The IMF, the World Bank, and Economic Policy in Bosnia

8/7/2019 The IMF, the World Bank, and Economic Policy in Bosnia

http://slidepdf.com/reader/full/the-imf-the-world-bank-and-economic-policy-in-bosnia 1/79



T h e IMF, the W orld Bank,

an d Econom ic Policy in Bosnia:

a Preliminary Assessment

David Woodward

An Oxfam Working Paper



©Oxfam GB1998

First published by Oxfam GB 1998.

ISBN 0 85598 396 5

A cata logue re cord for this publication is available from the British Library.

All rights reserved . R eproduction, copy, transmission, or translation of any part of this publicationmay be made only under the following conditions:

• with the p rior written permission of the publisher; or

• with a licence from th e Copyright Licensing Agency Ltd., 90 To tten ham Co urt Road, Lon donW l P 9HE, UK, or from another national licensing agency; or

• for quotation in a review of the work; or

• un de r the terms set out below.

This publication is copyright, but may be reproduced by any method without fee for teachingpurposes, but not for resale. Formal permission is required for all such uses, but will be grantedimmediately. For copying in any other circumstances, or for re-use in other publications, or fortranslation or adaptation, prior written permission must be obtained from the publisher, and a feemay be payable.

Available from the following agents:

for Canada and the USA : Hum anities Press Internation al, 165 First Avenue, Atlantic H ighlands,New Jerse y NJ 0771 6-1289, USA; tel. 732 872 1441; fax 732 872 0 717;for Southern Africa: David Philip Publishers, PO Box 23408, Claremont, Cape Town 7735

Sou th Africa; tel. 021 644136 ; fax 021 643358;for Autralia: Bush Books, PO Box 1370, Gosford South, NSW 2250 , Australia; tel. 043 23274; fax 029212248.

For the rest of the world, contact Oxfam Publixhing, 274 Banbury Road, Oxford OX2 7DZ, UK.

Published by Oxfam G B, 274 Banbury Road, Oxford OX2 7D Z, UK

Typeset by Oxfam CSUPrinted by Oxfam Print Unit

Oxfam GB is registered as a charity, no. 202918 and is a m em ber of Oxfam International.

This book converted to digital file in 2010



Contents

PREFACE 4

I N T R O D U C T I O N 5

A. POLITICAL AND ECONOMIC BACK

G R O U N D 7

Al Th e war and its consequences 7A2 Political structures un de r the Dayton

Agreement 9A3 Th e pre-war economy 11A4 Recent economic developme nts in the

Form er Yugoslavia 12

B. THE IMF, THE WORLD BANK AND

ECONOMIC POLICY 14Bl Th e constraints on economic policy 14B2 T he debt dimension 18B3 IMF and W orld Bank relations with the

authorities 20

C. DONOR-RELATED ISSUES25C l Disbursement delays 25

C2 Political conditionality 27C3 Th e distribution of dono r funding 29C4 Donor coordination and NGOs 31

D. ECONOMIC POLICIES 32Dl Th e Central Bank as currency board 32D2 Public finances 35D3 Privatisation and the settlement of

dom estic liabilities 39D4 Private sector developm ent 45

D5 Financial sector rehabilitation 46D6 Tra de policy 47D7 Labour marke t deregulation 50D8 Utilities and public services 52D9 Agricultural policies 53

E. SOCIAL DIMENSIONS 55E1 Employment and the Emergency Social

Fund 55E2 Pensions and benefits 59E3 Health and education 64

F. CONCLUSION 68

DILEMMAS1

9

3

4

5

6

7

8

9

10

Political conditiona lity 27Th e Central Bank as currency board 34Inter-re giona l inequality and transfers 38How to privatise in the federation? 43Free entry for foreign banks? 47T ra de policy 48Labour market deregulation 50Cost recovery 52Employment creation and the economicstrategy 57Reinte gration or partition? 70

BOXES

1 Econom ic and social effects of the war 152 Divergen t views on BH accession to the

Europe an Union 243 Prospects for the industrial sector 40

FIGURES

1 T he political structu re 102 Ethnic distribution of aid projects inimple me ntation, October 1996 30

TABLES

1 Composition of production and employme nt in pre-1991 Yugoslavia 11

2 World Bank programm es in Bosnia andHerzegovina 21

3 Pledges, com mitm ents, projectimplementation and disbursements bydon ors, Octobe r 1996 26

4 Financial requirem ents, comm itmentsand project implementation by sectors,October 1996 29

5 Regional economic indicators, 1996 376 Estimated num ber of pensioners relative

to popu lation and labour indicators 597 Pensions at alternative levels of unem ploy

me nt and real wages 628 Wage contributions required for minimu m

unem ployme nt benefit coverage 639 Wage contributions required to restore pre -

war per capita expenditure on health 65

10 Wage contributions required forminimum pensions, unemploymentbenefits and health expe nditure 66



Preface

This paper was written during the first quarterof 1997, on the basis of a visit to Bosnia inJanuary of that year. Unfortunately, it has notbeen possible to up-date the draft since. Inview of the rapidly evolving situation, thisinevitably means that it is likely to be some-what dated in parts. Hopefully, it will nonethe-less be of some use in helping to understandthe economic issues and policy dilemmas in

Bosnia.I am very grateful to all those in Oxford

and Sarajevo (and in Banja Luka and Bijeljina)who helped in the planning, preparation andexecution of my visit to Bosnia. I am particu-larly grateful to Marijana Aksin-Macak for her

excellent translation; to Indir Kurtovic for hispatience and endurance during our longer-than-expec ted dr ive f rom Banja Luka toZagreb; and to all of Oxfam's staff in Sarajevofor their exceptional friendliness and hospitali-ty. I would also like to thank all those whogenerously spared their time to meet me inSarajevo and Banja Luka.

It should be emphasised that the views

expre ssed in the pape r a re those o f theauthor, and should in no way be attributed toOxfam.

David Wo odwardDecember 1997



Introduction

This paper is the second in a series of studiesbeing commissioned byOxfam on the role of theIMF and the World Bank in economic policy-making in post-conflict situations. Like the firstpaper on Rwanda (Woodward, 1996), itconsiders the economic effectiveness of theproposed strategy, its potential social impactand the implications for the achievement of aviable solution to the political problems of

Bosnia and Herzegovina (BH) in the long term .This pap er is more preliminary than that onRwanda. This partly reflects the currentpolitical situation: the au tho r's visit to BH tookplace a little over a year after the end of the w arand the signature of the Dayton Agreement,which established the framework for a com-pletely new political structure, and just fourmonths after the September elections, whicheffectively star ted the process of forming most ofthe new institutions. As a result, the institutionsof gov ernm ent are less well-established; and the

emphasis is primarily on reconstruction andinstitution-building rather than on the consid-eration and development of detailed economicpolicies for th e long te rm.

As a result of this, and of the complexity of theeconomic and political constraints, the paperconcentrates on identifying a number ofdilemmas in economic and donor policies,rather than on making alternative proposals.There are no simple answers in BH, andwhatever course is chosen is likely to raise asmany prob lems as it resolves.

The paper begins with an outline of the

political and economic situation of BH. Th e keypolitical, economic and social constraints onpolicy are set out, and the current relationsbetween the IMF, the World Bank and thevarious autho rities are discussed in Part B. Thisincludes an assessment of BH's debt situation,and the prospects for a satisfactory solution. Adiscussion of various donor-driven problems,such as delays in disbursements and politicalconditionality, follows in Part C. Part D containsan extended discussion of the policy proposalscurrently envisaged, in terms of their economic

viability, social impact and political imp lications.Part E assesses the social dimensions of econ-omic policy. The Conclusion seeks to drawtogether the diverse threads of the p aper.



6



A. Political and economic background

A.I The war and itsconsequences1

Prior to 1991, Yugoslavia was a federation,consisting of six repu blics (of which BH was one)and two autonomous territories (Kosovo andVojvodina) within S erbia. Decision-making wasvery decentralised un der the 1974 Constitution,

and at the central level each republic andautonomous territory had one representativeon a collective presidency. Central controlandcoordination were maintained by the personalauthority of Tito, leaving a serious powervacuum following his death in 1980.

The result was increasing incohesion withinthe Federation through the 1980s, and thereaw ake ning (and political exploitation) of rivalnationalisms an d historical resentme nts. Thesearose primarily from Yugoslavia's traumaticexperience during the Second World War,2

when the Nazis brought the extreme Croatnationalist Ustasas back from exile in Italy andinstalled them in power in part of Croatia andthe whole of Bosnia-Herzegovina (BH). TheUstasas 'were terrorists who had overnight beenhanded total power' (Bennett, 1995, p43), andthey had as a primary objective a virtualgenocide of the Croatian Serbs. 'The objective,somewhat crudely but not incorrectly stated,was to convert on e-third of Ortho dox Serbs intoCatholics, to expel another third to Serbia, andto exterm inate the rest' (Crnobrnja, 1996, p65).Estimates of the nu m be r of Croatian Serbs killedvary enormously between Serb and Croatsources. Th e mo re m ethodical estimates suggesta figure in the o rde r of 300,000.

Opposition to the Ustasa regime cameprimarily from Tito's partisans (who wereYugoslav in orientation, and came from allethnic groups, although the greatest numberwere Serbs); and the Serb nationalist Cetniks.After an initial period of cooperation, the twogroups came into increasingly bitter conflict as

Tito's wider political ambitions became appar-ent. BH , because of its ethnic mixture, became aparticularly important area of conflict. Muslimsand Croats were faced with a choice between at

least passive collaboration with the Ustasaregime, or joining the partisans. Inevitably,most chose the quiet life, leading Serb nation-alists to identify Cro ats with the U stasas, and tosee Muslims as collaborators. These caricatureswere revived and accentuated by nationalistpropag anda in the 1980s.

Th e revival of nation alism, actively pro mo ted

by (often crude ) media pro pag and a, culminatedin the brea k-u p of Yugoslavia in the early 1990s;and it was this process which led to the war inBosnia, starting in 1991. However, the conflictin Bosnia, more than any other part of theformer Yugoslavia, arose primarily as a result offorces external to the Republic itself. It is ironicthat it was here that the war was longest andbloodiest.

Th e process of disintegration began in earnestwhen the Serb nationalist Slobodan Milosevicmanoeuvred himself into power in Serbia on a

tide of populist nationalism. Once in power, heprocee ded to ex pa nd the power of Serbia withinthe then Federal Republic of Yugoslavia. Byfomenting Serb nationalism, he succeeded inousting the existing auth orities in the two auton -omous territories (Vojvodina and Kosovo),installing his own supporters in their place.Together with the pro-Serbian Montenegroauthorities, this gave Milosevic effective controlover four of the eight votes on the collectivepresidency.

Slovenia and Croatia became seriouslyconcerne d at the increase in Serbian po wer, andthe extrem e Serb nationalism which un derlay it.Th e Slovenian authorities ultimately responde dby declaring independence, and, after a briefattempt by the Yugoslav national army (JNA) toforce them back into line, achieved it. Croatiaquickly followed suit, motivated partly bySerbia's still stronger position in the Federalinstitutions in Slovenia's absence; partly bygrowing nationalism among the large Serbminority in Croatia; and partly by the increasingCroat nationalism which this provoked .

Croatia had much greater difficulty inescaping from Yugoslavia than had Slovenia.The intervention of the Yugoslav army was



The IMF, the World Bank and economic policy in Bosnia

more forceful and less successfully resisted; andthe Serb minority within C roatia, with JNA sup-port, took control of substantial areas of the countrywhe re they were in the ascend ant (for exa mp le,

the Krajina and Eastern Slavonia). A long andbitter struggle followed, and one-third of thecountry's territory remained under Serb controlwhen Croatia was internationally recognised inJan uary 1992. In early 1997, the Zagreb auth ori-ties have only just re-established control overEastern Slavonia, the last Serb-controlled area .

With the withdrawal of Slovenia and Croatiafrom the Federal institutions, Bosnia was facedwith a choice between secession and rem ainin gwithin a Federation now heavily dominated bySerb nationalists. The situation of the BosnianSerbs, however, was the m irror image of this —they faced a choice between being part of themajority within a rump Yugoslavia, or being anethnic minority within an independent Bosnia.The latter option was made more unattractiveby the rise in nationalism am ong the C roats andMuslims (itself largely a response to increasingSerb nationalism both in Serbia and withinBosnia), and by vigorously nationalistic Serbianpropaganda.

These dilemmas were made more acute by

Bosnia's ethnic comp osition, and th e ge ogra ph-ical spread of the various ethnic grou ps.3 Whilethere was some degree of ethnic diversitythrou gho ut the whole of Yugoslavia, it was mostmarked in BH: uniquely, no single ethnic g rou prepresented an overall majority of the popula-tion. I n 19 91,44 per cen t of the po pulation wereMuslims, 31.5 per cent Serbs, and 17 per centCroats. Of the remainder, 2 per cent were ofother ethnic groups, while 5.5 per centconsidered themselves Yugoslavs (mostlypeople of ethnically mixed pare ntag e). In all, 16

per cent of children in BH were of mixedparentage, a higher proportion than anywhereelse in the former Yugoslavia.

There was a significant regional pattern inthe distribution of ethnic groups, with adisproportionate presence of Serbs in WesternBosnia, and of Croats in Herzegovina.However, there was a very high level of integra-tion. Only three of 112 opstinas (the localadministrative areas) were essentially hom ogen -eous in ethnic terms, while 30 had no overallethnic majority. Of the remainder, 37 had an

absolute majority of Muslims, 32 of Serbs, and13 of Croats.

Th e authorities respond ed to the political dilem-ma by holding a referendum on independence

in March 1992. Muslims and Croats votedoverwhelmingly for secession; but there was analmost com plete boyco tt of the vo te by the Serbs.The tension between the two sides escalated,

and the Yugoslav national arm y — virtually th eonly remaining institution holding thefederation together — moved in, ostensibly tokeep the peace. However, as Croats andSlovenians had largely d eserte d, the JNA was bynow Serb-dominated; and the remainingBosniacs (Bosnian Muslims an d oth er s up por tersof a united Bosnia) also left the army as it cameinto conflict with their compatriots. In May1992, the army was notionally withdrawn fromBosnia; bu t in practice its Bosnian Serb co mpo n-ent remained, together with much of theweap onry, to joi n with local Serb m ilitias.

Thus, while the war was largely broughtabout by exte rnal forces, it was prim arily foughtbetween Bosnians. It was therefore extremelydivisive, splitting Bosnian society into its threemain ethnic components. For most of the war,the conflict was between the Serbs (whose mainobjective was secession and unification withSerbia) and an alliance of Bosniacs and Croats.However, in April 1993, the Bosniac/Croatalliance broke dow n, as it becam e clear that any

negotiated solution would take the form of athree-way division of territory, based at leastpartly on military control on the ground at thetime. There was also a large faction, at least onthe Croatian side, for whom the alliance was atactical one and the ultimate objective wassecession of the Croat statelet of Herceg-Bosnafrom Bosnia and its unification with Croatia.Full-scale war between Muslims and Croatsfollowed in May of the same year.

The conflict between the Croats and theBosniacs was brought to an end, following

political intervention by the US, in February1994. Nonetheless, this conflict seriouslyembittered relations between Bosniacs andCroats, especially in mixed areas. The legacy,which is most clearly illustrated by the currentsituation in Mostar, is a serious im ped ime nt tothe establishme nt of new institution s an d policy-formulation in the Muslim/Croat Federation.

Relations have been particularly badlyaffected by the process of 'ethnic cleansing'.Fear, intimidation, and violence (includingmassacres in some areas) have brought the

country from a multi-ethnic mix developed overcenturies to an almost complete polarisation ofthe population into distinctly Muslim, Croat,and Serb areas. Some mixed areas (generally

8



Political and economic background

Muslim/Croat) remain; but, with the notableexception of Sarajevo, there is a high level oftension, and in some areas, such as Mostar,ethnic cleansing and polarisation c ontinue.

