dummies forEgypt and Israel and to make more country-specific observations regarding the influenceof donor interest in aid allocations to MENAcountries. In our base regression, we have ap-plied a fixed effects model to panel data to ana-lyze the determinants of aid allocations byvarious classes of donor (with 2,484 observa-tions covering 32 years and 138 recipients).The donor–recipient fixed effect coefficientsfor MENA countries are reported in Table 2.These coefficients capture donor–recipient spe-cific effects, that is, they show the linkagesbetween donor and specific recipients, whichinclude long-term strategic relations, economic

Table 2. Donor–recipients fixed effec

Country United States

Egypt 0.87Iran �1.61Israel 6.03Jordan 2.41Lebanon 0.91Morocco 0.35Sudan �1.9Syria �0.34Tunisia 0.65Turkey �0.6Yemen �2.24

Source: Harrigan and Wang (2004).a Non-United States is aid from all OECD DAC members

linkages, colonial ties, etc. As can be seen fromTable 2, donor interest, as represented by thefixed effects coefficient, has a strong positive ef-fect in the allocation of US aid to Israel andJordan, two of the most strategically importantUS allies in the region, and a strong negative ef-fect on US aid allocation to Iran, Sudan, andYemen, countries traditionally hostile to USforeign policy in the region. 9 As expected, thefixed effects coefficients for multilateral aid aremuch smaller, although links between multi-lateral aid and Egypt, as well as Jordan andLebanon, are evident. Interestingly, Israel isnot favored by the multilaterals, in contrast toits special relationship with the United States.Another recent study of aid allocation, which

enables conclusions specific to some of theMENA countries is that of Collier and Dollar(2002). In their paper, they compare the actual

ts coefficients for MENA countries

Non-United Statesa Multilateral

1.56 0.72�0.21 �0.831.4 �1.220.85 0.670 0.72

1.23 0.25�0.24 0.18�1.45 �0.491.34 0.391.46 0.15�0.55 �0.23

excluding United States.

Table 3. Optimal versus actual 1996 aid allocation in selected MENA countries (per cent of GDP)

Country Poverty-efficient allocation Actual aid

Jordan 0.0 3.26Egypt 0.0 1.31Morocco 0.0 0.70Tunisia 0.0 0.29Algeria 0.0 0.22

Source: Collier and Dollar (2002, Table 3).


allocation of aid with an optimal poverty-effi-cient allocation of aid, with the latter assumedto depend on each recipient’s level of poverty,the elasticity of poverty with respect to income,and the quality of its policies. Comparing ac-tual 1996 aid allocations with the optimum theyfind that a large number of middle-incomecountries with poor policy receive excessiveamounts of aid. As shown in Table 3, the signif-icance of their results in the context of MENAis that all of the MENA countries in their sam-ple of 59 developing countries should not re-ceive any aid on the poverty-efficient criteria.Jordan in particular stands out as a countryreceiving aid equivalent to 3.26% of its GDPas opposed to 0% under the optimal Collierand Dollar allocation.The above types of studies have led to the

general conclusion that MENA is over-aided.In the words of DFID:

‘‘In comparison with other regions, it (MENA) re-ceives substantially more aid per poor person . . .but poverty reduction is not the primary motivationfor many donors’ assistance to MENA. . .aid alloca-tions are substantially influenced by donors’ domes-tic political considerations, including commercialadvantage and foreign policy objectives such asmigration and terrorism.’’ (DFID, 2003, p. 11)

(c) Donor interest and the IMFand World Bank

Although there seems to be ample evidencethat bilateral aid flows to the MENA regionare influenced by donor interest, a commonview is that flows from multilaterals such asthe IMF and World Bank are less likely to beinfluenced in this way since they do not repre-sent the interests of any particular country.Indeed, the Bank and Fund Articles of Agree-ment explicitly state that lending decisionsshould not be influenced by political factors. 10

However, in reality, it seems that this has onlyreduced the level of politicization in theiractions rather than removed it altogether.Therefore, it is possible that IMF and World

Bank lending provide donors with anotherarms length instrument to pursue their owninterests. This might include the disbursementor withholding of IMF and World Bank fundsto reward or punish recipient behavior accord-ing to the interests of the major shareholders ofthese two institutions.Analysis of the voting power of the United

States in the IMF and World Bank clearlyshows that the United States has the capacityto exercise leverage in these two institutions. In-deed, a common criticism of the operation ofIMF and World Bank is that their decisionmaking process is dominated by the G-8 coun-tries, especially the United States. The BrettonWoods institutions have systems of governancebased on weighted voting. Each member pos-sesses a number of votes, which depend on itsquota allocation and which must be cast as abloc. This leads to a problem of democraticlegitimacy since a member’s influence or votingpower within such a decision-making systemdoes not in general correspond to its votingweight.Using voting power analysis, 11 Leech and

Leech (2003) and Leech (2002) show that theUnited States possesses considerably morepower than voting weight in relation to ordin-ary decisions requiring a simple majority. Theyconclude ‘‘Weighted voting tends to further en-hance the power of the United States at the ex-pense of all other members in both the board ofGovernors and the Executive board’’ (Leech &Leech, 2003, Abstract) and ‘‘Our principal re-sult is that the voting power of the UnitedStates turns out to be far greater than its quotawould warrant’’ (Leech & Leech, 2003, p. 3).Given America’s voting power advantage in

the IMF and World Bank, Mckeown, Palla-nsch, and Thacker (1999) have argued thatthere is no reason why American policy makerswould not be expected to use this power inorder to promote adherence to US alliances;to secure the strengthening of a regime friendlyto the United States or to weaken a hostile re-gime by removing a source of support; and to


win trade or investment concessions. Conteh-Morgan (1990) and Zimmerman (1993) havemade a similar argument. The influence of theUS government is further evidenced by the factthat the American Executive Director ‘‘isordered by law to clear his or her decision withthe Secretary of the Treasury’ (Swedberg, 1986,p. 379) and each major decision must have theapproval of the US Senate and Congress(Smith, 1984).A number of empirical studies have at-

tempted to provide more rigor to the aboveanalysis by trying to identify and quantify thespecific determinants of IMF lending, 12 Joyce(1992), Conway (1994), and Knight and Santa-ella (1997) concentrated on economic determi-nants. Some consensus has emerged withrespect to key economic variables. Declines inexport earnings, high debt service ratios, andthe presence of arrears on debts, as well ashistories of other financial problems are allassociated with a higher likelihood of an IMFagreement being signed.Based on a simple macroeconomic model,

Thacker (1999) conducted a test on whetherIMF lending is politicized by introducing vari-ables of political proximity (UN voting patternsimilarity between United States and recipient)and political movement (the shift of recipient’svoting pattern toward USA voting pattern).His results show that movement toward theUnited States within a defined internationalpolitical space can significantly increase a coun-try’s chances of receiving a loan from the IMF.The results from similar studies (e.g., Barro &Lee, 2001; Bird & Rowlands, 2001; Rowlands,1995) are generally consistent with Thacker’sfindings, but divergent regarding the level ofUSA influence. 13 For example, Rowlands(1995) concludes that the evidence for system-atic US influence is less strong than that com-monly expected.IMF lending may also be influenced by polit-

ical conditions in recipient countries, such asdemocratization or elections. Dreher and Vau-bel (2002) conduct a formal test of the politicalbusiness cycles using regression analysis. Theyfound that IMF credits in the more democraticrecipient countries are larger in pre-election andpost-election years, while the credits in moreauthoritarian regimes are marginally smallerin post-election years.It is not just the receipt of a loan that may be

influenced by western countries; the terms ofthe loan might also be affected. A series of casestudies conducted for a project by Killick

(1995, pp. 118–119) reveals that at least one-third of 17 countries studied secured favorableloan terms on their IMF programs due to theintervention of major shareholding countrieson their behalf. 14 Oatley and Yackee (2000)took the size of IMF loans as the dependentvariable and found that lending decisions wereresponsive to US pressure, with larger loansgoing to countries in which American bankswere highly exposed and to governments clo-sely allied to the United States.Regarding World Bank lending, Fleck and

Kilby (2001) used panel data to examine thegeographic distribution of the World Banklending from 1968 to 1992 and conclude that‘‘Countries with strong US trade ties receiveda significant share of the World Bank lendingthan comparable countries with weak ties;Countries which the United States favored withbilateral aid received a disproportionate shareof the World Bank funds as well’’ and such ef-fects ‘‘vary across US presidential administra-tions’’ (Fleck & Kilby, 2001, p. 16). In a morequalitative analysis of the World Bank, Scho-ultz (1982) has documented that the WorldBank’s ‘‘interests’’ in a given loan or aid pack-age are often influenced by the US Executiveand Congress.The above literature review has indicted that

donor influence, including strategic geo-politi-cal interests, influence aid allocations, espe-cially allocations in the MENA region, andthat the IMF and World Bank are not immunefrom the influences of their major shareholders,particularly the United States. We now turn tolook more specifically at the determinants ofIMF and World Bank program lending inMENA in order to assess to what extent ithas been determined by recipient need and towhat extent by donor interests. We adopt twoapproaches—a qualitative analysis of the tim-ing of the Bank and Fund loans and a moreformal quantitative analysis of IMF lendingbased upon a Probit model.