In addition to the death of some 250,000peop le, the w ar and ethnic cleansing gave rise toenormous population movements, bothinternally and externally. It is estimated thatbetween lm and 1.4m people left the country asrefugees, while a furthe r million were displacedwithin the country. In total, this representsabout half of the country's pre-war population.While the intern al displacemen t largely repres-ented an exchange of members of one ethnicgroup for those of another, it gave rise to majorincreases in population in some areas (forexam ple, the M uslim part of Mostar). This bothadded to the strain on housing and infrastruc-ture resu lting from war damage an d dislocation,and compounded tensions between ethnicgroups in these areas.

While the Dayton Agreement envisaged areturn of refugees and internally displacedpeople to their areas of origin, this has yet tohappen on any significant scale, and it seemsunlikely that the economic, social, and politicalconditions for large-scale return will be

achieved in the near future.As well as its dramatic effect on inter-ethnic

relations, the war was also very destructive inphysical terms (the economic damage arisingfrom the war is described in greater detail inBox 1 in Section B1). Much of the da ma ge was tocivilian targets, particularly to economic andsocial infrastructure, such as roads, railways,water and electricity supplies, schools, healthfacilities and housing. However, the destructionwas very localised: within areas which re ma inedunder the control of the same ethnic groupthro ugh ou t, th e physical damage was relativelylimited, alth ou gh such area s are still affected bythe broader effects of the war; in other areas,wh ere th e conflict was most acute and prolonged,the destruction is almost total.

While the Serb area sustained less physicaldamage than the Muslim and Croat areas as aresu lt of the conflict, it was seriously affected bythe more general economic disruption asso-ciated with war, and especially by UN-imposedeconomic sanctions. These factors bro ugh t about

a drastic downturn in economic activity and analmost complete lack of maintenance toinfrastructure and productive facilities, whichdete riora ted seriously as a result.

A.2 Political structures underthe Dayton Agreement

Four years of war in Bosnia and Herzegovina

were brought to an end by the DaytonAgreement in December 1995. This establishedBosnia and Herzegovina (BH) as a singlesovereign country divided into two Entities: theSerb-controlled Republika Srpska (RS), with 49per cent of the territory and a population ofaround 1.2m; and the Bosniac/Croat-controlledFederation of Bosnia and Herzegovina, with 51per cent of the territory and a population ofarou nd 2.3m.4 The Federation in turn is dividedinto ten cantons, with boundaries essentially

fixed on ethnic lines: five are predominantlyMuslim, three predominantly Croat, and twoethnically mixed. There is no equivalentdivision of Republika Srpska (RS). The lowestunit of government in each case is themu nicipality. (See Figur e 1.)

The ethnic division is strongly reflected in thedistribution of power between the different levelsof gov ernm ent: in effect, power is concentrated asfar as possible at the high est level consistent withethnic uniformity. The central government hasvery limited power, its role being limited to the

bare m inimum necessary for BH to be consideredas a single sovereign country (foreign policy,international trade and debt, monetary policy,immigration, internation al law enforcement, andair traffic control), plus relations between theEntities (inter-Entity transp ort, comm unications,and law enforcement).

The Entities have responsibility across amuch broader range of issues, including manywhich are more commonly held by nationalgovernmen ts. Th ese include:

• defence;• intern al affairs, policing and justice;• taxation and customs administration;• agriculture, industry, and other economic

policies;• health and social policies;• environm ental policies;• refugees and displaced persons; and• reconstruction program mes.

Under the Dayton Agreement, each Entity isalso allowed to establish its own relationsh ips with

neighbouring states, and to enter into agree-ments with states and international organisationswith the c onsent of the (national) ParliamentaryAssembly.

9



Th e IMF, the World Bank and economic policy in Bosnia

Within the Federation, the cantonal level ofgovernment is responsible for education,housing, public services, local land-use, andsocial transfer expenditures. However, these

responsibilities are automatically devolved tothe municipal level in municipalities where themajority ethnic group is different from that ofthe canton as a whole. More generally, m unici-palities in the Fed eratio n h ave 'self-rule on localmatters'. In RS, the responsibilities of munici-palities correspond to those of the cantons andthe municipalities in the F edera tion.

Financially, the Entities retain control ofcustoms and excise receipts within theirrespective territories, while in the Federation,

cantons retain revenue s from sales, income , andproperty taxes, together with revenuesgenerated by charges for public services. Thenational government has no independent taxbase, at least for the time being, but is whollydependent on transfers from the Federation(two-thirds) and RS (one-third). Ultimately, itwill be able to raise tax revenues, if this isapproved by the national parliament.

Under the Dayton Agreement, elections wereto be held at all levels in bo th Entities within six tonine months, and this deadline was met (just)

when elections were held on 14 Septem ber 1996,at all levels except for municipalities within theFederation, where elections were postponedfollowing allegations of fraud. The elections

resulted in an ove rwhelm ing victory for all threenationalist parties. While not surprising, thisseriously complicates the process of securing aconstructive, collaborative approach to theformation of institutions, bo th at th e nation al leveland within the Federation. The postponedmunicipal elections are now ex pected in April.

One important issue remains outstandingund er the Dayton Agreemen t, namely the statusof the Brcko corridor —a narrow strip of land innorthern Bosnia, which links the eastern and

western parts of RS, and which is contested bySerbs, Croats and Bosniacs. The corridor iscurrently under Serb control, but is subject tointernational arbitration. This was originally tobe completed by mid-December 1996, but thedeadline has now been extended (for the secondtime) until February 1998. Whatever the find-ings, the ann oun cem ent of the arbitration resultscould raise the political temperature substantially,due to the central strategic imp ortan ce of the area,the mutually-exclusive claims of the variousparties, and the stre ngth of feeling on all sides.

Figure 1 . The political structure

LEVEL:

National Bosnia and Herzegovina(3-person collective presidency;

Council of Ministers;bicameral parliament)

Entity

Canton

Local

IFederation of Bosnia

and Herzegovina(President; government;

bicameral parliament)

10 cantons

73 municipalities

Republika Srpska(RS)

(President; government;bicameral parliament)

64 municipalities

10



Political and economic background

A.3 The pre-war economy

After a brief experiment with central planningafter the Second World War, Yugoslavia

positioned itself both politically andeconomically between the communist East andthe capitalist West. This strategic positionenabled the country to receive substantialfinancial sup po rt from Western dono rs throu ghthe 1950s and 1960s. In 1965-8, the re was atemporary move towards market socialism,before the p rinciple of worker self-management— a form of decentralised socialism — wasadopted. At the same time, there was a shifttowards commercial borrowing, leaving thecountry seriously exposed when the 1982

Mexican crisis led to a generalised loss ofconfidence in lend ing to developing coun tries.

While Yugoslavia's debt situation wassubstantially less serious th an those of the majorLatin American debtors, the authoritiesresponded with a series of IMF- and WorldBank-supported adjustment programmes andrescheduling of bilateral and commercial debts.This entailed a very rapid adjustment of thebalance of payments, and very large netresource transfers to creditors, especially from

the mid-1980s. This contributed to a seriousslow-down in growth from the very rapid ratesprior to 1980, rising unemployment, andhyperinflation — and thereby, arguably to theconflicts of the 1980s.

The pre-1991 Yugoslav economy wasrelatively well deve lope d and diversified, with astrong industrial base and good infrastructure.The workforce was skilled and well-educated,with 9.7 per cent of 20-24-year-olds in full-timetertiary education — more than in the UK,

Ireland or New Zealand, though less than inother developed countries (UNDP, 1992, Table34). While agriculture was an important sourceof employment, representing about one-

qu arte r of all jobs in th e economy, the drivingforces of the economy were industry and (inCroatia) tourism (Table 1).

BH was poorer than Yugoslavia as a wholebefore th e war, with GD P per capita 32 per centless than the national average in 1989 (Vojnic,1995, Table 4-2). Agriculture was substantiallymore important, particularly in terms ofemploym ent: it employed around 40 per cent ofthe work-force (about half part-time), andaccounted for about 14 per cent of output(EU/EBRD/World Bank, 1996b, pi03). The

economic importance of the industrial sectorappears to have been broadly in line with thenational average, while services were somewhatless significant, possibly reflecting the limiteddevelopm ent of tourism.

Major industrial sectors in BH includedenergy (coal, coke and electricity); other rawma terials (especially w ood a nd b auxite); textiles;leather and footwear; and machinery andelectrical equipment. Much of the formerYugoslavia's armaments industry was alsorelocated in BH, following Tito's split withStalin, as it was seen as being safer in the mostcentral of the Republics. Industrial p roduction,both extractive and manufacturing, was domin-ated by about a dozen large socially-ownedconglomerates, which were mainly export-oriented, with about a thousand small andmedium enterprises, directed mostly towardsthe domestic m arket.

Electricity production was mostly thermal,using domestically produced coal, but with

Table 1 : Composition of production and employment in pre -1 99 1 Yugoslavia(percent )

GDP Employment

(1990) (1987)

Agriculture

Industry

Services

10.9

44.8

37.5

25.8

26.8

47.4

Notes: Employment data are from World Bank (1989 ); GDP data are from World Bank (1992 a).

GDP figures exclude 6.7 per cent representing indirect taxes.

11




some hydroelectricity. Prior to the war, a smallsurplus was generated for export to Serbia andCroatia. Coal production was 17.9m tonnes in1990, but is subject to serious environmental

concerns: BH's brown coal and lignite have ahigh sulphur content; and desulphurisationsystems are not in place in power stations, andwould be exp ensive to instal. As well as coal andbauxite, BH has substantial reserves of iron ore ,lead, zinc, barytes, rock salt, and ceram ic clays.

In the services sector, BH had a majorcapacity in civil engineering and construction,with 500 companies generating around 7 percent of GDP, including a significant pro por tionof export and overseas contracts. Tourism,

however, remained relatively under-developed,despite some efforts at promotion following the1984 Winter Olympics. The services sectoraccounted for the bulk of private en terprise s.

In 1990, about 1.6m ha (30 per cent of BH'stotal area) was farmed, 95 per cent of it by some540,000 small private farms (generally less than10 ha5); and the cultivated area was dividedmore or less equally between c rops, pas ture , andmeadow/fallow. BH's terrain is an importantconstraint on agricu lture, 45 per cent of the totalarea being forest, and much of the remainder

too steep for cultivation; and agriculture wastechnologically backward even before the war,during which much of the available machineryand equ ipment was destroyed.

Major crops were wheat and maize, whilelivestock and forestry were very important, aswas horticulture in the south of the country.Before the war, between one-third and half oftotal food requ ireme nts had to be imp orte d; andcereal production was concentrated in what isnow RS, which had a substantial surplus inwheat production, while the Federation was

heavily depend ent on imports.Forestry and wood products are important

sectors in both Entities, accounting for some 10per cent of GSP,6 and $200m of exports beforethe war, and employing some 22,000 people inforestry alone. However, there are againimportant environmental constraints: deforest-ation 'could be seen in some areas of centralBosnia even before the end of the war' (EIU,1996a, p20).

A.4 Recent economicdevelopments in the Form erYugoslavia7

Apart from BH, where economic performancehas been greatly affected by the end ing of the warand the reconstruction programme, economicperformance has been relatively consistent acrossthe various countries which formerly made upYugoslavia. Economic reform is proceeding, or atleast beginning, thoug h slowly and g rudg ingly inthe Form er Republic of Yugoslavia (FRY: Serbiaand Mon tenegro); economic growth is relativelymodest, at around 2-3 per cent pa, with someexpectation of a gradual improvement in the

next year or two in most cases; un em ploy m ent ishigh and rising; and the balance of paymentscur rent account is generally in deficit, tho ug h tovarying degrees, and with marked variation inability to finance it. T he grea test difference is ininflation, which is well under control in Croatiaand Montenegro, but a potential problem inSlovenia, and very high in the FRY.

Bosnia and Herzegovina has performed verywell, at least statistically. Economic growth in1996 was estimated at around 35 per cent,reflecting a recovery of output from the

artificially low level in 1995 because of the war,coupled with the inflow of aid to finance recon-struction. However, ou tpu t remain s far below thepre-war level, and unem ploym ent is estimated at50-60 per cent. Inflation has been kep t very low,reflecting the convertibility of the d ina r at a fixedexchange rate against the Deutschmark, coupledwith tight fiscal and m onetary policies; an d, whilethere is a large current account deficit, diis isentirely appropriate during the reconstructionphase. Serious problems remain, as discussedelsewhere in this pap er.

Slovenia remains the richest part of theformer Yugoslavia, with income per head ofaround $10,000 — more than Greece orPortugal. It has been moderately successfuleconomically since breaking away from theformer Yugoslavia. Economic growth was 3.5per cent in 1995, and is expected to return tothis level in 1997 after dipping to about 2 percent in 1996 (though with stronger growth ininvestment). Inflation has returned to below 10per cent, and the current account deficit isrelatively small. A successful bond issueindicates renewed access to international finan-cial markets; but this could yet be th rea ten ed bythe FRY's legal challenge to the country'sagreem ent with its commercial cre ditors.

12



Political and economic backgroun d

Around half of state industry has beenprivatised, but foreign investment has beenlimited, and rising real wages have driven outlow-wage industries, leaving unemployment at

13 per cent (and rising). Slovenia's greatestadvantages relative to its former Yugoslavneighbours are its political stability, its relativelyhigh level of development, its very limitedinvolvement in the war, and its geographicalposition on the fringe of the E U. How ever, thebenefits of Slovenia's growth to the othercountries in the region will be limited by itspreference for strong er econom ic links with theEU rather than any development of strongerties with its Balkan n eig hb our s.

In Croatia, economic growth resumed ataround 4 per cent in 1996, despite heavy (andrapidly increasing) debts in the business sector,widespread insolvencies, strongly deflationarymacroeconomic policies, and very high realinterest rates. Despite renewed grow th, employ-ment continues to fall, and the unemploymentrate remains above 20 per cent. Inflation is stillrelatively low at 3.5^1 per cent, but there is asubstantial current-account deficit. A three-yearIMF programme is in preparation, and com-mercial debts have been reneg otiated.

The FRY (Serbia and Montenegro) has beenpursuing conservative fiscal and monetarypolicies, and is expected to negotiate a stand-byarrangement with the IMF in 1997 (if theremain ing obstacles to IMF m em bership can belifted8), as a basis for debt renegotiation.Broader economic reforms are also possible,

despite marked reluctance on the part of theSerbian leadersh ip: there has already been someliberalisation of foreign investment laws, aprivatisation law has been approved, and some

trade reform is expected. Progress on reformhas been faster in Montenegro than in Serbia.Economic growth was arou nd 2.5 pe r cent in thefirst half of 1996 (down from 6.5 per cent in1995); but inflation rem ains aro un d 30 per centpa, unemployment is high and rising, and thereis a large ($2bn) trad e deficit, financed fromhard-currency resources held abroad. Botheconomic performance and the prospects foreconomic reform are likely to have be en seriouslyaffected by political instability following thecontested local elections in No vem ber.

The Former Yugoslav Republic ofMacedonia (FYROM) has been pro m oted by theWorld Bank as a case of successful structu raladjustment,0 with significant p rog ress on privat-isation and banking reform. However, this hasyet to be reflected in overall econom ic perform -ance. Moderate economic growth was achievedin 1996, after a decade of stagnation anddecline; and inflation is in single figures anddeclining. However, exports fell by 30 per centin the first half of 1996, increasing an already

large current-account deficit; access to foreigncommercial lending has not been re gain ed; andunemployment continues to rise. Inter-ethnicpolarisation represents a potential threat topolitical stability and continued growth. Theregime for foreign direct investment is to beliberalised, and trade policy rationalised.

13



B. The IMF, the World Bank and economic po licy

B.I The constraints oneconomic policy

The formulation of economic policy in BH iscomplicated by three interlocking sets ofconstraints: political/institutional, social andethnic , and economic constraints.

Political/institutional constraints

These arise directly or indirectly from the termsof the Dayton Agreement. Among the actualterms of the Agreement, the most important isthe ope ration of the Central Bank as a curren cyboard for six years. This means, in effect, thatdomestic currency can only be issued againstforeign currency held by the govern me nt, whicheffectively removes governmental discretion inmonetary and exchange rate policy, andprevents budget deficits from being financed

domestically. (This is discussed further inSection Dl.)

A second key com pon ent is the administrativearrangement of the country, specifically thevery high (and arguably excessive) degree ofdecentralisation of power, to the Entity leveland, in the case of the Federation, beyond.While the four levels of administration in theFederation may have been necessary politically,to secure the agreement, the arrangement willincrease administrative costs substantially, andmay well prove cumberso me.

The reliance of the political structures oncooperation between ethnically-based politicalleaderships gives rise to a considerable level ofpolitical inertia, due to the mutual suspicionboth between the two Entities and between theMuslims and Croats within the Federation (andwithin ethnically m ixed cantons such as Mostar).This is exacerbated by the fact that Serb nation-alists, whose ultimate goal is secession from BHand unification with the FRY, have everyincentive to ensure that their policies andeconomic systems are incompatible with those

adopted by the Federation, as a means ofpushing BH towards partition. It should beno ted, how ever, that some observers (including

some donors) see the political tensions withinthe Federation as being at least as problema tic asthose between RS and the Federation.

Social and ethnic constraints

There are tensions at a number of levels; themost obvious being between the three majorethnic groups within BH. Such tension repre-sents a serious obstacle to any lasting politicalsettlement, and has a central role in the cu rren tpolitical inertia. These tensions also representan important constraint to economic (or other)interactions between the Entities, for example,the intimidation of people who cross betweenEntities (or their fear of violence), an d the likelyreluctance by some to buy goods from the othe rEntity, if they were available. It is hoped thatethnic tensions can be eased by increasingeconomic interaction between the Entities; but

the tensions themselves limit the pace at whichthis can be achieved. For economic and short-and long-term political reasons, it is thereforeessential to avoid compounding inequalitiesbetween the ethnic groups, or creating new orexacerbating existing sources of tension.

Another potential source of tension,especially in the Federation, is between thosewho stayed in Bosnia du ring the war and thosewho left as refugees. Those who stayed,especially in enclaves such as Sarajevo, Zenica,and Tuzla, have in many cases been impover-ished and traumatised by the experience . Manyof those who left — who include most of thosewh o were relatively well-off initially an d m any ofthe m ost skilled and ed ucated — have in varyingdegrees escaped these ill-effects; and manyprofessionals have accumulated substantialsavings while abroad. The resulting disparitiesin economic opportunities could give rise toconsiderable resentment as the refugees ret ur n,in direct proportion to the opportunitiesavailable. This would be further compounded

by programmes providing additional assistancespecifically for returnees, to give refugees anincentive to return.

14



The IMF, the World Bank and economic policy

Overlaying these social constraints is themore g eneral need to reduce p overty and accessto basic necessities, such as food, h ea lth serv ices,housing, water supply, and sanitation. This is

partly a moral issue, but also contributes to thepolitical dynamics: increasing poverty was animportant factor contributing to the conflict inthe Former Yugoslavia; and the persistence ofchronic poverty would be a substantial obstacleto a durable political solution in BH. Reducingpoverty would require the creation of employ-ment opportunities, not only for the 50-60 percent of the workforce who are currentlyunem ployed, bu t also potentially for m ore thana million refugees if and w hen they re tu rn .

Economic constraintsEconomic constraints arise not only from theeffects of the war, but also from a number ofother changes in the economic environment

which coincided w ith it. T he war dam age is themost obvious an d severe constraint. BH's econom-ic and social infrastructure and produ ctivecapacity has been severely red uc ed . (See Box 1.)

In addition, the human capital base has beenseverely eroded by the exodu s of a large nu mb erof educated and skilled people. Of those whoremain, people with specialist skills have beeninternally displaced, and it is likely that manywill no longer be in (or able to ret ur n to) an areawhere their skills can be used productively.

The Former Yugoslavia's strategic positionbetween East and West was lost with the end ofthe Cold War, and w ith it the p rospect of favour-able financial and economic support from bothsides. Reconstruction assistance has been a

temporary substitute, but will be of limitedduration (until about 1999). A viable economicfuture is unlikely to be based on continuedfunding at the pre-war level.

Box 1: Econ om ic an d social effects of the war

Population: Some 250 ,000 pe op le w ere killed in the war. A fur the r 1—1.4 m illion left th e

country as refugees, of whom only 6 per cent have returned; around a further million wereinternally displaced, of whom 360,000 remain homeless; and some 2-300,000 entered thecountry as refugees from neighbouring states. More than 200,000 people were wounded, ofwhom 13,000 have permanent physical injuries, including 5,000 who have lost limbs. About245,000 people have been or are being demobilised in the Federation and 180,000 in RS.This implies that around two-thirds of the pre-wa r population of 4.4 million have been killed,disabled, internally or externally displaced or demobilised. An estimated one-third of thepopulation (about 1.2million people) are suffering from post-traumatic stress disorder.

Landmines and other hazards: Th er e estimated to be between 1.5 m illion an d 4 million lan d-mines in BH, of which only about 15,000 had been located as of November 1996. These aremostly concentrated in bands abo ut 5km wide along former lines of confrontation, includ ing

areas around Sarajevo, Banja Luka, Zenica, Vitez, Mostar, Srebrenica, Tuzla, Medugorje,Bihac, Zepa, Gornji Vakuf, and Gorazde. In addition, there is other unexploded ordnance inareas of conflict, and booby traps remain in areas of ethnic cleansing. Coupled with uncer-tainty about the location of mines, this is a serious deterrent to land use. Mines have also beenlaid in some areas to obstruct the use and reconstruction of infrastructure.

Transport and communications: More than 2,000km of main roads are in need of immediaterepair, all railway lines have been rendered inoperable, and Sarajevo airport was partlydestroyed and closed to civilian traffic. All bridges connecting BH with Croatia weredestroyed, and a total of 70 bridges have yet to be rebuilt. Public transport vehicles and facili-ties were destroyed or damaged, or have become run down due to lack of maintenance.Around 50 per cent of the telecommunications network has been damaged or destroyed,

including transmission and switching equipment, buildings, towers, overhead cables and theentire backbone transmission network. Call-completion rates in the worst affected areas are1-2 per cent at peak times, com pared with 35 -38 p er cent before the w ar.

15



Th e IMF, the W orld Bank and economic policy in Bosnia

Box 1: Econom ic and social effects of the war (continued)

Energy and heating: Almost all power stations, thermal and hydro, were damaged eitherdirectly or by lack of maintenance during the war, and more than 50 per cent of electricitygeneration capacity was put out of operation. At the end of 1995, electricity generation wasestimated to be opera ting at only 10 pe r cent of capacity. In ad dition, 60 per cent of the trans-mission network and control system was seriously damaged and about 50 per cent of the dis-tribution network damaged or destroyed. District heating systems were either seriously dam-aged (as in Sarajevo, whe re the n um be r of house holds served fell by about two-thirds) or dis-rupted by lack of energy supplies (as in Banja Luka, where only 5-6 per cent of installedcapacity was in use in 1995-6). Further damage has occurred due to non-use and lack ofmaintenance.

Water and waste management: War dam age and lack of maintenance, compounded by popu-

lation movements, imposed serious strains on the water and sanitation systems. Leakage-lossrates for water increased from 30 per cent to 50 per cent, and 27 of 69 municipalities in theFederation no longer have 24-hour supplies. Clogged sewage and drainage systems havecaused extensive flooding. While war damage has been relatively limited in RS, facilities havesuffered from lack of maintenanc e. Cross-b ound ary access to water supplies has been d enied ,fur the r limiting access. Solid waste services have bee n affected by extensive damag e to equ ip-ment; and limited access to landfill sites due to landmines has led to over-use of small urbansites.

Housing: In the Federation, 50 per cent of the housing stock was damaged, and 6 per centdestroyed; in RS, 24 per cent was damaged and 5 per cent destroyed. Deterioration of theremainder has occurred due to lack of maintenance. About 90 per cent of destroyed and

damaged housing has yet to be repaired, and some 30 per cent of inhabited housing has noglass in the windows.

Industry: As well as physical destruction, industry and finance have been affected by the dis-ruption of domestic and external trade links, and by the legal confusion created by the con-flict. The industrial sector is operating at about 15-20 per cent of its capacity in the Bosniacarea of the Federation, and at 8-10 per cent in RS. Industrial activity has also been reduced,though to a much lesser extent, in Croat areas. Non-agricultural unemployment is estimatedat 50 per cent in the Federation and 60 per cent in RS.

Finance: The freezing of foreign exchange deposits at the beginning of the war created adeep distrust of the banking sector, compounding its already weak financial position. Banks

have large foreign exchange liabilities, while household deposits are negligible, and commer-cial and public sector deposits very low; 90 per cent of bank assets are non-performing. Thecapital of the large state-owned banks has been wiped out, while the newer private banks aresmall, under-capitalised and inexperienced. As a result, only limited amounts of short-termcre dit a re available, and only at very h igh real interest rates.

Agriculture: 70 per cent of farm equipment and 60 per cent of livestock were lost during thewar, and farm buildings and irrigation equ ipm ent dam aged or destroyed. Some 15 per centof farm land and 20 per cent of forests remain inaccessible due to landmines, and high-valueorchards and vineyards were destroyed. In 1995, wheat, maize and potato crops were 30-40per cent below pre-war levels, and fruit production was down by about 50 per cent. Forestryand food-marketing systems were also disrupted.

Health: More than 13,000 people sustained permanent physical injuries during the war, andcom mun icable diseases increased by between 100 pe r cent and 400 per cent. Between 35 per

16




Box 1: Econom ic and social effects of the war (continued)

cent and 50 per cent of the health infrastructure was destroyed, and 35 per cent of hospitalbed-capacity was lost. There has been a 50 per cent reduction in the number of active healthpersonnel, and there are serious shortages of essential drugs and other supplies. Infant mor-tality rates and premature deaths have increased dramatically, doubling in some areas.

Education: Around 70 per cent of schools were damag ed, destroyed or requisitioned for mili-tary use or as centres for displaced people. In addition, budgetary resources for salaries andother recurrent expen diture dried up in many areas; many teachers depa rted, to be replacedby unqualified volunteers; materials such as textbooks and basic furnitu re remain scarce; andmany schools are without water and/or electricity supplies. Bosniac areas were worst affected,while education was sustained in Croat areas by financial support from Croatia. In RS, therewas significant physical dam age, but the m ain problem was lack of rec urr en t ex pe nd iture du e

to the UN embargo and hyperinflation. Many students left full-time education prematurely,or suffered a hiatus, during the war. The quality of education was seriously affected, and boththe school year and the school day were shortened to accommodate a shift system.

Sources: This Box is based mainly on EV'/EBRD/World Bank (1996, passim), with supplementary information from World Bank

(1996b) on population, Muratovic (1997) on population and housing, W orld Bank (1995c) on landmines, and EBRD (1996) on

electricity and telecommunications.

The economic system of the FormerYugoslavia was proving unviable long beforethe war. Even without the war, some form ofeconomic transition would have been nece ssary.Yugoslavia's debt burden was also heavy priorto the war. In the light of the w ar-damag e to theeconomy, BH's share of the debt is clearly uns us-tainable. While die World Bank has already actedto provide some debt relief, and action isexpected from bilateral and commercial bankcreditors, some potentially serious problemsrem ain. (See Section B2.)

In addition to the foreign debt, the govern-ment faces a heavy burden of domestic liabil-ities, particularly foreign exchange depositsfrozen during the war; and arrears owed topensioners, soldiers, and public servants. Th eseliabilities are very considerable (in the order ofDM13bn); and failure to settle them in a satis-factory manner could be politically destabilis-ing. Some 80 per cent of all households hadforeign exchange deposits of up to DM 1,000frozen; and public sector workers (includingsoldiers) are owed substantial salary arrears.

Settling arrears to pensioners should berega rded as a high social priority, in view of thevery high incidence of poverty among elderlypeople.

Globalisation and the transition process else-where in Eastern Europe represent additionalchallenges. In particular, BH's exports mustcompete with those of a number of otherdeveloping country and Eastern Europeaneconomies with (in some respects) similareconomic endow men ts, as all scramble to attractforeign investment and increase their foreignexchange earnings. A viable economic futurewill depend on the c oun try's ability to co mp eteeffectively in the world m arke t.

Finally, the cu rre nt level of aid shou ld also be

regarded as a constraint. Clearly, it would bepossible to provide a dditional aid; but, given th etight financial constraints on most donors' aidprogrammes, this would have to be at theexpense of curr ent (or increased) allocations toother, mostly poorer, countries. The currentlevel of aid to Bosnia is already very highcompared with many other post-conflictcountries in the recent past, although it is bothricher and better-endowed with economicresources. Aid comm itments to Bosnia for 1996(GNP per capita $250-500) amounted to

around $400 per capita, compared with $80 -90for Rwanda (GNP per capita $80) and $60-70for Mozambique (GNP per capita $60) at anequivalent stage of reconstruc tion.

17




B.2 The debt dimension

Yugoslavia faced a significant debt problem priorto the war, and efforts at external adjustment to

service the debts represented an importantcontributory cause of the conflict. Now, BH hasinherited its share of the foreign debts of theformer Yugoslavia; and its debt situation is agrea t dea l m ore serious than that of Yugoslavia inthe 1980s. Bosnia's debt situation is more thanusually complicated. It is not simply a que stion ofhaving too much debt and needing to find ways ofred uc ing it to a sustainable level; in Bosnia, th ereare two additional dimensions to the problem.First, the debts accumulated by the Former

Yugoslavia had to be allocated between theindependent states into which it is now divided.Seco nd, the cost of servicing the Bosnian share ofthe debt has to be divided between the twoEntities: while the central govern me nt is formallyrespon sible for servicing the de bt, itis ultimatelyde pe nd en t on contributions from the Entities todo so.

The first issue has, in principle, beenresolved. Of the Fo rme r Yugoslavia's $15.3bn ofexternal debt at the end of 1991, $ 12.2bn couldbe divided between the Republics according to

the 'final beneficiary' principle.10

The remaining$3.1bn was divided up in prop ortion to the fiveindependent countries' shares of the FormerYugoslavia's IMF quota. This resulted in BHtaking responsibility for a total of $ 1.9bn of 1991debt, which, with accumulated arrears andcharges, has increased to $3.4bn. There is,however, a potential complication with respectto the part of this debt owed to the commercialbank s, as discussed later.

The second issue — the division of theBosnian debt between the Entities — was

originally to be resolved in line with the two-thirds/one-third rule used in other contexts.However, the entities have now agreed to usethe final beneficiary principle instead. Theswitch to the final beneficiary principle has twomajor implications. On a financial level, it islikely that the RS will end up pay ing significantlym ore , and the Fed eration significantly less, thanunder the two-thirds/one-third rule. At firstsight, this m ight app ea r to suggest the possibilityof the RS side reneging on the agreement.However, there is also a significant politicaldim ensio n. Th e final beneficiary principle is theapp roach normally used to divide up the debtsof a country (such as the Former Yugoslavia)which splits into two or more separate

inde pen den t states. In other words, by acceptingthis principle, the RS authorities appear to beaccepting a greater financial burd en in orde r tobe treated as if they were seceding from BH,

thereby removing a potential obstacle to theirdoing so.