3. THE TIMING OF IMF AND WORLDBANK PROGRAMES IN KEY MENA

RECIPIENTS: A QUALITATIVEANALYSIS

(a) Recipient economic need

Although IMF and World Bank programloans have only been a small percentage of totalaid flows into the MENA region, they neverthe-


less have the capacity to influence the recipienteconomies via the loan conditionality. In thepast, such conditionality has brought with iteconomic reform via structural adjustmentand stabilization programs. A cursory glanceat economic performance in the region makesit clear that reform was indeed needed in manyMENA countries in the post-1980 period.In order to fully understand economic per-

formance in the region and the need for bothfinancial support and economic reform fromthe mid-1980s onwards (and hence a role forthe IMF and World Bank) it is necessary to di-vide performance into sub-periods. During the1973–81 period, the region’s wealth and indus-trial structure was concentrated on oil. Duringthis period, the region as a whole also enjoyedsubstantial inflows of so called Official Devel-opment Assistance (ODA) from DAC bilateraldonors, non-DAC bilateral donors, and multi-lateral donors. As shown in Table 4, the1970s was a golden period for the MENAregion. GDP growth averaged over 6% perannum, gross domestic savings and capital for-mation were a respectable 37% and 29% ofGDP, respectively, and both export and importcoefficients were high. On the back of thiswealth, public expenditure expanded with astrengthening of both state welfarism and stateeconomic activity.However, when oil prices softened in the

1980s, the structural weaknesses of the econo-mies in the region, especially the over-relianceon oil, became apparent. 15 As can be seenfrom Table 4, growth declined and per capitaGDP decreased by an average 1% per year inthe 1980s, a rate worse than any other develop-ing region, except sub-Saharan Africa. Othereconomic indicators also pointed in the samedirection: the saving rate and investment ratedropped in the late 1980s and the export/GDP and import/GDP ratios also declined.The disappointing economic performance of

the MENA region in the 1980s can be attrib-

Table 4. Selected economic indica

Year 1975–79 19

GDP growth (%) 6.2GDP per capita (constant 1995 US$) 1908.5 18Gross domestic savings (% of GDP) 36.7 2Gross capital formation (% of GDP) 29.5 2Export (% of GDP) 42.0 3Import (% of GDP) 34.8 3

Data source: WDI 2002 CD-ROM.

uted to a number of factors. Internally, a highpopulation growth rate, poor economic man-agement, corruption, and prolonged heavyprotection led to high unemployment and eco-nomic inefficiency. It is also worth noting thatduring the oil boom years, despite having highdomestic saving rates and high inflows of for-eign aid, investment in the MENA region wasmostly diverted toward consumption as wellas non-productive investment.The extent of the crisis can be seen when we

look at key macroeconomic indicators forcountries in the region who were to becomemajor recipients of IMF and World Bank pro-gram loans, namely, Egypt, Jordan, Morocco,Tunisia, and Algeria. Details of the IMF andthe Bank programs implemented in the abovecountries are listed in Appendices A and B,respectively. Macroeconomic indicators forthe above five countries indicate that inflationand current account imbalances were built upin 1970s and persisted throughout the 1980s,the central government debt and total debt ser-vice ratio were also high in the 1970s and be-came even worse in the 1980s. With the risingdebt and high inflation, gross capital formationstarted to decline from the early 1980s andnever looked like bouncing back to the peaklevel of late 1970s. In light of this fall in invest-ment, the decline in GDP per capita growthwitnessed in all five countries during the 1980sseems inevitable.However, a more nuanced analysis, which

looks at the specific timing of loans and corre-sponding macroeconomic indicators in eachrecipient, suggests that the determinants oflending often do not reflect recipient economicneed. Table 5 provides macroeconomic datafor each of the five countries for the year inwhich each received its first IMF loan and forthe previous five years (Egypt received two dis-tinct phases of loans and so is representedtwice). 16 By looking at the macroeconomicvariables for the year in which each country

tors of the MENA 1975–2000

80–84 1985–89 1990–94 1995–99 2000

2.3 1.3 4.4 2.96 3.982.3 1759.1 1842.7 1909.1 1983.18.7 18.7 22.2 23.1 30.57.4 23.6 24.6 21.7 20.46.2 23.9 32.0 30.7 37.95.0 30.1 34.5 29.3 27.9

Table 5. Macroeconomic indicators for selected countries

1984 1985 1986 1987 1988 1989

Algeria

Inflation, consumer prices 8.1 10.5 12.4 7.4 5.9 9.3Current account balance 0.1 1.8 �3.5 0.2 �3.5 �1.9Total debt service 36.8 35.6 56.4 76.6 76.6 66.8Gross capital formation 35.2 34.6 33.6 27.6 27.6 30.1GDP per capita growth 2.2 0.5 �2.4 �3.5 �3.6 1.8

1971 1972 1973 1974 1975 1976

Egypt

Inflation, consumer prices 3.1 2.1 5.1 10.0 9.7 10.3Current account balance �5.7 �5.3 �5.8 �17.7 �21.2 �10.2Total debt service 21.8 32.5 31.1 11.9 10.3 6.4Gross capital formation 13.2 12.3 13.1 22.5 33.4 28.4GDP per capita growth 1.6 0.2 �1.1 0.5 6.8 12.2

1982 1983 1984 1985 1986 1987

Egypt

Inflation, consumer prices 14.8 16.1 17.0 12.1 23.9 19.7Current account balance �9.9 �5.4 �8.2 �9.3 �9.4 �2.3Total debt service 19.3 20.1 21.4 25.8 27.0 17.9Gross capital formation 30.1 28.7 27.5 26.7 23.7 26.1GDP per capita growth 7.1 4.6 3.4 3.9 0.1 0.0

1984 1985 1986 1987 1988 1989

Jordan

Inflation, consumer prices 3.8 3.0 0.0 �0.2 6.6 25.7Current account balance �5.3 �4.9 �0.7 �5.4 �4.6 4.4Central government debt 49.6 56.4 59.0 70.1 100.1 126.1Total debt service 13.0 17.2 19.7 24.0 30.9 19.7Gross capital formation 28.8 20.5 20.5 23.3 23.5 23.7GDP per capita growth 4.6 �0.2 3.1 �0.8 �5.2 �16.5

1978 1979 1980 1981 1982 1983

Morocco

Inflation, consumer prices 9.7 8.3 9.4 12.5 10.5 6.2Current account balance �9.9 �9.4 �7.5 �12.0 �12.1 �6.4Central government debt 38.2 39.8 41.7 53.4 58.4 73.2Total debt service 22.9 26.6 33.4 38.4 45.4 40.3Gross capital formation 25.4 24.5 24.2 26.1 28.2 24.0GDP per capita growth 0.0 2.5 1.3 �4.9 7.2 �2.7

1981 1982 1983 1984 1985 1986

Tunisia

Inflation, consumer prices N/A N/A N/A 8.9 7.3 6.2Current account balance �5.4 �8.1 �6.8 �9.3 �6.9 �6.7Central government debt 35.1 38.2 41.5 42.3 45.5 56.5Total debt service 15.2 16.2 19.3 22.7 25.0 28.4Gross capital formation 32.3 31.7 33.5 35.9 30.2 26.6GDP per capita growth 2.8 �3.1 2.0 3.7 2.5 �4.5

Data source: WDI 2002.Note: Inflation, consumer prices (annual %); Current account balance (% of GDP); Central government debt (% ofGDP); Total debt service (% of exports of goods and service); Gross capital formation (% of GDP); GDP per capitagrowth (annual %).



received the first of its series of IMF loans andcomparing them with the previous period wecan see if there is evidence that the grantingof the first loan coincided with severe macro-economic distress.

(i) AlgeriaAlgeria received its first IMF Standby Loan

in 1989. Comparing 1989 with the previous fiveyears it does not seem that Algeria was inexceptional distress in terms of inflation, thecurrent account balance, gross capital forma-tion, or GDP growth. Indeed, GDP growthhad bounced back after three years of negativegrowth. The only variable that shows any signof significant deterioration is the debt serviceratio, which had doubled compared to 1984.However, in the year in which Algeria wasgranted its first IMF program the debt serviceratio was already beginning to improve. Thiscursory glance at the type of macro economicvariables that the IMF usually considers whendeciding whether a country is in need of a loanseems to suggest that Algeria’s first IMF loanin 1989 cannot be explained by the standardanalysis of recipient need. As will be arguedbelow, it is possible that important changes inAlgeria’s domestic politics and foreign policyprovide an alternative explanation of the timingof the 1989 loan.

(ii) EgyptEgypt received its first IMF loan in 1976

followed by a new phase of loans that com-menced in 1987. The first 1976 loan does seemto coincide with a period of increased inflationand deterioration in the current account,although debt service, gross capital formation,and GDP growth were not problematic. How-ever, given that inflation and the current ac-count are often regarded as critical indicatorsby the IMF, it would seem that the 1976 loanreflects a degree of recipient need. The picture,however, is very different for the 1987 loan.There is no indication of macroeconomicinstability in 1987. Indeed, all variables apartfrom the GDP growth rate were improvingin 1987. Although some of the improvementmay be due to the IMF program itself, giventhat the loan was signed in May, the previousyear’s data, apart from inflation, do not sug-gest any increase in macroeconomic distressimmediately prior to the loan. Again, it seemswe must look for other factors, which go be-yond recipient need to help explain the 1987IMF loan.

(iii) JordanJordan’s IMF programs began in 1989. In

this case there is much more evidence of macro-economic crisis. Inflation, the build up of cen-tral government debt and GDP growth allregistered a significant deterioration in 1989.The previous two years had also seen escalationin the debt service ratio. However, as we willargue below, the 1989 economic crisis in Jordanwas inextricably linked with changes in domes-tic politics and foreign policy, both of whichmay also have played a part in loan timing.

(iv) MoroccoMorocco commenced her IMF programs in

1983 but on the macroeconomic data containedin Table 5 would not seem to be an obviouscandidate for such programs. Inflation, the cur-rent account and debt service ratio all improvedin 1983 and this cannot be ascribed to the IMFprogram itself, as the loan was not signed untilSeptember. The only variable that worseneddramatically in 1983 was the GDP growth rate,which is not a variable IMF program tradition-ally responded to in the 1980s.

(v) TunisiaTunisia became an IMF Standby recipient in

1986. As with Morocco, there is little sign of re-cipient need in terms of the standard variablesof concern to the IMF. Inflation, the currentaccount, and debt indicators were showing nonotable deterioration. Again, evidence of needis limited, and it seems, as with Morocco andEgypt in the 1980s, the only obvious indicatorof need contained in Table 5 is the decline inGDP per capita growth.In summary, the 1980s, when most IMF pro-

grams in the MENA region commenced, was aperiod of generally deteriorating economic per-formance for the region as a whole. There was aclear need for both external finance to help withgrowing debt burdens as well as a program ofeconomic reform to restructure many of theeconomies in the region and generate sustain-able economic growth. Hence, the stage wasset for entry of the IMF and World Bank withtheir stabilization and structural adjustmentloans and we cannot deny the element of recipi-ent need in this respect. However, an analysis ofthe exact timing of the first IMF loan in the fivemajor recipients provides only limited evidencethat economic need, as illustrated by keymacroeconomic variables, was a determinant.Although it seems that Egypt’s loan of 1976and Jordan’s loan of 1989 were a response to


clear macroeconomic difficulties, the evidencefor Algeria, Morocco, and Tunisia as well asEgypt’s 1987 loan is much less clear cut. In-deed, if anything, in Morocco, Egypt, andTunisia it seems that the IMF was respondingto the growth rate variable, which was notone of the standard macroeconomic variablesone usually associates with IMF programs inthe 1980s. 17 This would suggest that in manyinstances other factors might well have influ-enced the decision as to when a country is eligi-ble to commence a series of IMF and WorldBank programs. It is to this issue that we nowturn.

(b) Donor interest and the influenceof the United States

We have argued in an earlier section thatthere are reasons to suspect that the politicalinterests of their major shareholders, particu-larly the United States, may well influence theflow of funds from the IMF and World Bank.Hence, in this section, we present a qualitativeanalysis of the timing of the signing of WorldBank and IMF program loans in the majorMENA recipients in order to see whether thereis any evidence that political factors have beeninfluential. The comparison of the timing ofeach loan with key domestic and internationalpolitical events is a simple form of qualitativeanalysis based on country case studies. Hence,it is both descriptive and speculative and assuch it cannot prove any causal link. We ad-dress this problem in Section 4 where we em-ploy more formal quantitative analysis.

(i) JordanWhile the 1989 agreement with the IMF and

World Bank largely reflected dire domestic eco-nomic conditions, Jordan’s experience withboth bilateral and multilateral aid, before andafter 1989, presents an excellent example ofthe subjection of such flows to the politicalinterests of and pressure from major westerndonors. Following the 1973 Arab–Israeli Warand an Arab oil embargo against the UnitedStates, Washington increased pressure on thelate King Hussein to sign an individual peacetreaty with Israel. But with more than half ofhis population being of Palestinian origin, andwithout tacit support, if not direct participa-tion, of the PLO, a separate peace treaty withIsrael would have been tantamount to politicalsuicide. The US ‘‘frustration with King Hus-

sein’’ led to the suspension of American aid toJordan in 1978 (Shultz, 1993, p. 454). Duringthis period, Jordan was neither favored by theIMF nor the World Bank.However, in the second half of the 1980s ten-

sion between Hussein and Arafat led to a rifton who should represent the Palestinians inany peace settlement with Israel. Husseinconsequently reverted to an old and by nowwell-documented approach of continuing secretcooperation with the Israeli state (Dallas, 1999)and in 1988 severed all economic and adminis-trative ties with the West Bank and Gaza Strip(WBG). With Jordan’s Palestinians dominatingthe private sector, severing ties with the WBGcreated uncertainty with regard to their politi-cal future and presence in Jordan. They thusengaged in extensive capital flight and curtailedtheir investment and economic activities in thecountry, contributing to Jordan’s first realbanking and financial crisis in 1989. Withinthe space of six months, the Jordanian Dinarlost almost 50% of its nominal value, Jordan’sexternal debt reached unsustainable levelsand per capita income was almost halved(Kanovsky, 1989). Subsequent riots resultedin a tactical move by King Hussein; he restoredparliamentary elections in 1989 (suspendedsince the 1976 war), and within a span of twoyears abolished martial law, legalized politicalparties, and sanctioned greater freedom ofpress. In July 1989, the IMF granted Jordanan SDR 60 million Standby Agreement andthe World Bank provided a US$160 millionSECAL for the industrial and trade sectors.In view of the above we can speculate that thecomplex interaction of the stance toward Israeland the WBG, the ensuing economic crisis, andthe attempted panacea in the form of politicalliberalization, all played a role in qualifyingJordan for IMF and World Bank support in1989.With strong domestic opposition to foreign

intervention in the region, Jordan took a neu-tral stand in the 1990–91 Gulf war and refusedto openly support US-led attacks against Iraq.This led to a complete halting of aid flows toJordan from the United States and its Arab al-lies in the Gulf and the temporary suspensionof the IMF and World Bank agreements withtwo-thirds of the IMF Standby funds not beingdrawn. Squeezed financially, isolated interna-tionally, and ostracized regionally the US-ledpressure had the desired effect of promptingan ‘‘alliance shift’’ (Brand, 1994, p. 20). Hus-


sein soon criticized Saddam Hussein andopenly talked about regime change in Iraq,and went further by hosting Iraqi oppositionleaders. Following this revised stance, a newSDR 44.4 million Stand-By Agreement wassigned with the IMF in February 1992.Since the mid-1990s, Jordan has been further

rewarded for her peace overtures to Israel. In1993, Jordan started direct negotiations with Is-rael under the Oslo Accord of 1991–92, leadingin October to a Common Agenda agreement onissues related to territory and water, refugees,and arms control. In the same month, theWorldBank granted Jordan a second US$80 millionSECAL for the energy sector. This was followedin May 1994 by a SDR 130 million IMF Ex-tended Fund Facility. In 1994, Jordan formallysigned a peace treaty with Israel in Wadi Araba,formally ending the 46-year-old state of war be-tween the two countries. Since then, Jordan hasnot only become one of the largest recipients ofUS aid in the world, but also the recipient of afurther six World Bank loans and three IMFloans. The political timing ofmany of these loansdeserves attention. The thirdWorld BankUS$80million SECAL for the agricultural sector cameonly three months after the Wadi Araba Agree-ment was signed. In the same year, the UnitedStates wrote off US$833 million of Jordan’s debtand began providing Jordan with advancedweaponry. The fourth SECAL for the sameamount came in October 1995, less than amonthafter Jordan’s support for the OSLO II Accordunderwhich Israel agreed to a partial withdrawalfrom the West Bank with administrative powersto be given to the Palestinian Authority.In light of the above analysis, it would seem

that the timing of at least seven of the 14 pro-gram loans that Jordan has received from theBank and the Fund since 1989 may well havebeen influenced by Jordan’s stance on MiddleEast affairs.

(ii) AlgeriaThroughout the 1970s and 1980s, Algeria was

considered a rogue state by the non-communistwest. In 1974 the country firmly rejected the‘‘open door policy’’ (infitah) adopted by Egyptand much of the Arab world which involved ashift in the balance of domestic economicpower to the private sector, opening up to wes-tern investment and accepting the hegemony ofthe United States. Instead, in 1976 PresidentBoumedienne adopted a new socialist constitu-tion and Islamic state, with the Islamist move-

ment further promoted by his successorPresident Chadli Benjedid. During this period,which witnessed rapid industrialization andthe successful development of domestic oiland gas, the west remained hostile to Algeria’santi-American regime (Pfeifer, 1996; Swearin-gen, 1996). In 1980, Algeria embarked on asuccessful liberalization program designed toovercome the inefficiencies created by the previ-ous import substituting industrialization strat-egy. Although the program was not dissimilarto a standard IMF and World Bank package,Algeria, unlike its Western friendly neighborsMorocco and Tunisia, received little assistance.While Morocco and Tunisia had more than athird of their annual external debt on conces-sional terms, Algeria was forced to finance itsreform program in the early and mid-1980swith market-based loans with only 3% of herdebt on concessional terms.But reforms, associated with external bor-

rowing on unfavorable terms, induced a bal-looning in foreign debt. The collapse of oiland gas prices in 1986 followed by a largedevaluation in 1988 contributed to escalatinginflation and unemployment, which triggeredstrikes, riots, and growing domestic oppositionto the regime. The economic and political cri-sis prompted a shift in both domestic andinternational policy. In 1988, a new constitu-tion restricting the military and allowingopposition parties was introduced. In early1989, Algeria joined the new Arab MaghrebUnion (UMA) which was committed to pre-venting the spread of radical Islam and foster-ing closer links between the Maghreb and theEuropean Union. As in Jordan, Algeria’s newpro-western stance combined with domesticpolitical liberalization was a signal for thearrival of the World Bank and IMF. In May1989, the Fund granted a SDR 156 millionStandby followed in August 1989 by aUS$300 million Structural Adjustment Loanfrom the World Bank. In total, during 1989–99 Algeria received four IMF stabilizationloans (as well as a Compensatory FinancingFacility) and four World Bank adjustmentloans. The timing of several of these, as ar-gued below, is noteworthy.President Benjedid’s political liberalization

backfired producing unexpected support forthe Islamic opposition. The Bank and the Fundresponded in June 1991 by offering the embat-tled regime a SDR 300 million Standby and aUS$350 million Structural Adjustment Loan