Th e no rmal sequence of debt reduction is forthe bilateral debt to be renegotiated first,followed by the commercial bank d ebt, and onlythen (if at all) the multilateral debt. In Bosnia'scase, this might well be reversed. The onlyconcrete progress made to date has been with theWorld Bank, which has consolidated BH's$625m de bt over 30 years at 0.5 per ce nt intere st.This was done by treating Bosnia as a new

member of the Bank (rather than as a successorto the Former Yugoslavia), and was presumablypreferred to the option of including Bosniain the recent multilateral debt initiative forhighly-indebted poor countries (the 'HIPCInitiative') as providing more immediatebenefits."

Bilateral debt cannot be reduced, or evenrescheduled, until an IMF programme is inplace. This m akes the timing very uncertain — itis hop ed, but by no m eans assured (see below),that a Fund programme will be in place by Ju ne .

However, there is a much greater need forurgency in the case of the commercial bankdebt, which must reach the 'agreement-in-principle' stage12 before 30 Ju ne 1997. Th ereason for this is that the previous reschedu lingof the Former Yugoslavia's debt (the NewFinancial Agreeme nt, or NFA, of 1988) includeda joint and several liability' clause between theRepublics. This means that each individualRepublic could be held liable for the entiream ou nt of the Form er Yugoslavia's debts in theevent of non-payment by the other Republics.

The banks agreed to waive this part of theagreement in June 1996; but the extension ofthe w aiver beyond 30 Ju ne 1997 is conditionalon an agreement-in-principle on BH's bankdebts before that date.

In principle, it would be possible for thebanks to extend the deadline. However, thiswould require the support of banks holdingtwo-thirds of BH's debt; and the FRY hasbought enough of Bosnia's commercial bankdebt on the secondary market to be able to blocksuch a move. It is therefore imperative to reachagreement-in-principle before the end ofJu n e ,even if this is before th e Paris Club ag reem ent.Th ere is concern in some quarters that the term sachieved on commercial bank debt may be less

18



Th e IMF, the World Bank and economic policy

favourable (to BH) if it is agreed before the ParisClub; but this may be a price which has to be paid.

If agreement-in-principle is not reached intime, BH's debt would be re-merged with the

FRY's; and the FRY could block any reversal ofthis move, unless (ironically) their legal actionagainst the Slovenian debt deal were successful.Any debt agreement would then need to benegotiated jointly by the FRY and BH authorities(and the Macedonians if they were in a similarposition), and provide equivalent treatment toboth (or all) parties. This would seriouslycomplicate the negotiation process, delaying afinal settlement, possibly for some years; and,given the much more limited goodwill towardsthe FRY, it would almost certainly provide less

favourable terms. The prolonged hiatus wouldalso limit BH's access to international financialmark ets until there was a resolution of the issue.

Even if agreem ent-in-principle is reache d, itwould seem possible, at least, that the FRY cou ldblock the finalisation of the deal, which gener-ally requires a greater than two-thirds majorityof participating creditors. However, the relaxedattitude of the authorities to this issue suggeststha t they expect a looph ole to be available. Evenif it were not, BH's debt would remain legally

distinct from that of the FRY; and the blockageof a final deal would seriously reduc e its value onthe secondary market. This could allow theauthorities to buy back their debt informally onthe secondary market and, in effect, to cancel it.13

There are various possible motivations forthe FRY authorities' actions in buying Bosniandeb t an d seek ing to block a deal:14

• They may be hoping that the FRY couldobtain a more favourable deal on its owncomm ercial debt if it were treated join tly w iththat of BH. In these circumstances, theremight well be some pressure on banks fromtheir governments to provide debt reduc tion,so as not to jeopa rdise BH's econom icrecovery; and this would allow the FRY tobenefit from equivalent terms. Such pressurefor favourable treatment would be unlikely ifthe FRY debt were treated separately.

• The FRY authorities may hope to reducetheir own debts more directly. While anysubsequent re-division of the debts wouldagain be based on the final beneficiary

principle (so that the FRY-held debts wouldstill be owed by BH ), it mig ht be possible forthe FRY to exchange the debts it holds withother creditors, possibly at a discount, for

FRY debts. After a re-division, the FRYauthorities would then be in a position tocancel a substantial part of their own debt.

• It is also possible that there is a politicalmotivation. By re-merging their commercialdebts with those of BH, the FRY authoritiesmig ht be trying to restore an obstacle — albeita largely symbolic one — to BH's autonom y. Ifthe debts held by the FRY are those attribut-able to RS under the final beneficiaryprinciple, the authorities could off-set anynegative effect on RS by cancelling th eir de bts.This would allow the RS authorities to b enefitfinan cially as well as politically from the iracceptance of the final beneficiary principle for

the inter-Entity allocation of debts.

Some of these motivations, at least, could givethe RS authorities themselves an incentive to blockany extension of the joint and several liabilitywaiver. In the third case, there would be a directfinancial benefit to RS, in the form of a greaterdegree of debt reduction than would otherwise beavailable, as well as the political advan tages. In thefirst two cases, the benefits would accrue mainly toFRY; but the RS might wish to support this, eitherfor reason s ofSerb solidarity or in return for a quid

pro quo ofsome kind. In view of the likely financialcosts (in the form of less favourable treatment ofBH debt), a quid pro quo would be essential un de rthe first case.

Th e question of the complicity o r otherwise ofthe RS authorities is critical, because this couldencourage them to obstruct the achievement ofagreement-in-principle before the end of June.Th e m ost obvious way of doing this would be todelay the approval of an IMF p rogram me . T heabsence of an IMF programme would notrepre sent an insuperable obstacle to agreement-

in-principle, as the procedures for commercialbank renegotiation are somewhat m ore flexiblethan those of the Paris Club. Nonetheless, ther ewould be substantially greater impetus towardsreduction of commercial bank debt with an IMFprogramme and the prospect of a Paris Clubagreeme nt than without. If an IMF p rogram mewere delayed, this could lead to a m ore gene raldisruption of BH's financial arrangements, forexample by delaying World Bank adjustmentlending, and would effect negotiations with thecommercial banks.

As noted above, the $825m owed to bilateralcreditors can only be reduced once an IMFprogramme is in place. The authorities arehoping to secure a reduction of 80 per cent in

19



Th e IM F, the World Bank and economic policy in Bosnia

net present value (NPV) terms;15 but it seemspos sible, at least, that they will be offered only 67per cent. This might be topped u p to 80 per centafter a track-record of economic adjustment

policies has been established (in line with theHIPC Initiative). Much will depend on thestrength of the political pressures for morefavourable treatme nt.

A 67 per cent reduction in the NPV ofcommercial and bilateral debt, coupled with theconsolidation of World Bank debt would reducethe overall NPV of debt to about $1.5bn, or 250per ce nt of estimated 1996 exports. The absoluteNPV figu re would be increased over time by newborrowing, although this should be relatively

limited, and mostly on concessional terms. Theratio of NPV to exports shou ld decline rapidly ifthe predicted rapid growth of exports isachieved : based on the World Bank's projections(World Ban k, 1996b, Table5.2), th e NPV of BH'sexternal d ebt might reach about $2bn— around80 per ce nt of exports16 — in 2000.

However, the export projections appear veryoptimistic (quite apart from the political risks),so that the rate of improvement may besubstantially slower. Th e effect of slower expo rtgrow th is threefo ld: as well as reduc ing th e level

of exp orts, and thus increasing the NPV /exportratio directly, the need for additional bo rrowingto finance the larger current-account deficitincreases the level of the debt, and the higherlevel of de bt increases in terest payments, addingto the ne ed for additional borrowing.

If export growth were reduced by one-third,without a corresponding reduction in imports,and the resulting financing gap were filled withnon-concessional borrowing, the NPV/exportratio would remain at about 250 per cent in 2000.This ou tcome could be avoided by reducing the

growth rate of imports; but to neutralise dieimpact would imply imports some 20-25 per centbelow th e levels currently projected in 2000; and ,withou t the possibility of devaluation (see SectionD1), this would imply a similar reduction in GDP.This would have a serious impact on incomes andemployment.

B.3 IM F and World Bankrelations with the authorities

BH join ed the IMF in December 1995, and theWorld Bank in April 1996 (although itsmembership was back-dated to February 1993,when th e me mb ership of the former Yugoslavia

was terminated). It is now represented on theExecutive Boards of both institutions by theExecutive Director for the Ne therland s.

The most important obstacle to IMF and

World Bank membership was the need to settleBH's share of the arrears accumulated by theformer Yugoslavia. This was a relatively minorproblem in the case of the Fund, as the arrearswere fairly small (less than $50m). They weresettled by a bridgin g loan from bilateral don ors,enabling BH to join the Fund and make a firstcredit tranche drawing of 25 per cent of quota,without policy conditionality. This was thenused to repay the bridging loan.

Th e problem was much greater on the WorldBank side, as arrears to the Bank amounted tosome $450m. However, this problem wasresolved flexibly and constructively by the Bank,which consolidated the entire amount ofBosnia's $625m de bt to the Bank over 30 years,at 0.5 per cent annual interest. This could bedone only by treating BH as a new member ofthe B ank, rath er than as a successor state to theformer Yugoslavia, following the precedentestablished by Bangladesh in 1975.

Relations between the IMF and the WorldBank and the authorities were complicated

initially by the Muslim domination of thenational government prior to the September1995 elections, and the resulting suspicion ofthe RS and Croat political leaderships. Since theelections, problems have included the split ofresponsibilities between th e Federation a nd theRS, and the strong mutual suspicion betweenthe two; and the tensions between the Croat andMuslim sides within the Federation, resulting inthe continued de facto operation of the C roatianstatelet of Herceg-Bosna . These factors contrib-uted to a prolonged hiatus in the formation of

the national government, and a continuing stateof political ine rtia.

Nonetheless, the IMF and the World Bankmaintained contacts with at least some of theauthorities during this period. The IMF, inparticular, was among the first to establish linkswith the RS authorities, at the beg inning of theirrelations with BH immediately after the DaytonAgreement. The Bank was initially in contactonly with the Federation side, but h as, since theelections, established a more symmetricalrelationship. Both the Fund and the Bank are

now in regular contact with the authorities ofboth Entities as well as the national gov ernm ent.

The World Bank started to supportreconstruction in BH before the country actually

20




joined the Bank, through a Trust Fund managedby the Bank, funded by donor contributions.This also allowed p art of the Bank financing to beprovided on gran t terms. Th e first project was th e

Emergency Reconstruction Project in February1996. A further twelve projects have beenapproved since, for a total of $325.6m of IDAfunds ($1,045.5m including cofinancing by otherdon ors). (See Table 2.) Only five of the 13 loanshave included RS: the remainder, accounting for80 per cent of total funding and 83 per cent ofWorld Bank resources, have gone exclusively tothe Federation.

Further support isexpected to take the form ofthree approximately annual policy-basedstructural adjustment credits (SACs). BH is

expe cted to make the transition from concession-al IDA credits to non-concessional IBRD loans in1998, reflecting its increasing per capita incomeand improving creditworthiness.

An IMF programme is also expected in thenear future. While BH is eligible for concession-al funding under the enhanced structuraladju stm ent facility (ESAF), the first pro gra m m ewill be a twelve-month stand-by arrangement(SBA), on non-concessional terms. The SBAoption is preferred because it takes less time toprepare than an ESAF, which would entaillonger and more detailed policy discussions.The important consideration at this stage is toget an IMF programme in place, to allow debtnegotiations to proceed (see Section B2), rather

Table 2 : World Bank Programm es in Bosnia

Emergency Recovery

Emergency Farm Reconstruction

Water, Sanitation and SolidWaste Urgent Works

Emergency Transport Reconstruction

War Victims Rehabilitation

Emergency Education Reconstruction

Emergency District HeatingReconstruction

Emergency Landmine Clearance

Emergency Housing Reconstruction

Emergency Demobilisation andReintegration

Emergency Public W orks andEmployment

Emergency Electric PowerReconstruction

Transition A ssistance Credit

Total

DateEffective

3/96

4/96

5/96

6/96

6/96

6/96

7/96

8/96

8/96

8/96

8/96

9/96

9/96

Coverage

F

F

F

F

F

F/RS

F/RS

F/RS

F

F/RS

F/RS

F

F

F

F/RS

al l

and Herzegovina

AmountWorld Bank

45.0

20.0

20.0

35.0

10.0

10.0

20.0

7.5

15.0

7.5

10.0

35.6

90.0

270.655.0

325.6Notes: Data are from World Bank (1996c), and cover loans effective in November 1996.

Coverage: F = Federation only; F/RS = Federation and Republika Srpska.

(USSm)

Total

160.0

50.4

70.0

163.0

30.0

32.8

40.5

67.0

60.4

20.0

45.0

196.4

110.0

840.2205.3

1,045.5

21




than the nature of the support provided.However, the high cost of the funds providedsuggests either that they should not be drawn (a'classical stand-by'), or tha t they should be ad ded

to the reserves, rath er th an b eing used to financeimpo rts. Th e inten tion is to negotiate an ESAF,which w ould cover a further three years, whilethe SBA is in place .

The policy conditionality included in an SBAwould effectively be limited in BH's case. Theconditions relate only to macroecono mic variablessuch as the bud get deficit, th e m oney supply, andthe exchange rate; and these are in any caseeffectively imposed by the currency boardarrange me nt man dated un der Dayton. An ESAF

would go beyond this into structural policies.While this would require more extensive policydiscussions, with greater scope for dispute withand between the Entities, the policies entailedwould overlap significantly with the World BankSACs.

Th e exten t of pro-active policy-makingby th eauthorities app ears to be limited. Th e process isone not so much of negotiation as of consulta-tion: the IFIs consult the various authorities,formulate policy propo sals, and p ut them to theauthorities, who accept them. Moreover, the

donors (and other international agencies, suchas UNOHR) appear to be presenting a veryunited front: all are firmly committed to thegeneral p rinciple of econo mic reform, and seemagreed on virtually all the details. The result isvery heavy conditionality: the authorities arereminded that if they do not comply with theprogram me outlined, there will not be anotherdon or conference — or, by implication, furthersubstantial assistance for reconstruction.

It is clear that the threat is not limited to amajor failure of the reform. The OHR/EU/UST

(1997) paper, 'elucidated with the cooperationof the IM F and the W orld Ban k, states explicitlythat:

There is no room or hesitancy or halfmeasures and the

reform program must be implemented quickly and in

full...[I]f it were limited in urgency or scope...the

willingness of the international comm unity to provide

large-scale resources would...be likely to diminish.

As a general principle, such heavy-handedconditionality seems less tha n ideal. How ever, itwould probably be necessary to ensure the imple-men tation of any kind of policy framework — no tbecause of political oppo sition, b ut because of th einertia generated by the current politicalarrangements. Much of the initial part of the

reform process is eithe r essential to th e establish-men t of BH as a state or man dated by the DaytonAgreement (for example, the formation of theCentral B ank, customs autho rities, and so on).