(the largest ever Bank program loan to anycountry in the MENA region). The army, sha-ken by the Islamists’ overwhelming victory inlocal elections, stepped in on January 1992,deposed Bendjedid, and cancelled the elections.A retired army general, Liamine Zeroul, wasappointed by the army as head of the state inJanuary 1994, and confirmed President afterthe 1995 elections. The disposition of Benjedidand cancellation of elections ushered in a newand bloody era in Algeria’s history, character-ized by two main features. First, brutal repres-sion of Islamists and other opposition in whatbecame known as the ‘‘dirty war.’’ Second,the growth of a radical opposition Islamicmovement coincided with the new post-ColdWar American view of Islam as the new com-munism. Therefore, the new Algerian military-backed regime used the so-called war on terrorto build closer relations and links with the Uni-ted States. In March 1993, Algeria broke offdiplomatic relations with Iran, after years ofcultivating links with Tehran, which it nowblamed for exporting Islamic revolutions tothe Arab World. At the same time, Algeriawithdrew its ambassador from Sudan,another country described by US officials as arogue and anti-western state (EIU, 2001, pp.14–15). In May 1994, Algeria was rewardedwith a SDR 457 million IMF Stand-ByArrangement.In late 1994, Algeria also vigorously sup-

ported, along with Egypt, an anti-terrorist codeof conduct at the Casablanca Islamic Summit.A few months later in January 1995, Algeriareceived a US$150 million World Bank Eco-nomic Rehabilitation Support Loan, followedin May 1995 by the largest ever IMF ExtendedFund Facility in the region of SDR 1.2 billion.The latter came to an end in May 1998, and wasassociated with unexpected macroeconomicsuccess, mostly caused by improved globalprices and demand for gas.In late 1998, President Zeroual announced

that he would stand down and that electionswould be brought forward to early 1999. Oneday before the voting began in the delayedApril 1999 elections, all candidates, exceptMr. Abdelaziz Bouteflika, pulled out due tocredibility problems. This left Mr. Bouteflika,who was supported by a pro-western, anti-Islamist powerful coterie of senior army officersand state officials, to become Algeria’s seventhPresident. His first years of rule were markedby intensified violence and further crackdownon Islamists. To help bolster Mr. Bouteflika’s

position, particularly following the decline inoil prices in late 1998 and the rise of debt ser-vice ratio to 46%, the IMF extended a Compen-satory and Contingency Financing Facility ofSDR 223.5 million in May 1999.

(iii) EgyptEgypt was the first Arab state to sign a peace

treaty with Israel in 1978, hence formally end-ing the state of war between the two countries.Since then, Egypt has become a favored recipi-ent of US aid, despite its appalling humanrights record. Although Egypt has had rela-tively few Bank and Fund program loans, 18

what is noteworthy is that Egypt has continuedto receive such loans despite the disappointingpace and quality of reform which falls signifi-cantly short of that in Morocco and Tunisiaas well as in Jordan after 1999. In addition,the timing of two of these loans was undoubt-edly influenced by political factors. WithoutEgypt, there would have been no Arab stancesupportive of the US-led coalition in the1990–91 war against Iraq. Egypt mobilizedArab support for the war and held an emer-gency Arab Summit for that purpose in 1990.Unlike Jordan, Egypt also sent troops to fightalongside the American forces in liberating Ku-wait. Three months after the war ended in May1991, Egypt was rewarded with a SDR 234 mil-lion IMF Standby Loan and a US$300 millionBank Structural Adjustment Loan. Egypt alsoreceived more than $15 billion of debt write-off from the west for its efforts and strong sup-port for the allies during the 1990–91 war, thehighest level of debt forgiveness in the historyof MENA.

(iv) Tunisia and MoroccoTunisia, like Morocco, has long been re-

garded as a friendly pro-Western regime withinMENA. Consequently, both Morocco andTunisia have been treated favorably first bythe European Union, and later, by the IMFand World Bank. During 1982–2003, Moroccohad six debt rescheduling agreements with theParis Club and three with private internationalbanks, received 15 World Bank Structural andSectoral Adjustment Loans and seven Stand-byand Extended Facilities from the IMF.Morocco has also long been the recipient ofgenerous American military support. Over thesame period, Tunisia received nine World BankStructural and Sectoral Adjustment loans inaddition to five years of continuous IMF finan-cial support. Such treatment compares very


favorably with their less America-friendlyneighbors such as Algeria in the 1980s andLibya in the 1980s and 1990s.It has been argued that Washington not only

used its influence inside international financialinstitutions to soften IMF and World Bankconditionality in Morocco and Tunisia as wellas the WTO’s entry requirements, but also,along with the European Union and Japan,‘‘repeatedly and generously lubricated’’ theirreform efforts by ‘‘financial assistance to easethe pain and political costs to the regime ofearly austerity phases’’ (Pfeifer, 1999, pp. 23& 25–26). It would seem that in these MENAcountries US officials hoped, by providingfriendly regimes with financial and militarysupport and by developing them into regionalshowpieces of globalization, that this wouldstabilize the regimes of their Arab allies (Alex-ander, 1996; Waterbury, 1998).Morocco has been such a massive and con-

tinuous recipient of Bank and Fund programloans that it is difficult to link key domesticand international political events to the timingof such loans. However, its efforts in supportingthe 1991 Gulf war, including sending 1,200 ofher troops, was rewarded handsomely withmore than US$5 billion in debt forgivenessfrom the United States and Arab oil-rich states.Political liberalization in late 1997 and early1998, with parliamentary elections resulting inthe first change over of political power in theKingdom’s history, was followed by threeWorld Bank loans over the next year totalingUS$450 million.Tunisia has long pursued a pragmatic pro-

Western foreign policy (Murphy, 2002). How-ever, a significant shift in domestic politicsoccurred in late 1987 with the coming to powerof General Zine Ben Ali. Ben Ali’s new regime,claiming an attempted Islamic coup, rapidlycracked down on the Islamic movement, arrest-ing the head of the main opposition the Move-ment de la Tendance Islamique. In early 1989,the regime signaled a further shift against Isla-mic politics in favor of a pro-western stance byjoining the Arab Maghreb Union designed toprevent the spread of radical Islam and fostercloser links with the European Union. The re-sponse of the Washington-based internationalfinancial institutions mirrors that in Algeria.The crackdown on Islam was followed by botha Bank and Fund loan the following year, whilejoining the Maghreb Union was followedwithin four months by two further Bank Sec-toral Adjustment Loans.

4. LOAN TIMING: A MORE FORMALQUANTITATIVE ANALYSIS

The above has provided a simple descriptiveanalysis of the timing of IMF and World Bankprogram loans to MENA recipients in order totry and isolate the influence of both economicneed and key domestic and international polit-ical events. The results suggest that although insome cases recipient economic need, as signaledby a deterioration in key macroeconomic vari-ables, is a determinant of the start of IMF(and usually World Bank) lending, such loansalso seem to be influenced by political eventsin the recipients that curry favor with US policyin the region. However, so far the analysis lacksrigor. In order to strengthen and advance ourargument we now use a more formal quantita-tive approach.In this section, we employ a Probit model to

investigate what factors influence IMF loans toMENA countries. 19 The dependent variable isIMF coded either one, if a country signed anagreement (including SAF, ESAF, and PRGF)in year t, or zero, otherwise. Based on the pre-vious discussion, three sets of variables areincluded in the regression: economic need, USinfluence, and domestic political factors. Inde-pendent variables representing economic needconsist of GDP per capita, GDP growth rate,debt service ratio, short-term debt as percent-age of total debt, balance of payments, andchanges in national reserves. 20 US influence iscaptured by the dummy variable PEACE,which indicates whether a country signed apeace treaty with Israel or not. Two variablesare used to capture the domestic political fac-tors: DEM—the democracy index (rangingfrom one to seven, one is the highest level ofdemocracy, seven the lowest), and DELEC—the legislative election year. 21

The estimation is conducted in a pooled sam-ple with 11 countries from 1975 to 2000. The de-tailed variable definition, data sources, and thecountry list are reported in Appendix C. Fol-lowing the standard procedure, the sample waslimited to years in which a country was notunder a previously agreed IMF program, andthe explanatory variables—LGDPPC, GDPG,CAB, TDEBTS, and SDEBT are lagged byone year to avoid simultaneity bias. 22

The first stage of the analysis focuses only onthe economic variables, which may reflect thepresence of a financial or macroeconomic prob-lem that might prompt the government to ap-proach the IMF for resources and which may


be used by the IMF to decide on loan eligibility.The results for the corresponding economicmodel are reported in Table 6. The coefficientson GDP per capita (LGDPPC[�1]), total debtservice ratio (TDEBTS[�1]), and changes innet reserves (D(RES/GDP)) are all significantand with the expected signs indicating that aMENA country with low GDP per capita, highdebt service ratio, and experiencing a sharp de-cline in reserves is likely to receive IMF assis-tance. The coefficient on the current accountbalance (CAB[�1]) is not significant and thecoefficients on GDP growth (GDPG[�1]) andshort-term debt (SDEBT[�1]) are significantbut with unexpected signs.

Table 6. Probit analysis of the de

Variables Economic m

LGDPPC[�1] �0.659(0.279)*

GDPG[�1] 0.058(0.033)

CAB[�1] �0.015(0.023)

TDEBTS[�1] 0.039(0.013)*

SDEBT[�1] �0.060(0.023)*

D(RES/GDP) �5.731(3.46)*

DEM[�1]

DELEC[�1]

DELEC

DELEC[+1]

PEACE

Constant 3.197(1.941)

Number of observation 165Predicted

0Actual 0 139

1 18

Log likelihood �48.8**

Chi squared 35.6***

R2 (ML) 0.194a Standard errors are in parentheses.***, **, * and � indicate that the coefficient is significantlyrespectively.

Although the above economic approachprovides some useful insights into the deter-minants of IMF programs, it suffers fromspecification error due to the omission of rele-vant variables (e.g., variables that capture polit-ical factors which we know from our aboveliterature review and qualitative analysis arelikely to be important). Consequently the eco-nomic model has low explanatory power (mea-sured by an R2 of 0.194) and a low correctprediction ratio. As can be seen from thebottom of Table 6 the number of observa-tions for receipt of a loan is 23 but oureconomic model only correctly predicts 5 ofthese. 23

terminants of IMF agreementsa

odel Supplemented model

�0.337* (0.445)

0.042* (0.045)

0.038(0.037)0.033

** (0.017)**

�0.065** (0.037)*

�6.972(4.426)�

�1.008(0.445)*

0.6000(0.535)0.694(0.601)0.832

(0.505)*

1.052(0.481)**

5.804* (3.089)**

154Predicted

1 0 13 125 65 7 16

* �31.4***

67.1***

0.353

different from zero at the 1%, 5%, 10% and 15% levels,


In order to try and improve the model, thetwo sets of political variables are added intoregression. The results for the correspondingsupplemented model are reported in the thirdcolumn of Table 6. The coefficients on GDPper capita (LGDPPC[�1]) 24 and GDP growth(GDPG[�1]) are no longer significant; thecoefficient on the current account balance(CAB[�1]) is still insignificant; and the coeffi-cients on short-term debt (SDEBT[�1]) 25 andthe change in reserves (D(RES/GDP)) changedslightly in terms of magnitude. For those polit-ical variables, the dummy PEACE is positiveand significant, the democracy index DEM isnegative and significant, and the dummy forthe year after a legislative election DELEC[+1]is positive and significant. This suggests thatMENA countries which have signed a peacetreaty with Israel, that have just had a legisla-tive election and are democratic are likely to re-ceive an IMF loan. In the supplemented model,the R2 statistic is much higher at 0.353 and thecorrect prediction ratio of the positive value(sign an agreement with IMF) is 16 out of 23,which is a big improvement, compared to theeconomic model. 26

The above results clearly show that in tryingto predict when the IMF will sign a loan agree-ment with a MENA country the model whichincorporates both political and economic vari-ables is superior to a purely economic model.Our supplemented model indicates thatwhether a country receives an IMF programis influenced by both economic and politicalfactors, particularly the latter. The only eco-nomic variables in the supplemented modelthat have the predicted sign and are significantare the change in foreign reserves and total debtservice—a decline in reserves or a high debt ser-vice ratio are good predictors of an IMF pro-gram. This finding can be further supportedby the fact that 20 out of 28 IMF programsin MENA were accompanied by a Paris Clubdebt relief or reschedule agreement. Hence,IMF programs seem to clearly coincide withdebt problems and the need to save foreign re-serves. Along with the two economic variables,signing a peace treaty with Israel and improv-ing democracy also increase the likelihood ofreaching an agreement with the IMF. The exis-tence of political business cycle also plays itspart in that we have found that MENA govern-ments are more likely to enter into an agree-ment with the IMF in the year after thelegislative election.

The formal results from the Probit modellend support to our more qualitative analysis.They show that economic need alone does notreally explain the timing of IMF loans. How-ever, political liberalization, which often seesthe incumbent regimes challenged by Islamicopposition, seems to have an influence asshown by the significance of the democracyand election variables. Likewise, a change inforeign policy stance represented by signing apeace treaty with Israel is a good predictor ofIMF loans.

5. LINKING THE QUALITATIVE ANDQUANTITATIVE ANALYSIS

The quantitative Probit model work can beused to identify outliers (those cases that themodel fails to predict) which, in an iterative re-search process, could then be re-examined byway of the qualitative case study type work 27

Cases were no program was introduced despitethe model’s prediction of a loan are: Egypt1985, 1986, 1989; and Morocco 1979, 1994,1997. Cases were a program was put in placebut not predicted by the model are: Egypt1976, 1987, 1996; Jordan 1989, 1992; Tunisia1986 and Yemen 1997. It is beyond the scopeof this paper to return to extensive qualitativecountry case study analysis to explain theseoutliers. However, two general factors helpaccount for the outliers. Firstly, the problemof missing variables, that is, our model mayexclude some important variables which aredetermining factors in individual cases. Forexample, the cases of IMF loans to Egypt in1976 and Jordan in 1989 may be partially ex-plained by the use of such loans by the UnitedStates and other Western powers to help facili-tate the peace process between these two coun-tries and Israel prior to the signing of a PeaceTreaty. Jordan’s 1992 loan was largely theproduct of the Gulf War and Yemen in 1997is a special case since its IMF loan took theform of an ESAF accompanied by a PRGF.A second general explanation of outliers

might be the fact that in some instances thesigning of an IMF agreement might be delayeddespite the existence of predictive events. A typ-ical example might be Egypt where our modelincorrectly predicted agreements in 1985 and1986 yet failed to predict the 1987 agreement.The fact that Egypt’s IMF program in themid-1970s sparked serious food riots, as well


as Egypt’s access to generous assistance fromthe United States in times of economic need,might explain why there was a delay in reachinganother IMF agreement. As can be seen fromTable 5, economic conditions started to deteri-orate in 1982 and yet an agreement with theIMF was not reached until 1987. This was de-spite the fact that in 1985 and 1986, when themodel incorrectly predicted an agreement,overall economic conditions were much poorerthan the following year when the agreementwas eventually signed.

6. CONCLUSION

Our qualitative and quantitative analysis en-ables us to conclude that both recipient needand donor interest influence the granting ofIMF and World Bank program loans to coun-tries of the MENA region. This is not surpris-ing given that our literature review indicatedthat most empirical studies of aid allocationfind that donor interest, including geo-politicalinterest, influences who gets what in terms ofaid. The generally accepted view is that donorinterest plays a more important role in bilateralaid allocation than in multilateral aid alloca-tion. This may be so, but we have identifiedimportant reasons why the major westernshareholders might be able to influence the flowof funds from the two major Washington-basedmultilaterals. Given its voting power in boththe Bank and the Fund, the United States isin a particularly influential position.Our qualitative analysis focused on the five

major MENA recipients of IMF and WorldBank program loans—Algeria, Jordan,Morocco, Tunisia, and Egypt. Looking at eachcountry’s macroeconomic performance in theyear in which they commenced their first phaseof program loans, we see very little evidence ofeconomic need. Only in the case of Jordan inthe late 1980s and Egypt in its first phase ofloans during the mid-1970s do we see any clearsign of recipient economic need in terms of asignificant deterioration in the macroeconomicindicators that the IMF is usually concernedwith. It seems therefore that we must look toother factors to explain the IMF and WorldBank engagement with Egypt in the 1980sand with Morocco, Tunisia, and Algeria. Inall cases a cursory political analysis would indi-cate that a shift toward a pro-western foreignpolicy, peace overtures to Israel, domesticpolitical liberalization, and the often related

challenge to the regime by Islamic oppositionprompt an inflow of funds not just from theUnited States but also from the Bank andFund. Even in the case of Jordan, whichbecame a recipient of such loans in 1989, thesevere economic crisis of that year was inextri-cably linked with such foreign policy anddomestic political events.The above findings are further supported by

our more formal quantitative analysis. Usinga Probit model to estimate the determinantsof IMF lending in the region we found that amodel that only includes variables representingrecipient need performs very poorly. However,once we include foreign policy and politicalvariables the model performs extremely well.In this supplemented model the only economicvariables that help to predict whether a MENAcountry will be granted an IMF loan are achange in foreign reserves and total debt ser-vice—a decline in reserves or a high debt serviceratio are good predictors of an IMF program.Signing a peace treaty with Israel improves acountry’s chance of a loan as does improvingdemocracy. Related to the latter, we also foundthat holding an election is likely to be followedby an IMF loan in the post-election year.The above findings are important, not just

because they add to an already large body ofempirical work on the determinants of aid allo-cation, but also because they have importantpolicy implications. The fact that IMF andWorld Bank lending in MENA seems to be ori-entated toward pro-western regimes that intro-duce western-style democracy, and adhere toUS foreign policy interests in the region suggestthat factors other than recipient need are influ-encing global aid allocations. This has twoimportant implications, which go beyond thescope of this paper. Firstly, it may well reducethe developmental impact of a scare resource,namely aid. Low income countries or those thatcan use aid to the best effect, may not receive asmuch aid as wealthier countries or countrieswith weak policies, where aid has been shownto be less effective (Burnside & Dollar, 2000).As Collier and Dollar (2002) have argued, amore poverty-efficient allocation of aid hasthe potential to double the number of peoplelifted out of poverty from 10 to 20 million.Secondly, the politically motivated flow of

funds to MENA may well trigger adverse socialand political effects. Program loans from theIMF and World Bank have economic liberal-ization conditions attached to them. Such re-form conditions, although they often have the


potential to bring significant economic gains,may well have negative social ramifications inthe recipients unless adequate social safety netsare in place. For example, reforms such as priv-atization, removal of state subsidies on food-stuffs, devaluation, and trade liberalizationcan potentially increase unemployment and in-come inequality as well as reduce real incomesof the poor. This, in turn, may lead to thegrowth of anti-reform movements challengingincumbent regimes. There is already ampleanecdotal evidence that this has occurred. The1990s and the first four years of the 21st cen-tury have witnessed a rise in the number andforms of distributive conflicts in the ArabWorld, including riots, demonstrations, strikes,violence, assassinations, clashes with labor un-ions and university students in addition to anincrease in crime rate (Ayubi, 1995; Economist,September 5, 2002; El-Ghonemy, 1998; Rich-ards & Waterbury, 1996; Shafiq, 1998). Quiteoften this unrest has an explicitly anti-western,anti-globalization, and anti-IMF focus. Insome instances, such as in the riots in Jordanin April 1989 and August 1996 prompted bythe IMF-induced lifting of price supports, theIMF and the World Bank were viewed by manyof the opponents of reform as synonymouswith the American presence and interests inthe region. If this persists, the very regimes thatAmerica and the west are trying to supportwith funding and reform packages may wellnot survive.Many such opposition movements have cen-

tered on Islamic-based political parties. Politi-cal Islam and Islamic fundamentalism shouldnot be confused. But a vicious cycle of decliningsocial welfare caused by possible effects ofeconomic liberalization, increased domesticopposition to pro-western local regimes imple-menting such programs, and repression of suchopposition by the same regimes is likely to force

frustrated religiously based political groupsinto increasingly extremist responses as well asenhancing their appeal to impoverished anddisaffected members of society.The general influence of the IMF and World