Nonetheless, some elements of the policiesbeing thus enforced are less clear-cut; and nodistinction is made between the essential state-forming components of the policy proposalsand those relating to the particular economicstrategy to be adopted. Thus, while theestablishment of a regulatory framework forinternationa l trade is necessary, the abolition ofadministrative controls and the setting of tariffsat a low and uniform rate, with m inimal exem p-tions, is not. O ther me asures , such as privatisa-

tion, are wholly unrelated to the Dayton terms.However, the authorities appear to accept theprinciples of reform which have been putforward, most of which are probably eithersensible, or inevitable in the current circum-stances (subject to the concerns discussed inSection D). It is difficult to assess the balancebetween genuine commitment and compliancebased on financial dependence, but there is atleast an e lem ent of the form er. RS (1996) wouldappear to suggest that the financial aspect maybe of greater im portance, and genuine commit-

men t m ore limited, in RS.The unanimity of the donors is presumably

based on th e fact that much of the package beingproposed is essential, and that it will not beimplemented unless they speak with a singlevoice, and make it impossible for any of theautho rities to dilute or obstruct it.The elementswhich go bey ond the essential are at least accept-able enough to the rest of the internationalcommunity to be seen as a price worth payingfor securing the essentials — although there islittle evide nce of dissen t.

Three more specific concerns may be notedwith respect to the international community'sapproach to promoting the reform process:

• The underlying ideology does occasionallyshow through. The OHR/EU/UST (1997)paper, for example, concludes that: 'die timehas come to shift the emphasis of economicpolicy-makers [NB not for policy-makersthemselves to shift their emphasis] fromreconstruction after the ravages of war to policyreform after the ravages of socialism.'

• EU/EBRD /World Bank (1996a, p lO l)proposes using the media to promote theprinciples of policy reform and transition,through soap operas as well as educational

22



Th e IMF, the World Bank and economic policy

programmes, documentaries, and debates.This seems h ighly questionable, particularly inthe ligh t of the role of the m edia in th e conflict,and would seem to be at odds with the objec-

tive of increasing its political ind epe nde nce .

As well as the stick of reducing reconstructionassistance and withholding debt reduction, acarrot has been dangled before the audioritiesin the form of accession to the EuropeanUnion. While this is seemingly an almostuniversally-held objective, and there appea rs

to be a g reat deal of optimism on this score, theprospect of achieving it remains u ncertain, andat best distant. There is also a conspicuousdiverg ence of views, both between the EU itself

and other donors, and between the Federationand national authorities on the one hand andthe RS authorities on the other (Box 2). Thistactic would therefore seem to carry with it ahigh risk of disillusionment and frustration inthe medium term; and could create a riftbetween the two E ntities.

23




Box 2: Divergent views on BH accession to the European Un ion

Th er e is a ma rked disparity in th e views of different parties on the likelihood an d desirability of

BH's accession to the European Union in the foreseeable future. This can be seen from thefollowing quotes.

Th e donor vision of the ben efits:

Dono rs, in putting forward their policy recom mend ations, have put substantial emp hasis on thepotential benefits of EU accession and the necessity of market-oriented economic reforms as aprerequ isite for its achievem ent:

Joining the European Union and becoming increasingly integrated with the rest of Europe is the common

vision shared by the people of Bosnia. It is a vision ofprosperity for every citizen of Bosnia, and a vision of a

modern market economy, and underpinned by a small but effective governme nt ....To realise this vision of

integrating with Europ e w ould require forward-looking policies and reform.

(EU/World Bank, !99f>b, pi)

The National Government view:

Th e national gov ernm ent app ear s to aspire to EU accession.

The Governm ent of Bosnia and Herzegovina w ill continue its negotiation with the EU in relation to

gaining access to the PHARE programme and to conclude a Trade and Cooperation Agreement, and

Financial Protocol. The Government of Bosnia, and Herzegovina reiterates its position that Bosnia and

Herzegovina deserves special treatment, as an unprecedented case, for Associate Membership in the European

niOn- (World Bank, 1996d, plO3)

Howev er, there are two dissenting voices:

Th e broader EU view:

Despite its involvement in th e prep aration of the first docu me nt quo ted, elsewhe re the EU hasbeen much m ore ambivalent:

In the case of Bosnia-Herzegovina, the peace process, reconciliation between the various sides in the civil

war and reconstruction must irst be seen to bear ruit before further steps, such as the negotiation of a Europe

Agreement, can be taken. Only when they have borne fruit could the Pliare programm e be used as a means of

preparing Bosnia-Herzegovina ...for accession to tfieEU.1

's

°J (European Parliament, plO)

This suggests that the E uro pea n Parliament, at least, may be less enthusiastic about th e prospect

of BH membership than the Commission. Some other donor documents co-authored by theCommission itself have also been noticeably reticent on the possibility of BH accession:OHR/EU/UST (1997, p6), for example, says only that the EU is 'prepared to considerestablishing a contractual relationship' with BH.

T h e RS v iew:

While the second quote above is probably a realistic representation of the aspirations ofFederation representatives on the national government, it is by no means clear that theirenthusiasm is shared by the RS. Th e RS Policy Statemen t on 'Fore ign Econo mic Relation s Policy'(RS, 1996, pp l 19-120), for examp le, does not mention the EU, referring instead to a free-tradezone within the former Yugoslavia and 'projects for integration within the Balkans'.

24



C. Donor-related issues

C l. Disbursement delays

As is almo st invariably the case in reco nstruc tion(or other aid) programmes, there has been asubstantial difference between pledges andactual disbursements and project implementa-tion: as of November 1996, of $ 1,851m

committed, only $720m had been disbursed. Afurther $456m was under implementation, of

which contracts had been signed for $330m(EU/World Bank, 1996a, Table 3). However,this is no worse than is normal in suchcircumstances. A substantial p art of the $ 1,851 mcommitments — including all of the WorldBank programm es — was designed to be pha sedover a period beyond the end of 1996.Moreover, w here delays have occurred, this hasoften bee n as a result of political conditionality.(See Section C2.)

Disbursem ent delays may have ham pered thereconstruction process to some extent, which in

turn suggests some negative effect on employ-ment. However, their wider economic effectsare likely to have been limited, particularly asbalance-of-payments support has not beengreatly affected. The delays to the reconstruc-tion process may have some significance,however, if they m ean that other com ponen ts ofthe transition process, such as privatisation, takeplace at a lower level of infrastructure andeconomic activity than was anticipated. (Thepace of privatisation is discussed in Section D3.)

Disbursement delays, and particularly those

related to political conditionality, have alteredthe balance of assistance between the F eder ationand RS, as the latter has been much moreaffected. (See Section C3.) However, this doesnot so far appear to have done any seriousdamage to the political situation; and the needfor international financial support has been a

major factor in the development ofa more coopera-tive app roach on the p art of the RS authorities.

There are, inevitably, marked differencesbetween donors in the translation of pledgesinto actual disbursements. (See Table 4). Whilethis no doubt partly reflects differences in thenature of their support and its distributionbetween sectors, it seems likely that some moregeneral variations, for example, in political

commitment or bureaucratic efficiency, alsoplay a role.Overall, multilateral agencies have performed

somewhat better than bilateral donors in trans-lating pledges into commitments and projectsunder implementation; but a significantly largerproportion of pledges have been disbursed bybilaterals, partly du e to the ap pare nt failure of theEBRD to get any projects to the implem entationstage. The EU has also performed relativelyweakly in terms of disbursements, while theWo rld B ank has been well above average in bothimplementation and disbursement.

Among bilaterals, the G7 have mostlyperformed reasonably well, with the exceptionsofJap an and Italy. Smaller developed co untrieshave been more variable, the Netherlands per-forming well (possibly reflecting their status asBosnia's representative within the IMF andWo rld Ban k), while Spain has perform ed partic-ularly badly. The weakest performances areregistered by countries which might beexpected to be particularly partisan — ie theFRY, Russia and the Islamic countries (SaudiArabia, K uwait, Tu rke y an d Malaysia)" —

possibly reflecting lack of experience in aid orlimitations in their administrative capacity.Between them , these six countries made pledgesof $183 .5m (nearly 10 pe r cent of the total), bu tha d d isburse d only $5m (2.7 per cent) of this byOctober 1996, compared with an average of48.4 per ce nt for other bilateral donors.

25




Table 3: Pledges, commitments, project implementation and disbursements

by donors, October 1996

Donor

EU

World Bank

EBRD

Islamic Dev. Bank

other multi lateral

TOTAL MULTILATERAL

USA

Japan

Netherlands

Italy

Russia

Saudi Arabia

Norway

UK

Germany

Kuwait

Switzerland

Sweden

Turkey

Canada

Spain

Malaysia

Austria

FR Yugoslavia

France

other bilateral

TOTAL BILATERAL

ALL DONORS

Pledges

($m)

367.1

330.0

80.2

15.0

25.0

817.3

281.7

136.7

100.0

63.7

50.0

50.0

40.8

39.7

39.3

35.0

33.5

30.4

26.5

25.4

17.5

12.0

11.5

10.0

9.3

164.2

1,077.1

1,894.4

a:

commitments

105

99

10 9

12 7

95

103

10 5

70

10 0

111

0

84

10 4

98

99

100

82

10 3

43

88

101

10 0

200

0

14 2

12 6

94

98

> per cent of pledges

implementation

58

89

0

40

55

64

82

32

70

12

0

40

92

98

9 1

0

78

65

8

46

27

100

196

0

125

85

60

62

disbursements

28

49

0

40

5 1

35

44

27

58

12

0

10

92

86

72

0

36

6 1

0

46

27

0

1 0 1

0

118

57

40

38

Source: EU/World Bank (1996a), Annex 1.

26



Donor-related issues

C.2 Political conditionality

A critically important element in the equation ofdonor financing has been direct and indirect

political conditionality. Aid commitments havebeen seen by donors as an important means ofapplying pressure on the various politicalfactions and authorities (most notably in RS) tosecure progress on the implementation of theDayton Agreem ent. While it has not b een whollysuccessful, the actual or implied threat ofwithholding funding seems to have had someeffect in ter ms of political prog ress; and withoutit, it is unlikely that even the current level ofcooperation would have been achieved.

Political conditionality has been applied tovarious issues. Probably the clearest, as well asthe most important, has been on the formationof national institutions. Th e RS authorities haveno real political commitment to BH, or tocooperation with the Federation. Left tothemselves, they would therefore be unlikely tocooperate to construct institutions at thenational level. On the contrary, they have everyincentive (and, given the level of decentralisa-tion, every opportunity) to undermine theprospects of reunification by adopting policies

inconsistent or incompatible with those of theFederation. It may be possible to push themsome way towards cooperation by the thre at ofwithholding financial support.

In some areas, political conditionality hasbeen directly relevant to particular projects.This applies particularly to infrastructurerehabilitation projects, where donor supporthas been conditional on the linking of infra-structure between the two Entities. Where the

viability of the infrastru cture to be rehabilitateddepends on the reintegration of the Federationand RS com ponen ts, their continued separationrepresents an important obstacle to achieving

the objectives of such aid. Moreover, the natureof the support required depends critically onwh ether one integra ted or two separate systemsare being supported. To design rehabilitationon the basis of two ind ep end en t systems wouldnot merely provide implicit backing for theirseparation, but would entrench the separationin practical terms.

In other area s, political conditionality has beenless effective. For example, while donors weresuccessful in securing the removal of Radovan

Karadzic and Ratko Mladic from public office,ther e is little prosp ect of the ir extradition to facethe Hague war-crimes tribunal. Conditionalityon the freedom of movement and the return ofrefugees and displaced people appears to havebeen given less weight by donors, and has hadminimal effect.

There is a potentially important cost topolitical condition ality. Not only does it give riseto delays in disbursements, it also skews thepattern of disbursements between Entities; andthis applies particularly where, as at present, its

application is asymmetric. The main aim ofpolitical conditionality, from the donors' pers-pective, is to influence the policies of the RS. Butthe result has been that only a tiny proportion oftotal aid flows have reached RS. (See SectionC3.) While this could be effective in puttingpressure on the RS authorities, if economicrecovery in RS were jeopardised, it couldequally well strengthen the political position ofthe extremists.

Dilemma 1: Political conditionality

Some form of political conditionality is necessaryfor don ors. Providing supportto all parties irres-pective of their political actions wouldbe an acuteembarrassment, as it might well lead to donorsproviding financial support to policies in directcontravention of the Dayton Agreement. It hasalso proved to have some (though by no meanscomplete) success in influencing policies and pro-

moting the implem entation of the Dayton terms.However, some caution is needed in the

application of political conditionality. It cannot

necessarily be assumed that the RS authoritieswill comply with political conditionality, evenwhere they ap pea r to have a very strong finan-cial incentive to do so. They have in the pastproved very stubborn in the light of donorthreats , not least in the case of extradition. Evenif conditionality is successful in promotingcoope ration, if this is forced on d ie RSauthorities

by financial hreats w ithout any genu ine politicalcom mitm ent, it is likely to be superficial: it maybe possible to ensure the creation of national

27



Th e IM F, the World Bank and econom ic policy in Bosnia

Dilemma 1: Political conditionality (continued)

institutions, but it would be difficult to ensure a

constructive and cooperative approach withinthem (as Muslim/Croat relations within theFederation d emon strate). Only in the long termcould this be expected to con tribu te to politicalprogress, and then only if there were a shifttowards more moderate political leadership inboth Entities.

Moreover, an over-zealous application ofpolitical conditionality could result in inade-quate financial support to reconstruction andeconomic recovery, particularly in RS. Thiswould have a direct negative effect on the

economy, and thereby on living standards;and this, coupled with the opportunity toallege 'persecution' by the donors, wouldprovide considerable ammunition to politicalextremists. 'All-or-nothing' political condition-ality, if its requirements were seen by theauthorities as excessive, could lead the RS toturn their backs on the Dayton process, andlook to Belgrade for sup po rt instead . Th e verylow level of disburse me nts to RS so far suggestsa real possibility of m ov em ent in this direc tion.

It is also possible that the RS authorities mayhave an incentive to derail or delay someparticular aspects of don or sup po rt (eg die IMFprogramme; see Section B3). In such cases,political conditionality may simply providethem with an opportunity for do ing so.

This suggests a need for carefully grad uatedpolitical cond itionality, so that the amo untof aidreceived dep ends on the exten t of compliance,as against an all-or-nothing ap pro ach (howevertempting the latter might be to donors).Cond itionality sh ould also be selective, prioritis-

ing those areas which a re of greatest importanceto the peace process, which are also within thecontrol of the authorities. This suggests that thecreation of national-level institutions and diereunification of infrastructure should be regard-ed (as they are at presen t) as high priorities.

Unnecessary conditionality, or conditioningof aid on factors not fully within th e con trol ofthe authorities, would be counterproductive,and should be avoided. It would reduce theamount of aid the authorities would expect toreceive, weakening their incentive to comply

with other political conditions; it would risk

reinforcing the perception of the donors asbiased towards the Federation; and it wouldslow recovery and worsen living conditions inthe RS. All these factors would contribute toperpetuating extremism.

Conditionality based on the return ofrefugees and displaced pe ople is arguably a casein point. The obstacles to retur n ariseas much atthe popular level — the fear of violence andintimidation — as they do from the actions ofthe authorities. To impose political condition-ality based on the rate of return, when this is

largely beyond the control of the authorities,would have little effect. In fact, to th e exte nt th atliving conditions were worsened by die with-holding of aid, die (social) obstacles to returnwould be increased. Any condidonality in thisarea should be confined to specific adminis-trative and policy m easures .

The question of war criminals is moredifficult. T he re may be a case for conditionalityon the extradition of Radovan Karadzic andRadco Mladic, if they are seen as retainingpositions of influence with the RS authorities.

How ever, this is a m atter of political judg me nt,on which views differ. In oth er cases, economicand political considerations would appear tosuggest that conditionality should not beapplied, at least for the present, as the costswould be greate r than the benefits. It must bestressed, however, that considerations ofnatural justice also need to be taken intoaccoun t (and no t only because of the ir politicalimportance to donors). The relative import-ance of extradidon for political reconciliadon

is likely to increas e as (or if) oth er prob lem s a reresolved.Overall, do no r policies on political condition-

ality seem broadly appropriate. There is a riskdiat they are erring on the side of under-funding RS, but this issue is well recognised, andit is expected that disbursements to RS willincrease substantially in 1997. However, it willbe impo rtant to keep this issue un de r review, tomaintain a realistic assessment of what can beachieved, and to proceed with considerablecaution.