Bank in MENA, the welfare effects of IMF andWorld Bank programs in MENA countries aswell as local perceptions of these two institu-tions is hence an area that deserves further re-search. It may well be that in view of the factthat IMF and World Bank funds have ac-counted for only a small percentage of the flowof funds into the region (see Table 1) that theirinfluence over policy and liberalization is corre-spondingly small. In addition, the very fact thatthe flow of funds is politically motivated maymean that the conditions attached to thesefunds are weak. Indeed, both these factors helpexplain why the Bank and the Fund have hadrelatively little influence over the economic re-form process in Egypt in comparison with theinfluence of USAID. On a more general globallevel, there is a growing body of literature thatsuggests that IMF conditionality is not effectivein obtaining intended reform outcomes (Dollar& Svensson, 2000; Goldstein, 2003; Mercer-Blackman & Unigovskaya, 2004; Mussa &Savastano, 2000) and that it does not have acatalytic effect in terms of access to the interna-tional capital market (Bird, 1996; Bird & Row-lands, 1997, 2000, 2001, 2002; Rodrik, 1995;Rowlands, 1996). If the latter is the case, theneffectiveness of granting or withholding IMFand Bank loans in terms of the carrot and stickeffect is severely weakened. On the other hand,there are many, including many groups andindividuals in MENA countries, who continueto believe that the IMF and World Bank wieldconsiderable influence over their economies.The basis, origin, and accuracy of such beliefsin the specific context of MENA countries areessential area of future research.

NOTES

1. The World Bank definition of MENA includes:Algeria, Djibouti, Egypt, Iran, Iraq, Jordan, Lebanon,Libya, Malta, Morocco, Oman, Palestine, Saudi Ara-bia, Syria, Tunisia, and Yemen. It does not include thehigh-income countries of the Gulf, nor Israel andTurkey, nor Sudan and Mauritania which althoughpredominantly Arab countries face challenges moretypical of sub-Saharan Africa. In our general discus-sions of aid allocations we use the same countrygrouping as the Bank, although also include reference

to Israel, a major recipient of US aid. Although Turkeycould be considered part of MENA in political,cultural, and geographic terms, and is also one of themost important countries in the region in terms ofUSA foreign policy, she is not generally consideredpart of MENA in economic terms. This fact, alongwith her close ties with Europe and moves to join theEuropean Union in the near future, mean we havetaken the standard definition of MENA and excludedTurkey from our analysis.


2. For an excellent survey and methodological critiqueof this work, see McGillivray and White (1993).

3. Western aid to Mobuto’s Zaire or Marcos’s Philip-pines designed to bolster anti-communist pro-westernregimes are good examples.

4. Post 9/11 the United States has been increasinglyforthright in suggesting that the War on Terror and USsecurity are important reasons for foreign aid (www.usaid.gov/fani/overview, p. 2). Likewise, when PresidentG.W. Bush proposed the first significant increase in USdevelopment assistance in a decade, he offered thefollowing justification when speaking at the UnitedNations Financing for Development meeting in Mont-errey, Mexico in March 2002: ‘‘We fight poverty becausehope is an answer to terror.’’

5. In terms of destination of aid in the MENA region,DAC aid is quite concentrated. During 1961–2001 Egyptand Israel accounted for more than 60% of DAC aidinto the region, followed by Morocco 8.0%, Jordan5.3%, and Tunisia 4.5%.

6. Donor interest includes pursuit of commercialinterests via the promotion of donor trade or investmentopportunities by allocating aid to countries most likelyto absorb donor exports and investment. It also includesthe pursuit of political, diplomatic and strategic objec-tives in order to create an international environment,which favors the donor. According to Feeny andMcGillivray (2002, p. 3): ‘‘This can involve allocatingaid to countries which are in a strategic geographiclocation or which have particularly close diplomatic tieswith the donor. It can even involve rewarding countriesfor particular actions with increased aid or punishingothers with reduced or continually low or zero levels ofaid.’’

7. We ran our regressions for three dependent vari-ables—US aid, bilateral aid excluding the United States,and multilateral aid and found that donor interest hasthe strongest effect on the allocation of US aid, and alsothat non-US bilateral aid responded more to donorinterest than did multilateral aid.

8. The US–Egypt dummy parameter in Berthelemy andTichit’s study was particularly large in the 1980s sub-period following the Camp David Peace Accord withIsrael and the analysis suggested the privileged assistanceenjoyed by Egypt from the United States translated intoan aid bonus of US$49 per capita.

9. A surprising result is that Turkey has a negativefixed effect coefficient for US aid compared to thepositive coefficient for non-US aid, despite the geopo-

litical importance of Turkey to the United States. Thismight be explained by two factors. Firstly, the fixedeffect coefficients capture time invariant linkages. It maybe the case that compared to the United States otherDAC countries, especially the European donors whichare major donors to Turkey, have more stable relation-ships with Turkey which are stronger over time.Secondly, fixed effects coefficients are only a proxy fordonor–recipient long-term linkages and may not alwaysbe accurate in individual cases.

10. IBRD Articles of Agreement IV: operations, Sec-tion 10; IDA Articles of Agreement V: operations,Section 6.

11. A country’s voting power is not the same as itsvoting weight: its power is its ability to decide the issuewhen a vote is taken whereas its weight is just thenumber of votes it has the right to cast. Voting power iscalculated by analyzing all the voting outcomes that canoccur, and in each case investigating the ability of everymember to be decisive—that is to be the one memberwho can decide whether the vote leads to a decision ornot.

12. The bulk of the studies have concentrated on thedeterminants of IMF rather than World Bank lending.The reason seems to be twofold. Firstly, an IMFagreement is usually a pre-requisite for a Bank programloan and most IMF agreements are followed by such aloan. Hence, many of the determinants of an IMFagreement will also be determinants of the Bank’sactivities. Secondly, the aspects of recipient need thatthe IMF is meant to respond to, namely inflation,balance of payments and budget deficits are much easierto measure than the more medium term supply sidedeterminants of World Bank program loans (structuralimbalance, developmental indicators etc.). Hence, froma methodological standpoint it is much easier toconstruct the independent variables in an equationestimating the determinants of IMF loans than thosethat would need to enter such an equation for WorldBank program loans.

13. It should be noted that Bird and Rowlands reportin a footnote to their paper that they were unable toreplicate some of Thacker’s results.

14. Dreher and Vaubel (2004) also took loan condi-tionality as the dependent variable and found that thenumber of conditions per IMF loan was positivelyrelated to prior use of Fund credit relative to quota andto the number of World Bank adjustment loans, as wellas being positively related to world interest rates andrecipient monetary expansion and negatively related torecipient international reserves.

http://www.usaid.gov/fani/overview

http://www.usaid.gov/fani/overview


15. Reliance on oil took two forms, direct and indirect.Direct reliance refers to the oil export countries whichinclude OPEC countries in the region, along with Egyptand Yemen. Indirect reliance refers to those countries,especially Jordan, Egypt, and Yemen, who received largeremittances from the oil rich GCC countries.

16. We look at IMF loans rather than World Bankprogram loans because the former are almost always aprerequisite for the latter.

17. The traditional division of labor in the 1980s wasthat the IMF would take care of balance of paymentsproblems while medium term growth would be theconcern of the World Bank (Mosley, Harrigan, & Toye,1995, vol. 1, pp. 51–56).

18. Egypt received only three World Bank SALs (twoof which were in the 1970s) and four IMF programloans.

19. Again only IMF programs are considered here dueto the fact that the World Bank’s SALs or SECALs aregenerally preceded by an IMF program. The Probitmodel can be used to determine the eligibility ofreceiving aid as opposed to the amount received. Hence,the dependent variable takes the value of one or zerodepending on whether each country in the samplereceives aid in a give year or not. In our case we usethe Probit model to predict whether a country receivesan IMF loan.

20. In our descriptive case country analysis of theprevious section, we also used inflation and investmentrates as indicators of macroeconomic instability. Wehave excluded these variables from our Probit model dueto the problem of multicollinearity. For example, it isnot appropriate to use both the investment rate andGDP growth variable in such an analysis since theformer is an important determinant of the latter. We didhowever run the regression with an inflation variable(both consumer price index and GDP deflator) but bothwere insignificant and did not improve the performanceof the model. This is probably due to the problem ofcollinearity, with the correlation between inflation andtotal debt service being �0.606. The way to overcomethis problem is to drop one variable. We kept debtservice instead of inflation for two reasons. Firstly, theresult is slightly better when debt service is included.Second, MENA countries tend to consult the IMF fordebt problems in the first instance—20 out of 28 IMFprograms in MENA were accompanied by Paris Clubdebt relief or rescheduling agreements.

21. In the interest of parsimony we had to restrict thenumber of explanatory variables introduced in the

model. Much of the aid allocation literature uses avariety of variables to capture political influences, forexample, political instability variables, UN voting pat-terns. However, increasing the variables enhances theproblem of multicollinearity and reduced degrees offreedom due to both increased number of explanatoryvariables and an increased number of missing observa-tions. Exclusion of such variables does not imply thatthey are not important in predicting the likelihood of anIMF loan. However, it is not our purpose to identify alldetermining factors and their weights in the loangranting decision process. Rather, we aim to establishthat loan disbursement is not exclusively determined byrecipient economic need but is also influenced bypolitical factors.

22. The results using a one year lag are consistent withthe results when economic variables are not lagged. Theuse of two or three year lags is not suitable for this typeof study due to the nature of IMF lending. The IMF ismeant to provide assistance to a country only at theactual time when the general macroeconomic conditionis deteriorating; hence a longer lag structure wouldsignificantly reduce the explanatory power of the mac-roeconomic variables.

23. The actual number of observations for no IMFloan is 142 and of these the economic model correctlypredicts 139 (and incorrectly predicts 3). But it should benoted that most Probit model analysis scores highly onpredicting the zeros in the observations of the dependentvariable.

24. The coefficient on LGDPPC[�1] is sensitive towhether we include a constant or not. When the constantis excluded, the coefficient is negative and significant.

25. Short-term debt as percentage of total debt has anegative impact on receiving an IMF loan, which is notexpected. However, when we checked the debt structureof MENA countries, we find short-term debt has been indecline since the early 1980s, and more than two-thirdsof their long-term external debt is from official sources.If the government’s desire to sign an agreement withIMF is designed to help initiate the process of long-termdebt relief from major donors this would explain thenegative coefficient on short-term debt. The smaller thepercentage of short-term debt (and hence the largerthe percentage of official long-term debt) the more likelyit becomes that an IMF agreement will be signed.

26. Given this study is based on a relatively smallsample, those figures are very respectable. Bird andRowlands (2003) have shown that a common feature ofthis genre of research is its arguably low explanatorypower overall. Although the percentage of correctprediction was often 80% and 90%, it has to be recalled


that these numbers corresponded roughly to the per-centage of countries without agreements.

27. Some authors, for example, Bird and Rowlands(2002) have had to analyze the outliers in such a waybecause their addition of political variables into theregression did not improve the explanatory power of the

model or the rate of correct predictions. This is not thecase in our analysis. As shown in Table 6, the R2

increases from 0.194 to 0.353 and the rate of correctprediction of the signing of an agreement also increasessignificantly from 5 to 16 once political variables areintroduced. Nevertheless, it is interesting to identify anddiscuss outliers.

REFERENCES

Alesina, A., & Dollar, D. (2002). Who gives foreign aidto whom and why. Journal of Economic Growth, 5,33–64.

Alexander, C. (1996). State, labour and the new globaleconomy in Tunisia. In D. Vandewalle (Ed.), NorthAfrica: Development and reform in a changing globaleconomy (pp. 177–202). New York: St. Martin’sPress.

Ayubi, N. (1995). Over-stating the Arab State. London:I.B. Tauris.

Barro, R. J., & Lee J.-W. (2001). IMF programs: Who ischosen and what are the effects. Paper presented atSecond IMF Research Conference, Washington, DC,IMF.

Berthelmy, J., & Tichit, A. (2002). Bilateral donor’s aidallocation decision. World Institute for DevelopmentEconomic Research (WIDER) Discussion Paper,No. 2002/123, United Nations University.

Bird, G. (1996). The International Monetary Fund anddeveloping countries: A review of the evidence andpolicy options. International Organization, 50(3),477–511.

Bird, G., & Rowlands, D. (1997). The catalytic effect oflending by the international financial institutions.The World Economy, 20(7), 967–991.

Bird, G., & Rowlands, D. (2000). The catalyzing role ofpolicy-based lending by the IMF and World Bank:Fact or fictions? Journal of International Develop-ment, 12(7), 951–973.

Bird, G., & Rowlands, D. (2001). IMF lending: How is itaffected by economic, political and institutionalfactors? Journal of Policy Reform, 4, 243–270.

Bird, G., & Rowlands, D. (2002). Do IMF programshave a catalytic effect on other international capi-tal flows? Oxford Development Studies, 30(3),81–98.

Bird, G., & Rowlands, D. (2003). The demand for IMFassistance: What factors influence the decision toturn to the Fund. Department of Economics Work-ing Paper Series, University of Surrey.

Brand, L. (1994). Jordan’s inter-Arab Relations: Thepolitical economy of alliance making. New York:Columbia University Press.

Burnside, C., & Dollar, D. (2000). Aid, policies andgrowth. American Economic Review, 90(4), 847–868.

Collier, P., & Dollar, D. (2002). Aid allocation andpoverty reduction. European Economic Review, 46(8),1475–1500.

Conteh-Morgan, E. (1990). American foreign aid andglobal power projection. London: Dartmouth Pub-lishing Co.

Conway, P. (1994). IMF lending programs: Participa-tion and impact. Journal of Development Economic,45, 365–391.

Dallas, R. (1999). King Hussein: A life on the edge.London: Profile Books.

DFID (2003). Middle East and North Africa RegionalAssistance Plan. London: Department for Interna-tional Development.

Dollar, D., & Svensson, J. (2000). What explains thesuccess or failure of structural adjustment pro-grammes? Economic Journal, 110, 894–917.

Dreher, A., & Vaubel, R. (2002). Does the IMF causemoral hazard and political business cycles? Evidencefrom panel data. Economic Working Paper Archiveat WUSTL in its series International Finance, No.0207002.

Dreher, A., & Vaubel, R. (2004). Causes and conse-quences of IMF conditionality. WUSTL EconomicWorking Paper Archive, International FinanceSeries, no. 0309004.

Economist. Revolution delayed. September 5, 2002.Economist Intelligence Unit (EIU) (2001). Country

Profile Algeria 2001. London, EIU.El-Ghonemy, M. R. (1998). Affluence and poverty in the

Middle East. London: Routledge.Feeny, S., & McGillivray, M. (2002). Modelling inter-

temporal aid allocation. CREDIT Research Paper,No. 02/10, University of Nottingham.

Fleck, R. K., & Kilby, C. (2001). World Bank indepen-dence: A model and statistical analysis of USinfluence. Vassar College Department of EconomicsWorking Paper Series, No. 53.

Frank, A. G. (1969). Latin American, underdevelopmentor revolution. New York: Monthly Review Press.

FreedomHouse. Available from www.freedomhouse.org.Goldstein, M. (2003). IMF structural conditionality:

How much is too much? In M. Feldstein (Ed.),Economic and financial crises in emerging marketeconomies. NBER conference series (pp. 363–437).Chicago: University of Chicago Press.

Harrigan, J., & Wang, C. (2004). A new approach to aidallocation among developing countries: Is the USmore selfish than the rest? School of EconomicStudies Working Paper Series No. 04/11. Universityof Manchester.

Hayter, T. (1971). Aid as imperialism. New York:Penguin Books.

Hayter, T. (1981). The creation of world poverty: Analternative view to the Brandt report. London: Pluto.

Hensman, C. R. (1971). Rich against poor: The reality ofaid. London: Allen Lane.


Hinnebusch, R. (2003). The international politics of theMiddle East. Manchester and New York: Manches-ter University Press.

Hubbell, S. (1998). The containment myth: US MiddleEast policy in theory and practice. Middle EastReport, Fall 1998, p. 9.

Jalee, P. (1968). The pillage of the third world. New York:Monthly Review Press.

Joyce, J. P. (1992). The economic characteristics ofIMF programme countries. Economic Letters, 38(2),237–242.

Kanovsky, I. (1989). Jordan’s economy: From prosper-ity to crisis. Mosha Dayan Centre, Tel Aviv Univer-sity.

Killick, T. (1995). IMF programs in developing countries:Design and impact. London: Routledge.

Knight, M., & Santaella, J. (1997). Economic determi-nants of IMF financial arrangement. Journal ofDevelopment Economics, 54, 405–436.

Leech, D. (2002). Voting power in the governance of theInternational Monetary Fund. Annals of OperationsResearch, 109, 375–397.

Leech, D., & Leech, R. (2003). Voting power in theBretton Woods Institutions. Paper presented to theDevelopment Studies Association Conference, Glas-gow, 10–12th September, 2003.

Maizels, A., & Nissanke, M. K. (1984). Motivations foraid to developing countries. World Development,12(9), 879–900.

McGillvray, M. (2003). Modelling aid allocation: issues,approach and results. Journal of Economic Develop-ment, 28(1), 171–188.

McGillivray, M., & White, H. (1993). Explanatorystudies of aid allocation among developing countries:A critical survey. Institute of Social Studies WorkingPaper No. 148, The Hague: Institute of SocialScience.

Mckeown, T., Pallansch, L., Thacker, S. (1999). Politicalconditionality in US bilateral and multilateral for-eign assistance. Paper presented at the 40th AnnualConvention of the International Studies Association,Washington, DC, February 16–20.

McKinlay, R. D., & Little, R. (1977). A foreign policy ofUS bilateral aid allocation. World Politics, 30(1),58–86.

McKinlay, R. D., & Little, R. (1978). The French aidrelationship: A foreign policy model of the distribu-tion of French bilateral aid, 1964–70. Developmentand Change, 9, 459–478.

McKinlay, R. D., & Little, R. (1979). The US aidrelationship: The test of recipient need anddonor interest models. Political Studies, 27(2), 236–250.

Mercer-Blackman, V., & Unigovskaya, A. (2004).Compliance with IMF program indicators andgrowth in transition economies. Emerging MarketsFinance and Trade, 40(3), 55–83.

Mosley, P., Harrigan, J., & Toye, J. (1995). Aid andpower. London and New York: Routledge.

Murphy, E. (2002). The foreign policy of Tunisia. In R.Hinnebusch, & A. Ehteshami (Eds.), The foreignpolicies of Middle East States. London: LynneRienner.

Mussa, M., & Savastano, M. (2000). The IMF approachto economic stabilization. NBER MacroeconomicsAnnual, 14, 79–122.