28




C.3 The distribution of donorfunding

A greater problem than disb ursem ent delays

has been that of financing gaps for particularprogrammes and components of programmes.Th e availability of don or funding and the rate ofproject implementation have varied consider-ably between sectors, as shown in Tab le 4. Whilesome areas of supp ort (for exam ple, housing and'government and social') have been over-sub-scribed relative to the original estimates of requ ire-ments, oth ers hav e faced serious shortfalls.

Because of the nature of some of theseshortfalls, there is a risk that this could substan-

tially worsen the social impact of the reformprocess.

• As of Novem ber, none of the expected $31mcofinancing from non-Bank-administeredfunds for the Emergency Public Works and

Employment programme had materialised,suggesting a reduction in the scale of the

project of mo re than two-thirds. This implies

a reduction in employm ent of about 5-6,000

jobs for the two years of the project, and a

small redu ction in real wage levels as a resultof the lower de ma nd for labour.

• The re has reportedly been a substantial short-fall in cofinancing for the Emergency SocialFund (ESF), although this is difficult to evaluateas figures for ESF financing are not p rovidedseparately by the B ank. The entire of the $44mfund was to be financed by othe r d onors;18 an dat the time of approval, only $13.6m of thishad been com mitted (World Bank, 1996c, p29).

• Com mitments to the health sector am oun tedto $83m, 43 per cent less than the estimatedneeds of $145m (EU/EBRD/World Bank,1996, pi3 2) . A particular gap h as been in th earea of health workers' salaries, where no

significant support has been provided,although 'most health workers are paidinsufficient salaries' (ibid, pi 37).

T he effects of these shortfalls are discussed inSections El (EPW and ESF) and E3 (health).

Table 4: Financial requirements, commitments, andimplementation by sectors, October 1 9 9 6

InfrastructureElectricity/Coal

District Heating/GasTelecommunicationsTransportWater

Productive SectorsAgricultureIndustry/Finance

Social SectorsEducationGovernment/SocialHealth

Housing

Miscellaneous

Landmine ClearancePeace Implementation

Balance of Payments

Commitmentsa s % o f

Requirements

8535175 160

65140

118136

58

16 5

6 1n/a

n/a

Implementation

as%of

Commitments

75553 15866

9653

637842

78

5289

8 4

project

Implementation

as%of

Requirements

64

195

3040

6274

75106

25

129

32n/a

n/a

Notes: All f igures are approximate, being based on figures transcribed manual ly from EU/World Bank (1996a), Chart 1 .

29




Figure 2 : Ethnic distribution of aid projects in imp lem entatio n,October 1996

Bosniac Croat Serb

National and Inter-Entity $165m

Federation:

General/Multi-Canton $462m

Federation:

Mixed Cantons $137m

Bosniac Cantons $361m

Croat Cantons $27m

RS $24 m

Source: Ell/World Bank (1996a), Annex 3, Table 2.C

Other particularly acute shortfalls in fundingcommitments have occurred in telecommunica-

tions (83 per cent) and natu ral gas (69 per cen t).In the former case, the problems have beenlargely a result of the lack of progress inreunifying the national network.

The regional distribution of disbursementshas also been strongly skewed. This is illustrated

in Figure 2, in which the area of each block isproportional to the amou nt of aid. Only about 2per cent of projects being implemented as ofOctober 1996 were in RS, with a similar

30




proportion in Croat cantons in the Federation.Most of the aid from which Serbs and Croatsbenefited w ere at the n ationa l level or (for Croats)to the Federation or mixed cantons — that is, it

represents their notional shares of aid providedto administrative units in which the Bosniacs arealso represen ted. It seems quite possible that thebenefits of such aid ar e skewed to wards Bosniacs;and , to the extent that it is aimed at p romo tingreunification and cooperation, this may be seenas furthering a primarily Bosniac objective.

Some skewing of aid towards Bosniac area s isjustifiable in economic term s, to the exten t thatit is here that the physical dama ge, hu m an costs,and economic disruption caused by the warwere greatest. However, the ex tent to which the

aid is skewed seems less clear-cut, since theeconomic situation in RS is now n o be tter — an dmay well be significantly worse — than inBosniac area s.

Th ere may be impo rtant costs from a politicalpoint of view. Firstly, the almost exclusive focusof aid on Bosniac and part-Bosniac adminis-trative units risks comprom ising p erception s ofpolitical neutrality on the part of the donors.This could weaken the legitimacy of theirattempts to influence the political process.Secondly, the minimal levels of financialsupport to the RS and to the Croat cantonsmea ns that th eir au thor ities a re likely to see littlefinancial incentive to comply with politicalconditionality. This could seriously jeopardisethe whole political process: quite simply, theSerb and Croat authorities may see little reasonto comply with the political objectives ofagencies they believe to be siding with theirop po nen ts, and little to lose by failing to d o so.

C.4 Donor coordination andNGOs

A third don or-driven problem arises from thefailure of donors to coordinate their activitieseffectively with the a utho rities. Accord ing to theBH Minister for Foreign Trade and EconomicRelations:

Overall reconstruction coordination is still not

satisfactory. The dono r sector task forces still operate

too much in isolation from the Governm ent.

(Muratovic, 1997, p4)

This applies particularly to NGOs, which insome cases fail even to inform the auth orities oftheir activities. This complicates the planningprocess, and results in geog raphical imbalances

in social provision. If the projects establishedwere ultimately to be continued with govern-ment resources, it could also skew governmentbudgets towards those areas favoured by NGOsin the future. This could, in principle, contrib-ute to increasing tensions within or betweenEntities.

This situation is now improving; and thebud get prob lem, at least, should be eased by theFederation authorities' intention that NGOprojects should ultimately be taken over by local

NGOs rather than by the public sector. N onethe-less, NGOs should, where necessary, increasetheir consultation and coordination with therelevant authorities; and the lesson should belearned for the future.

More generally, there is a certain feeling ofsuspicion with reg ard to the motivation s of bothNGOs and official international agencies. Manyare seen in some quarters as having hiddenagendas of one kind or another (particularly,but by no means exclusively, religious-basedNGOs); and both types of organisation, and at

least some of their staff, are seen as being m oti-vated by financial rather than purely altruisticmotives. The high salaries they pay (by localstandards) contribute to this problem, as doesthe habit of flying in expa triate c onsultants (ateven greater cost), often to perform tasks whichcould have been done (and in some cases areseen as already having been done) equally wellby locally-recruited per son ne l. Officials, at leastand quite possibly many ordinary people, areacutely aware tha t th e cost of all this comes ou t of

the funds pledged by donors, so that the moreextravagant foreign agencies are, the lessreconstruction gets don e.

Th e tendency of some N GOs to go their ownway, without consulting or informing the rele-vant authorities, does nothin g to he lp, particul-arly in a country with a strong bureaucratictradition. The fact that international agencieswere virtually absent from Bosnia until the lastfew years (in contrast to many poorer post-conflict countries) probably also contributes tothis problem .

31



D. Economic policies

D.I The Central Bank ascurrency board

Macroeconomic policy for the whole of BH islargely dictated by the requirement of theDayton Agreement that the Central Bank shouldoperate (for the next six years) as a currencyboard. Asnoted above, this means, in effect, thatcurrency can be issued only against the Central

Bank's h oldings of foreign currency; and this inturn prevents the authorities from running abudget deficit, unless it is financed by foreignloans or grants. It also means that the currency isconvertible (ie that there are no restrictions onbuying or selling foreign currency), and that theexchange rate is fixed (at 100 Dinars to 1Deu tschm ark), with no possibility of adjusting itup or down.

This arrangement is not unique—A rgentina,for exam ple , has op erate d a similar system since1991. Its main advantage from an economicpoin t of view is that it provides a mu ch highe rlevel of confidence in the currency, both locallyand internationally, than would be likely with amore discretionary arrangement—anyone whoholds D inars knows that he or she can exchan gethem at any time for Deutschmarks at a fixedexchang e rate. Together with the tight constraintit imposes on monetary policies, this means thatthe arrangement is very effective in keepinginflation down. (Hence its use in Argentina,after th e inflationary pro blem s of the 1980s.)

The currency board approach raises somepotential problems, in particular through thelimits it imposes on exchange rate policy (WorldBank, 1996a, pp 17-19). Normally, one wouldexpe ct to find a real appreciation of the ex chang erate d ur ing reconstruction, as the large-scale aidinflows increase the supp ly of foreign exch ang e.As reconstruc tion progresses, and th e associatedaid inflows fall, the appreciation would bereversed, and the exchange rate would fall.However, under the currency board arrange-ment, the exchange rate cannot be adjusted in

this way — instead, real exchange rates adjustthemselves through price changes. Th us, r ather

than an appreciation of the nominal exchangerate during the reconstruction phase, any realexchange rate adjustment would have to takeplace throu gh price inflation.

Th is is not, in itself, a very serious p roble m. Areal exchange rate appreciation will only benecessary if the economy app roac hes full utilisa-tion of capacity; and this seems unlikely ov er th ethree to four years of the reconstruction pro-

gra mm e. (See Section El.) Moreo ver, if the ca pa-city constrain t is reached, and a real ap preciationis necessary, this will occur, as excess dem an d willpush up prices for non-tradeable goods.

A much greater problem will arise when theeffect is reversed: in order to achieve a realdepreciation as reconstruction aid winds down,inflation would have to be lower than inGerm any (since the currency is fixed against theDeutsch mark ). Even with zero inflation, the rateof real depreciation would be inadeq uate .

This means, that, during the later stages ofreconstruction (say from about 1999 onwards),either domestic policies would have to be sodeflationary as to achieve a strongly negative rateof inflation, which would severely depress ou-tput, increase unemployment, and reduce realincomes; or the exchange rate would becomeincreasingly over-valued, making exports anddomestically-produced import substitutes lesscom petitive, which would result in faster accum-ulation of foreign debt or, if financing were notavailable, a foreign-exchan ge crisis.

Since unemployment will almost certainlystill be very high at this stage (see Section El),the former option would imply an extremelyhigh social cost. Apart from the general role ofworsening social conditions in exacerbatingethnic and political tensions, the effects ofdeflation would be seen as directly attributableto the national institutions. It seems almostcertain that this would be used by Serbnationalists as a further argument for secession— and it could be a fairly co nvincing on e. T hesame might also apply on the Croat side.Certain ly, the effect on the very fragile cohesionof the BH state would be extremely destructive.

32



Economic policies

The second option — allowing the exchangerate to become over-valued — might thereforeapp ear m ore attractive. However, this is at best atemporary solution; and the scope for taking

this path will be critically d ep en de nt on Bosnia'sdebt situation. There would again be a cost interms of employment, as export and import-substituting ind ustries would be less able to com-pete than with a more com petitive excha nge rate.

Th e over-valuation option is only viable to theextent that the country is able to borrow inter-nationally to finance its balance of paymentsdeficit. If it already has an uns ustain able level ofdebt, creditors will be unwilling to lend. More-over, their reluctance will be greatly increased(and the sustainable level of debt g reatly reduc ed)if creditors see the coun try's debt-servicing capa-city as being compromised by institutionalarrangements which prevent adjustment of anexchange rate which is already over-valued.

The problem of over-valuation could becompounded in the event of adverse politicaldevelopments, such as rising social or politicaltensions. During the reconstruction phase, thiscould disrupt financial flows, either because ofpolitical conditionality or by impeding policychanges or specific reconstruction projects. In

the longer term, it would represent a seriousobstacle to private borrow ing. In eithe r case, theeffect would be to reduce the supply of foreignexchange, and thus reduce the (notional)marke t exchange rate; and this would w iden thegap between the fixed official rate and thenotional market rate; in other words, it wouldincrease the deg ree of over-valuation.

Adverse political developments, or theperception of exchange rate over-valuation,could also cause a loss of domestic confidence. Ifpeople believe that the exchange rate is over-

valued to such an ex tent that some form of down-ward adjustment (ie a change in the currencyboard system) is likely, this will encourage peopleto convert their local currency into foreignexchange, to avoid losing out. A loss of confidencefor political reasons could well have a similareffect, as well as a more direct effect in making theexchange rate even mo re o ver-valued.

In principle, people have the righ t to converttheir local currency to foreign exchange at theman dated e xchang e rate. This is not an immed-iate problem, precisely because the Central

Bank has to hold enough foreign exchange tomeet the resulting demand. However, currencyboard arrangements generally require theCentral Bank only to hold enough foreign

exchange to match the value of base money (ienotes and coins in circulation), and not, forexample, bank deposits. While it would intheory be possible to require the currenc y board

arrangement to op era te against th e total stock ofmoney (including bank deposits), 'this wouldeffectively mean a shutdown of the bankingsystem' (World Bank, 1996c, pi8).

With the narrower system, a general loss ofconfidence domestically would create seriousproblems. If people try to withdraw their bankdeposits and convert them into foreign exchang eas well as their cash holdings, 'the authoritieshave only two options...: they can try to block th edeposit withdrawal or refuse the exchange ofmoney, that is, they can choose between a bank-ing crisis or a foreign exchange crisis' (WorldBank, 1996c, p i 8).

The currency board also means that thecountry will face a binding constraint on publicexp end iture, which can only be financed fromgovernm ent revenues and receipts from dono rs.While the tax base has not collapsed to the extentto which it has in many other post-conflictcoun tries, and the exceptionally high level of theauth oritie s' domestic and foreign liabilities wouldmake further significant borrowing both difficult

and unwise, this will represent an importantlimitation on the flexibility of policy.

Thu s the currency board arra nge me nt raisessome potentially very serious problems.However, from a political perspective, it isprobably inevitable over the medium term,precisely because of the co nstrain ts it impose oneconomic policy. In the cu rren t political circum-stances, a discretionary monetary policy wouldonly be viable if all decisions were mad e jointlyby representatives of all three ethnic grou ps, ona consensus basis. Disagreements would be

inevitable, leading to a very d am aging inertia indecision-making, which would undermineeconomic policy, macroeconomic stability, andconfidence, with potentially serious knock-oneffects. For example, there would be a strongtemptation to resolve inter-Entity (and intra-Federation) disputes over resources bycompetitive inflationary financing. In view ofthe extent of the needs for resources and thevery low level of cooperation, there is a seriousrisk that this would lead ultimately to hyper-inflation, as in the form er Yugoslavia.

It should also be noted that the currencyboard broadly represents a continuation of thepolicies being pursued in both Entities since1994.

33




Dilemma 2: The Central Bank as currency board

The currency board principle is probably the

only way to ensure effective macroeconomicpolicy-making in a decen tralised system with ahigh level of political conflict and inertia.Howev er, it imposes important c onstraints oneconomic policies (specifically fiscal andexchange rate policies); and it raises a seriousrisk of a financial crisis after about 1999. Thelatter could have a major social and politicalimpact, potentially jeopardising the wholeDayton process.

Abandoning the currency board principle isnot a realistic option, as it is mandated for sixyears under the Dayton Agreement, and itwould almost certainly lead to serious prob lemsin macroeconomic policy, either through thetwo Entities printing money competitively tofinance their own programmes (leadingultimately to hyper-inflation), or throughcom plete inertia. Either would lead to a loss ofconfidence, domestically and internationally,which would be very dam aging economically.

Shortening the period of operation of thecurrency board might be more feasible polit-ically, but it could only be achieved, an d wouldonly be viable, if there were a much greaterdegree of mutual trust between the variousparties th an exists at present — or than seemslikely over the next few years. Otherwise, thesame problems would arise. Even discussionsof changing the Central Bank arrangements(which would almost certainly be protracted)could seriously undermine confidence,increas ing the risk of a financial crisis, even in arelatively favourable political climate.

Even ending the currency board arrange-ment after its six-year term could be difficultwithout a substantial improvement in thepolitical climate; and while elections in the

mean time might moderate political positions,

it is at present far from clear that this willhappen. In any event, serious economic riskswill arise long before the currency board's six-year term has elapsed.

In view of these problems, serious andurg en t consideration should be given as to howthe political benefits of the currenc y bo ard canbe retained while minimising the associatedeconomic risks. Two options might meritfurther (and more expert) consideration:

• A self-financing system of im po rt tariffs a nd

export subsidies. This could be used toadjust the effective exchange rate throughits effects on the prices paid for imports andreceived for exports. When an exchangerate depreciation is required, subsidies andtariff's would be increased, averting theproblem of exchange rate over-valuation.However, this would be at odds with thetrade policies currently being recommend-ed, and, in the absence of some form ofexemption, with WTO membership. Thisoption is discussed further in Section D6 .

• A tem porary sup port facility for the Cen tralBank. This could take the form of a fund,financed by donors, to provide balance ofpayments and fiscal supp ort in the e vent of acrisis beyond the authorities' control.Clearly donors would not be willing to writea blank cheque: support should be grad-uated according to need, as determine d bysome previously agreed formula; and thefacility would need to be subject to a fixedceiling. Some conditionality on app rop riate

adjustment measures would also berequ ired in the event of its activation.

34



Economic policies

D.2 Public finances

The financial position of the authorities repres-ents a critical constraint on economic policy;

and this constraint is rigidly enforced by the

currency board principle mandated by the

Dayton A greeme nt (as discussed in Section Dl).

The authorities, taken as a whole, will only be

able to spend what they receive, either fromtaxation or from donors, for at least the ne xt sixyears.

While the tax base has not collapsed to the

same extent as in some other post-conflictcountries (for example, Uganda, Mozambique,and Rwanda), it has nonetheless narrowedsubstantially. The sales tax base had declined by

about 75 per cent from its pre-war level(roughly in line with GD P) by 1995; and pay-rolltaxes have plumm eted due to the reduction in

both formal sector employment and wagelevels. As a result, BH has 'extremely limitedfiscal capacity'. At the same time the prewarfiscal and tax collection system has collapsedand been replaced by different (and incompat-ible) tax systems and separate ad ministrationsin the two Entities (EU/EBRD/World Bank,1996, pl41).

This suggests a need, firstly, to reconcile thetwo tax systems so that they can operate effect-ively within the same country; and secondly, to

increase revenues sufficiently to meet the

country's financial needs.

The development of compatible taxationsystems in the two Entities is problematic in

itself. The RS authorities, in particular, becauseof their separatist aspirations, have everyincentive to ensure that their taxation (and

other) systems are incompatible with those in

the Federation, so as to create an additional

barrier to reunification of the country. Theremay be a similar temptation for the Croatnationalists within the Federation itself.

There is also a need for compatibility in the

structures and rates of taxes, particularlyimport tariffs and the proposed value-addedtax, if BH is to operate as a single nationaleconomy. This suggests that the donors willhave to take a very proactive role, not only in the

development of tax collection and administra-tion systems, but also in the details of tax

policies, to ensure their compatibility.

Strengthening the public sector finances willbe com plicated by two major tax reductions (or

constraints) being pressed by donors for reasonsof econo mic efficiency: on imp ort tariffs; and on

wage-related taxes and social security contribu-tions. (These recomm endation s are discussed in

Sections D4 and D5 respectively.) In 1995, thesetwo sources represen ted mo re than one-third of

the authorities' total revenu es in the Federation(World Bank, 1996b, Table 3.1).

Because of the currency board arrange me nt,any reduction in revenues from these sourceswill feed through directly either into lowergovernment expenditure or into higher rates of

other taxes. It has been suggested (EU/WorldBank, 1996b, p20) that the cut in payroll contri-butions should be compensated by increases in

excise and sales taxes. However, these are

already very high: in the Federation, theyrepresent around 16 per cent of GDP (World

Bank, 1996b,Table 3.1); and in RS, the generalrate for the goods and services tax (with someexemptions and reductions) is already 31 per

cent (RS, 1996, pi 13). This suggests that therewill be little if any scope for further increases,and that the main burde n of the adjustment willfall on public expend iture.

The trade-off between tax rates and govern-ment expenditure will be particularly starkbecause both customs duties and payroll taxesand contributions are ear-marked for specific

uses. Customs duties are to be used to financethe central government. Reducing importtariffs could thus leave a shortfall in financingfor the central governm ent (unless it is fully off-

set by a reduction in external debt-servicepayments); and this could substantially weakenthe forces for political cohesion. Pay-roll taxesand contributions are the sole source of financ-ing for pensions, benefits, and health services.Any limit on the rates of such taxes wouldtherefore also constitute a limit on social safety-nets. This will become still more problematic

over the longer term , as the proposed transitionfrom a pay-as-you-go pension system to a

system based on individua l acco unts will entail asubstantial increase in costs over a transitionperiod of perhaps 40 years. (See Sections El

an d E3.)

Over-tight fiscal constraints represent an

important obstacle to effective policy-making,efficient administration, and the provision of

high quality public services. A particularproblem in this contex t is that of public sectorsalaries: donors have shown themselves willing

to finance shortfalls in other types of recurrentspend ing, such as money for essential drug s and

other consumables in the health sector, and for

teaching and learning materials; but they are

35




not generally willing to contribute towards w agecosts. As a result, the tight fiscal constraints inBH limit both the numbers and the quality ofstaff; and this represents a potential bottleneck

in project implementation and policy-making:

While most of the Bosnian implementing agencies

have competent staff, their numbe r is very limited,

and their capacity to handle a large number of

projects may be stretched to the limit. Furthermore,

there is compe tition for highly-qualified people

among foreign institutions located in Bosnia and a

growing demand coming from the private sector,

resulting in a salary structure that makes it hard to

keep staff in the public sector.

(EU/World Bank, 1996a, p6)

The World Bank's approach to public-sectorsalaries has been based primarily on raisingsalaries at the Fede ration level into line with tho sein the Croat areas (and at die national level toabout 25 per cent higher), so as to attract anethnically balanced civil service. Th is is an essen-tial objective; but it still leaves some pro blem s. Inparticular, the financial support for this policytotals only $5m, compared with an annual addi-tional cost of $9m pa.19 This leaves die a uthoritiesto find an extra $6.5m pa until March 1998, and

$9m pa diereafter. Also, no support is planned forthe cantons or municipalities; and no account istaken of salary levels available in die priva te sectoror from do no r agencies and NGOs themselves.

It is interesting to compare the level ofsup po rt for salary supplem ents with the cost ofproviding consultants under the EmergencyReconstruction Credit. The former is $5m, tosupport 2,520 staff over the two years of theproject's duration — an average of about $ 1,000per person-year. Almost the same amount($4.53m) is allocated for some 26 person-years

of consultancy services — an average of$174,000 per person-year (World Bank, 1996a,Annex I Appendix 5). In other words, theexternally-financed salary supplement for theaverage civil servant for a year is enough tofinance the average consultant for about oneand a half working days.

An othe r potential problem is the distributionof reven ues between levels of gov ernm ent. Th emost sensitive aspect of this problem is therequirement of contributions from the Entitiesto the n ational gov ernm ent. This is essential to

the operation of the BH government, which hasno independent source of revenue (until suchtime as the national parliament approves taxesat the national level); but, as the World Bank

(1996 b, p34) points out, the system 'clearly ru nsthe risk of vulnerability to the transfer frombelow no t being forthcom ing'. This problem willbe the more acute if, as one World Bank study

sugg ests, 'Tax rates will need to be hig he r in theSerb Republic than in the Federation to generatethe respective revenue contributions because ofthe gre ater aggregate population and incomes inthe Federation' (Foxand Wallich, 1997, pi2).-°

The implications of this vulnerability areextremely serious: 'The state's sustainability is aconcern in this environmen t, wherea strong degreeof distru st ex ists' (Fox and Wallich, 1997, p 11).

RS, like the Federation, itself faces a tightfinancial constraint. At the same time, the RS

autho rities have no real political com mitmen t tothe BH government: given the choice, theywould prefer autonomy or unification with theFRY. This provides them with little incentive todivert resources from their own spending prior-ities to finance the national government. In1996, the RS authorities made no contributionto the national government, and their initialbudget was sent back for adjustment by theParliam ent because it envisaged exp end iture 50per cent above anticipated revenues, with nomeans of financing the deficit. Further

spending reduction was required because of a15 per cent shortfall in revenues (Fox andWallich, 1997, pi 6) .

There are signs that the RS authoritiesrem ain very reluctant to fund the national gov-ernm ent; and, while it is possible that this is onlypolitical posturing, this can by no means beassumed. Overcoming this reluctance may wellrequire heavy-handed donor conditionality;an d to som e extent this will be automatic, since aviable national budget would be an essentialprecond ition for an IMF programm e, and thus

for debt reduction (see Section B2), quite apartfrom m ore direct dono r conditionality.

However, it is not clear that this form ofconditionality will be effective; in fact, therewould appear to be a risk that it could proveseriously counter-productive. As noted earlier,it is possible that th e RS authorities w ould see apolitical advantage in derailing the debt reduc-tion process, and delaying the IMF programmeuntil after June could contribute significantly tothis objective. This in turn would require nomore than procrastination on their commit-m en t to funding the national governm ent. Moredirect a nd explicit (and possibly heavy -handed )donor pressure may therefore be required tosecure a favourable outcom e.

36



Economic policies

A second issue in RS is that of transfers fromthe Entity level (where most revenues arecollected) to the municipalities. While Serbnationalists dominate the Entity authorities, the

dependence of municipalities on such transfersprovides an opportunity to manipulate thesystem to penalise m ore mod erate autho rities atthe local level. The re are signs that this is alreadyoccurring: the authorities in Banja Luka (thecentre of the more moderate political forces inRS) have 'argued strongly that [Banja Luka] isreceiving less than its legislated share of taxes'(Fox and W allich, 1997, pl 6) .

A related issue is the direct or indirect transferof resources between administrative units at thesame level (for ex am ple, between the Entities o rbetween cantons within the Federation). This iscritically important because of the high level ofdecentralisation and the ethnic base of admin-istrative units. As the W orld B ank has o bserved:

as the failure of the Former Yugoslavia demon -

strates, a decentralized structure is only viable if

inter-Entity and cantonal links are mutually

beneficial. It is essential...that there no t be too much

cross-subsidization.

(World Bank, 1996b, p xx)

Th e problem is that there are marked differ-ence s in the levels of economic activity, incom esan d e mploy men t between different areas of thecountry; and these differences are, to a great

extent polarised along ethnic lines. Most Croatarea s escaped b oth the worst of the dire ct effectsboth of the war (which afflict the Muslim areas)and of the UN sanctions (which have severelyaffected the RS economy); and they have receivedsubstantial support from Croatia. As a result,these areas have substantially higher livingstandards than the rest of the country (Table 6).If, as seem s likely, the effects of UN san ction s onRS can be reversed more quickly than thedestruction and disruption resulting from thewar in Bosniac areas, a similar gap could openup between these two regions.

In the absence of transfers between the variousadministrative units, the very tight fiscal con-straints imposed by the currency board limitspen ding within each Entity, canton, and munici-pality to the amount of revenue which can beraised by its authorities. This would allow a m uchhigher level of spending (and thus of socialprovision, and support to production) and/orlower tax rates in richer areas, while imposing

Table 5:

Total production(% of pre-war)

Industrial production

(% of pre-war)

Total unemployment (%)

Non-agriculturalunemp loyment (%)

Average net wage (DM pm)

Per capita income ($, 1995 )

Regional Economic

Source

1

2

1

2

2

3

Croatareas

85%

n/a

n/a

i

330

1,800

Sources: 1 : EU/EBRD/World Bank (1996 ), p77 .

2: World Bank, 1996b, p51 (except average wage in RS:3: Fox and Wallich (1997), p3.

Indicators, 19 96

Bosniacareas

5-10%

15-20%

50-60%

i

50%

160

500

unofficial IMF estimate).

NB data from different sources may not be directly comparable.

RS

n/a

8-10%

n/a

60%

70

1,000

37




very harsh financial constraints and/or highertaxation in poorer areas. The result would be anincreasing polarisation between the areas inhab-ited by the three ethnic groups. This would both

increase poverty (as income growth would be

lower in the areas with the higher rates of pov-erty), and widen the political gaps between theethnic group s (since richer areas seldom want theencumbrance of poorer areas within the same

country).

Dilemma 3: Inter-regional inequality and transfers

In principle, the extreme inequality betweenthe different ethnic regions of BH suggests aneed for some inter-regional redistribution ofresources, both between the Entities andwithin the Federation. However, political

considerations, combined with binding fiscalconstraints, impose a very tight limit on thescope for any such redistribu tion.

Given the level of inter-ethnic politicaltensions, any substantial transfer of resourcesbetween administrative units of differentethn ic compo sition co uld give rise to consider-able resen tmen t. As the W orld Bank (1996b, pxxv) has noted, 'tolerance for interregionalsubsidies is low, if not nil'. Once again, thisprob lem is exacerbated by the currency boa rd.

Since the public accounts need to be balancedoverall, a deficit in one area is possible only ifthere is a corresponding surplus in anotherarea. This makes the transfer much moreapparent than if there were merely differentlevels of deficits. Any country with a fiscalsurplus would immediately know that it waseffectively subsidising those areas w ith deficits;and, if they were areas with a different ethnicbase, this would give considerable amm unitionto political extremists.

O n e way of circum venting this problem is by

adjusting the geographical distribution of aidin accordan ce with local economic cond itions.This is already happening in terms of thedistribution of aid between Bosniac and C roatareas; but the exceptionally low levels ofassistance to RS are seriously at odds with thisobjective. (See Fig ure 2 in Section C3.)

However, there are three importantlimitations to this so lution:

• Correcting regional imbalances is not theonly criterion for th e geographical distribu-

tion of aid. In p artic ular , it may conflict with

wider political objectives, such as applyingpolitical condition ality to secu re the level ofcooperation necessary to the smooth func-tioning of institutions. This means, firstly,that aid to each eth nic area mu st be sufficient

to provide an incentive for compliance, inorder to avoid losing it;21 and secondly, thatit may be nece ssary to w ithhold aid from lowincome are as as a mea ns of applying politicalpre ssu re (as in RS in 1996).

More generally, an explicit geographicaltargeting of aid, largely corresponding withethnic bou ndarie s, could m erely transfer partof the burd en of resentment from other ethnicgroups towards the d onors. Th ey could cometo be seen m ore explicitly as being edinically

partisan, rath er tha n m erely as partisan in termsof their pre ferre d political outcom e; and thiscould further undermine their position inseeking to influence the political process.

Aid is concentrated very heavily on capitalexpenditure rather than recurrent spending;and what support there is for recurrentexpenditure is selective, missing a numberof critical areas, such as wage costs. Capitalspending for reconstruction and rehabilita-tion is helpful to the local economy, gener-

ating employment and providing the facilitiesfor renewed provision of essential services;but if inad equ ate resources a re available forrecurrent spending, poorer areas will remainat a distinct disadvan tage, and there may beserious inadeq uacies in essential pub lic servicesand social provision. Over-compression ofrecurre nt spen ding may also unde rmine theviability of investment and reconstructionprogrammes and weaken administrativecapacity, due to inadequacies in staffinglevels, wages, other necessary inputs and

maintenance.