Niva, S. (1998). Between clash and cooptation: USforeign policy and the spectre of Islam. Middle EastReport, Fall 1998, pp. 26–29.

Oatley, T., & Yackee, J. (2000). Political determinants ofIMF balance of payments lending: The curse of thecarabosse? IMF Working Paper 2000. IMF: Wash-ington DC.

Perthes, V. (1998). Points of differences and cases forcooperation: European critique of US Middle Eastpolicy. Middle East Report, Fall 1998.

Pfeifer, K. (1996). Between rocks and hard choices:International financial and economic adjustment inNorth Africa. In D. Vandewalle (Ed.), North Africa:Development and reform (pp. 26–53). New York:St. Martin’s Press.

Pfeifer, K. (1999). How Tunisia, Morocco, Jordan andeven Egypt became IMF success stories. The MiddleEast Report, 29, 23–26.

Richards, A., & Waterbury, J. (1996). A political econ-omy of the Middle East. Colorado: Westview Press.

Rodrik, D. (1995). Why is there multilateral lending.NBER Working Paper (June) No. w5160.

Rowlands, D. (1995). Political and economic determi-nants of IMF conditional credit allocations: 1973–89.Norman Paterson School of International AffairsDevelopment Working Paper, Ottawa: NPSIA.

Rowlands, D. (1996). New lending to less developedcountries: The effect of the IMF. Canadian Journal ofEconomics, 29.

Schoultz, L. (1982). Politics, economics, and US partic-ipation in multilateral development banks. Interna-tional Organisation, 36, 537–574.

Shafiq, N. (Ed.) (1998). Prospects for Middle Eastern andNorth African economies: From boom to bust andback? London: McMillan.

Shultz, G. (1993). Turmoil and triumph: My years assecretary of state. New York: McMillan.

Smith, F. Jr. (1984). The politics of IMF lending. TheCato Journal, 4(1), 211–247.

Swearingen, W. D. (1996). Agricultural reform in NorthAfrica: Economic necessity and environmentaldilemma. In D. Vandewalle (Ed.), North Africa:Development and reform (pp. 67–93). New York:St. Martin’s Press.

Swedberg, R. (1986). The doctrine of economic neutral-ity of the IMF and the World Bank. Journal of PeaceResearch, 23(4), 377–390.

Thacker, S. C. (1999). The high politics of IMF lending.World Politics, 52(1), 38–75.

US Agency for International Development’s (USAID).Website: Foreign aid in the National Interest: Pro-moting freedom, security, and opportunity. Avail-able from www.usaid.gov/fani.

Waterbury, J. (1998). The state and economic transitionin the Middle East and North Africa. In N. Shafik(Ed.), Prospects for Middle Eastern and North Afri-can countries: From boom to bust and back? London:McMillan.

Zimmerman, R. F. (1993). Dollars, diplomacy anddependency. Boulder: Lynne Rienner Publishers.

APPENDIX A. HISTORY OF IMF LENDING ARRANGEMENTS IN MENA

Date ofarrangement

Date ofcancellation

Amountagreed

Amountdrawn

Amountoutstanding

AlgeriaExtended fund facility May 22, 1995 May 21, 1998 1,169,280 1,169,280 525,487Standby arrangement May 27, 1994 May 22, 1995 457,200 385,200 0Standby arrangement Jun 03, 1991 Mar 31, 1992 300,000 225,000 0Standby arrangement May 31, 1989 May 30, 1990 155,700 155,700 0

EgyptStandby arrangement Oct 11, 1996 Sep 30, 1998 271,400 0 0Extended fund facility Sep 20, 1993 Sep 19, 1996 400,000 0 0Standby arrangement May 17, 1991 May 31, 1993 234,400 147,200 0Standby arrangement May 15, 1987 Nov 30, 1988 250,000 116,000 0

JordanStandby arrangement Jul 03, 2002 Jul 02, 2004 85,280 10,660 10,660Extended fund facility Apr 15, 1999 May 31, 2002 127,880 127,880 127,880Extended fund facility Feb 09, 1996 Feb 08, 1999 238,040 202,520 113,738Extended fund facility May 25, 1994 Feb 09, 1996 189,300 130,320 35,205Standby arrangement Feb 26, 1992 Feb 25, 1994 44,400 44,400 0Standby arrangement Jul 14, 1989 Jan 13, 1991 60,000 26,800 0

MoroccoStandby arrangement Jan 31, 1992 Mar 31, 1993 91,980 18,396 0Standby arrangement Jul 20, 1990 Mar 31, 1991 100,000 48,000 0Standby arrangement Aug 30, 1988 Dec 31, 1989 210,000 210,000 0Standby arrangement Dec 16, 1986 Apr 30, 1988 230,000 230,000 0Standby arrangement Sep 12, 1985 Dec 15, 1986 200,000 10,000 0Standby arrangement Nov 15, 1959 Mar 15, 1985 833,250 0 0Standby arrangement Sep 16, 1983 Mar 15, 1985 300,000 300,000 0

TunisiaExtended fund facility Jul 25, 1988 Jul 24, 1992 207,300 207,300 0Standby arrangement Nov 04, 1986 May 31, 1988 103,650 91,000 0

Source: IMF (amount in thousands of SDR).Note: Seven cases are not reported here: Djibouti, 1996 SAF and 1999 PRGF; Yemen 1996 SAF and 1997 PRGF andESAF; Egypt 1976, SAL and Morocco, 1980 SAL.


APPENDIX B. HISTORY OF WORLD BANK LENDING ARRANGEMENTS IN MENA

Project name CommitmentUS$ million

Country Date ofapproval

Economic Reform Support Loan 300 Algeria Aug 31, 1989Enterprise and Financial Sector Adjustment Loan 350 Algeria Jun 21, 1991Economic Rehabilitation Support Loan 150 Algeria Jan 12, 1995Structural Adjustment Loan 300 Algeria Apr 25, 1996Agricultural Industrial Imports 70 Egypt Dec 03, 1974Agricultural Industrial Imports (02) 70 Egypt Jun 14, 1977Structural Adjustment Loan 300 Egypt Jun 21, 1991Industry and Trade Policy Adjustment Loan 150 Jordan Dec 14, 1989Energy Sector Adjustment Loan 80 Jordan Oct 07, 1993Agriculture Sector Adjustment Loan 80 Jordan Dec 08, 1994Economic Reform and Development Loan 80 Jordan Oct 24, 1995Economic Reform and Development Loan (02) 120 Jordan Dec 11, 1996Economic Reform and Development Loan (03) 120 Jordan Jun 01, 1999Public Sector Reform Adjustment Loan 120 Jordan Jun 21, 2001Public Sector Reform Adjustment Loan (02) 120 Jordan Jul 02, 2002Industrial and Trade Policy Adjustment Loan 150.4 Morocco Jan 31, 1984Agricultural Sector Adjustment Loan 100 Morocco Jun 20, 1985Industrial and Trade Policy Adjustment Loan (02) 200 Morocco Jul 16, 1985Education Sector Reform Program 150 Morocco Mar 20, 1986Public Enterprise Rationalization Loan 240 Morocco May 26, 1987Agricultural Sector Adjustment Loan (02) 225 Morocco Nov 24, 1987Structural Adjustment Loan 200 Morocco Dec 01, 1988Financial Sector Development 235 Morocco Jun 25, 1991Structural Adjustment Loan (02) 275 Morocco Apr 30, 1992Financial Markets Development Loan 250 Morocco Jul 27, 1995Contractual Savings Development Loan 100 Morocco Jun 09, 1998Post Information Technology 101 Morocco May 06, 1999Policy Reform Support Loan (PRSL) 250 Morocco Jun 01, 1999Information Infrastructure Loan 65 Morocco May 31, 2001Asset Management Reform Loan 45 Morocco Jun 05, 2003Agricultural Sector Adjustment Loan 150 Tunisia Sep 18, 1986Industrial and Trade Policy Adjustment Loan 150 Tunisia Feb 24, 1987Structural Adjustment Loan 150 Tunisia Jun 16, 1988Agricultural Sector Adjustment Loan (02) 84 Tunisia Jun 01, 1989Public Enterprise Reform Loan 130 Tunisia Jul 11, 1989Economic and Financial Reforms Support Loan 250 Tunisia Dec 12, 1991Economic Competitiveness Adjustment Loan 75 Tunisia Jul 25, 1996Economic Competitiveness Adjustment Loan (02) 159 Tunisia Apr 20, 1999Economic Competitiveness Adjustment Loan (03) 252.5 Tunisia Dec 20, 2001

Source: World Bank Project Database.Note: Six cases are not reported here: Iran, 1957 SAL; Lebanon, 1977 SAL; Djibouti, 2001 SAL; Yemen, 1996, 97, 99SAL.


APPENDIX C. VARIABLE DEFINITION AND DATA SOURCE

D(RES/GDP) Changes in net reserves (BoP, current US$)/GDP(current US$)CAB Current account balance (% of GDP)TDEBTS Total debt service (% of exports of goods and services)GDPG GDP growth (annual %)LGDPPC Log[GDP per capita (constant 1995 US$)]SDEBT Short-term debt (% of total external debt)

(World Development Indicator 2002 CD-ROM)DEM Democracy index (1–7)

(Freedomhouse: www.freedomhouse.org)DELEC Dummy, is there a legislative election? (1 if yes)

(World Bank DPI database)PEACE Dummy, the year that a country under peace treaty with Israel (1 if yes)Country list: Algeria, Djibouti, Egypt, Iran, Jordan, Libya, Morocco, Oman, Syrian, Tunisia,


and YemenYears covered: 1975–2000

http://www.freedomhouse.org

the economic and political determinants of imf and world

Documents