38



Economic policies

Dilemma 3: Inter-regional inequality and transfers (continued)

On e possible approa ch to this problem w ouldbe for do nors to establish a fund for su ppo rt ofrecu rrent expe nditure, on a regional basis. Thiscould be done using anticipated balance-of-

payments support, for example, under WorldBank adjustment loans. Resentment againstdo no rs could be limited by basing the geo grap-hical distribution of funds clearly on an explicitformula measuring needs22 (for example,providing funds in direct proportion to theshortfall of local income in the jurisdic tion of

each adm inistrative unit relative to the n ationalaverage).

This would not eliminate potential politicalproblem s. It would require the initial appro valof all pa rties, which would be difficult to se curewithout additional aid flows; it would meanskewing funds strongly away from Croat areas;and some areas would inevitably receive lessthan they are currently expecting. However,the use of a transparent formula based onincom e levels would at least mak e the ratio naleof the process explicit, and the use of aid flowsrather than domestic resources would make

th e trade-off" between the interests of differentareas somewhat less acute.

D.3 Privatisation and thesettlement of domestic liabilities

It is currently proposed that there should be arapid and wholesale privatisation of the econ-omy. This idea, and the outline of the ap proachto privatisation, appears to originate with theWorld Bank, appearing in their first post-warpaper on BH (World Bank, 1995b, ChapterIV).

As elsewhere, the recommendation forprivatisation is partly m otivated by the view thatit will lead to a more efficient industrial sector,promoting faster economic recovery, whilelimiting the cost to the public sector. In partic-ular, coupled with efforts to promote thedevelopment of new enterprises (see SectionD4), it is intended to avoid any tendency torecreate the pre-war industrial sector. Thisobjective is entirely a pp rop riate . (See Box 2.)

In the Federation, at least, there is also asecond motivation for privatisation: to providethe means for the authorities to settle thedom estic liabilities accum ulated du rin g the war.This linkage is im por tant: since it is unlikely tha tthe proceeds even of the full privatisationpr og ram m e will be sufficient to settle all of theliabilities in full, any redu ction in its scale would

need to be matched by a reduction in thepayments to holders of such liabilities; and, inview of the nature and spread of the liabilities(which include a rrears on wages and pensions,

and frozen foreign exchange deposits held by80 per cent of the population), failure to settlethem could have high social and political costs.This problem does not arise in RS, where theauth oritie s claim to have no such liabilities.

It is prop osed that state companies should selloff functioning assets (eg vehicles and otherequipment) as quickly as possible, retaining theprocee ds, before the en terprises themselves areprivatised. The companies would also hive offnon-viable operations (which would beliquidated) and assets not related to their coreactivities, such as wo rkers' housing, which wouldbe sold w here possible. Th e large conglom erateswould then be split into their compo nent parts,which would be privatised. The intention is thatsmall-scale privatisation should begin during

the first half of 1997, and large-scale privatisa-tion du ring the second half; and th at at least onebank should be privatised by the end of the year(OHR/EU/UST, 1997).

Th ere is a high level of decen tralisation in theproposed approach to privatisation, at leastwithin th e Fed eration. Each enterprise (or com-ponent of a conglomerate) is to be sold by thecanton or g roup of cantons in which it ope rates.This is largely intended to preserve the ethnicidentities of enterprises, so as to pre vent resent-ments arising from ownership transfersbetween ethnic groups, or the operation ofenterprises by different ethnic groups fromthose in the areas where they operate.

39




On the buyers' side, the proposal is thatoutstanding liabilities (for example, salary andpension arrears, restitutions for war damage,and frozen foreign exchange deposits) should be

settled by issuing their holders with vouchers.Th ere has been some suggestion of an additionalissue of vouchers to the population at large, aselsewhere in Eastern E ur op e; but the scale of theliabilities (estimated at more than DM]3bn23)relative to the likely proceeds of privatisation

suggests that this is unlikely to be viable. Thevouchers could the n b e used, together with cash,to buy privatised assets, including housing , andcompany assets such as vehicles and equipm ent,

as well as shares in privatised companies. Insome cases, priority may be given to bids whichinclude a cash component, or full payment incash may be req uired .

There is a great deal of urgency in therecommendations on privatisation. Wherever it

Box 2: Prospects for the industrial sector

Even before the war, ther e w ere problems in the industrial sector, in terms of whatwas producedand of how it was produ ced; moreover, the economic environ men t faced by the country has gonethrough some importan t changes over recent years(as discussed in Section B1), quite ap art fromthe effects of the w ar itself.

Th e p roblems of the p rew ar ind ustries extend beyond conside rations of efficiency, an d applyto a num ber of major ind ustries:

• Th e coal reserves are of relatively low quality, with a high sul ph ur con tent.

• Electricity production is based substantially on burning this coal, and is therefore verypolluting.

• The armaments industry has been particularly affected by the war, both through being

targeted for strategic reasons, and because much of the po rtable eq uipm entwas transferred toSerbia or Croatia by the resp ective com batants. Many do no rs w ould find it politically difficultto promote the rehabilitation of this sector; and direct financial support would be legallyunten able in many cases.

• Th e long-term competitiveness of the textiles sector may be questiona ble. While there hasbeen a significant recov ery in this sector, reportedly sup por ted in par t by foreign investment,this is based in p art o n the cu rren t very low level of wages. Inter natio nal competition in thetextiles sector is likely to become increasingly intense, as low-income countries seek to prom oteexport-oriented textile prod uction, and as the Multi-Fibre Arran gem ent is phased out. As realwages rise, BH is likely to beco me a progressively less attractive location for foot-loose textileoperations.

More viable sectors include light engineering and wood products. However, there will beintense competition in these sectors from other Eastern European countries. To prosper, itwould be necessary to get ahead quickly in these sectors, and the damage and disruptionresulting from the war will be a serious obstacle in this ende av ou r. If, as seems inevitable, othe rEastern Euro pean coun tries prec ede BH in acceding to the EU , this will be a further set-back.

A mo re general prob lem , affecting all indu strial sectors, is the d isplac em ent of skilled workers,both externally as refugees and internally (especially between RS and the Federation andbetween Muslim and Croat areas). While this problem may be eased in the long term, it will besome time before the conditions for a general return are created; and the likely scope of anyultimate return remains unclear. Meanwhile, the previous pools of skilled labour have beendispersed and workers with specialised skills separated from the areas where they can be mostproductively employed.

40



Economic p olicies

is proposed, the emphasis is always oncom pleting the process as quickly as possible. Thereason advanced is that privatisation will benecessary for a resump tion of production by theen terp rise s co ncern ed (which will itself be essen-tial to recovery, given the ir impo rtance), since thegovernment does not have the resources tofinance their rehabilitation.

The need for rapid privatisation is generallysup por ted by a strong emphasis on the severityof the en terprises' plight:

Most of these conglomerates were...heavily

damaged during the conflict.(EU/World Bank, 19 96b, p8)

The prewar conglomerates were especially hard hit

by the war.... Most of them were split between the

Entities, their industrial facilities have been

damaged, and their managem ent methods are

inappropriate in the ne w market economy.

(EU/EBRD/World Bank, 1996, Box 3, p82)

However, the speed at which privatisation isexpected to take place has some severe disad-vantages:

• Selling off assets an d enterp rises quickly willgreatly reduce the proceeds. In the current

situation , w here the economy is still at a verylow ebb, domestic d eman d is weaker th an it islikely to be in a few years. This will sub-stantially re duc e th e profitability of the ente r-prises. At the sam e time, incomes are low an dsavings in m any cases dep leted. Th is will limitthe extent of cash contributions to purchasingassets.

• Where cash contributions are required orpreferred, rapid privatisation will skewparticipation towards those with money. InMuslim areas, at least, these broadly includereturnees (who are, in some areas at least,resented by those who stayed to suffer theprivations of the war); war profiteers andcriminals (who are resented even m ore); andemployees of international agencies (whomay be seen as profiting further from thepolicies of their emp loyers).

• Selling off functioning com pany assets (unlessthey are genuinely surplus to requirements)is likely to redu ce the value of the companiesthemselves, since they will ultimately have to

be rep laced , almost certainly at higher cost.

It is also unclea r th at the justification pu tforward for rapid privatisation is wholly valid.

The strong statements of the plight of socially-owned conglomerates is at odds with assess-ments made elsewhere — sometimes elsewherein the same documents — where a m ore positivepicture is presented, for exam ple as evidence of'indic ator s of a revival of the ind ustrial sector'.

Only a small proportion of the overall industrial

capacity has been damaged, an d some regions have

been almost completely spared (Banja Luka and

southern Herzegovina, for example)....[MJany

plants are still relatively well-equipped with

serviceable machinery capable of producing and

marketing good quality products.

(EU/EBRD/World Bank, 1996, Box 1, p78)

Equally, the financial constraints on thegovernment's ability to rehabilitate socially-owned enterprises arise primarily because thedonors are unwilling to provide support for thispurpose. This, in turn, is presumably becausethey have a preference for the rapid privatisationoption.

Confidential survey evidence for the donors'Industry Task-Force suggests a picture of thesituation of socially-owned conglomeratessomewhere between the two extremes quotedabove. Of the 50 largest socially-owned enter-

prises in the Federation (ie those 'especially h ardhit' by the w ar, in the E ntity with the mo re severewar dam age), enterprises accounting for 35 pe rcent of pre-war employment had 70 per cent ormore of their pre-war capacity still operable in1996 — in some cases more than 95 per cent.Tho se accounting fora further 45 per cent of pre-war employm ent had between 40 per cent and 65per cent of their previous capacity.

A greater problem than the damage to pre-warcapacity is und er-utilisation of the capacity whichremains. Of die 50 enterprises, 20 (accounting

for 60 per cent of pre-war employment) wereope rating at 20 per cent or less of their presentcapacity, while only n ine (5.5 per cen t of pre-waremployment) had capacity utilisation of morethan 50 per cent. Only two enterprises, repres-enting just 1 per cent of pre-war employment,were operating at more than 75 per cent ofcapacity.

In many cases, the capital investmentreq uir ed to achieve full u tilisation of the rem ain-ing capacity, or the restoration of pre-warcapacity, is very limited. By providing capital

selectively to socially-owned enterprises to thisend only where the cost is DM 11,000 ($6,500) o rless per job, it would be possible to generatenearly 45,000 jobs at a total cost of less than

41




DM200m ($120m). This represents a one-timecapital cost of DM4,350 ($2,570) per job — lessthan half the cost per person-year of employ-ment created by the Emergency Public Works

programme.24 T he scale of the ben efits could beincreased further by expanding the scheme tocover smaller enterprises and those in RS.

At present, ortho dox commercial lending isvery limited, almost entirely short-term, andextremely expensive, so the main sources offinancing for the industrial sector are variouscredit schemes being sup ported by donors suchas the World Bank a nd USAID. However, theseschemes are exclusively for small and mediumenterprises and focused heavily on the privatesector, and a re the refore likely to exclude mostof the above investments. Socially owned enter-prises are eligible for support under the WorldBank programme only if they are 'have goodprivatisation poten tial'.

To the extent that the larger socially-ownedenterprises a re exclud ed from access to these (orother affordable) funds, the result will be toimpede their rehabilitation, by limiting theirability to invest; to force them into moreexpensive sources of funding, damaging theirlong-term financial position; and to force them

to sell assets which may be needed for theirfuture operation, or which might be sold inmore favourable market conditions later. Theoverall effect will be to weaken their financialposition, possibly to jeo pa rd ise the ir viability,and almost certainly to red uce the procee ds if orwhen they are ultimately privatised.

As well as the pace of privatisation, the mech-anisms proposed also raise some potentiallyserious problems:

• Sale to domestic buy ers thro ug h the voucher

system will no t pro vide the capital necessary tofinance rehabilitation. Enterpriseswill be under-capitalised and critically dependent on theirability to borrow; but the availability ofcommercial lending is likely to remainrelatively limited over the m edium term . Thiscould represent a significant obstacle to econ-omic recovery.

• Because of the war, enterprises have fallenseveral years behind in technology comparedwith foreign com petitors; and for some indus-tries, catching up in technology will be essen-tial to regaining their competitive position.However, selling enterprises to domesticinvestors is unlikely to prov ide the technology

transfer which could be gained in the case offoreign investors. There is thus a risk ofcreating an economy based on technologicallybackward and under-capitalised companies,

which would substantially weaken BH's long-term prospects.

• The weak financial position of most of thepopulation would give rise to considerablepress ure to sell shares , raising the risk that themarket would be flooded and prices wouldplum me t. Once ag ain, those with cash wouldbe able to buy up assets very cheaply, skewingecono mic po we r to a relatively small elite, andgreatly reducing the benefits to the bulk ofthe population (who might understandably

feel hard don e by).

• Th e decen tralised basis of privatisation will en-trench the ethn ic division of ownership, imply-ing a strong p olarisation of ownership,with majorenterp rises clearly divided along ethnic lines.

Politically, there was initially some degree ofambivalence towards privatisation on the part ofthe RS authorities. The 1996 policy statement,for example, seems somewhat at odds with theobjectives of rapid and near-universal privatisa-

tion and th e operatio n of market principles:In the sector of the econom y with m ajority state

ownership (state corporations), the state will

determine prices. The essential principle gove rning

this...is 'cost plus'....The role of the state is to 'iron

out' losses an d income between state corporations.

(RS, 1996, pp 120-1)

However, this now appears to have beenresolved, and there seems to be a reasonabledegree of commitment to privatisation —although the de pth of such commitment may be

open to question.The privatisation process envisaged in RS is

very different from that in the Federation.Un der the Privatisation Act of July 1996, 55 pe rcent of share s in enter pris es to be privatised willbe transferred to six social funds,25 which willultimately be transformed into stock-holdingcompanies, but will not participate in theirman ageme nt; 30 pe r cent of shareswill be distrib-uted free to the adult population, through avoucher system; and the remaining 15 per centof shares will either be offered for sale on the

capital market, or be sold to strategic investorssubject to conditions relating to businessactivities, investment, e mp loym ent, an d so on.

42



Economic policies

Dilemma 4: How to privatise in the Federation?

In the absence of strong domestic opposition,there is little reason to oppose the principle ofprivatisation for non-utility com panies. Largestate-owned companies are present through-out the productive sectors, in activities forwhich there is no strong argument for stateinvolvement. However, the questions of howand whe n to privatise are more open to debate.

Domestic versus foreign investment: T h eproposed sale of enterprises to domesticinvestors has some advantages: it keeps thecontrol and profits of the enterprises in local

hands (at least in the short term), and avoidsthe potentially high long-term foreignexcha nge cost of foreign investment. H oweve r,it will limit the proceeds of privatisation, thecapital available to enterprises for rehabilita-tion, and the technological advantages of priv-atisation. The proceeds of privatisation areimp ortan t, as a limited settlement of outstand -ing claims on the public sector will have asignificant social and political impact.26

Rapid versus gradual privatisation: A veryrapid privatisation process is proposed as a

mean s of getting the economy moving quickly.Ho wev er, this will further limit the proc eed s ofprivatisa tion: in view of th e low level of econ-omic activity at present, and the depletion ofsavings, peop le will be less able to pay for stakesin privatised enterprises than m ightbe the casein the future. The rapid approach will alsoskew the benefits of privatisation away fromthose w ho fared worst in the war, while a mor egradual a pproach could providea wider ethnicownership base if freedom of movement andresidence were reestablished.

Dece ntralised versus national approaches:Th e dece ntralised privatisation process shouldavoid dispu tes and r edu ce conflict,by ensuringthat each enterprise will be owned by themajority ethnic group in the area in which it islocated. However, the polarisation of majorcompanies along ethnic lines (rather thanhaving a broad mix of ownership) is moreconducive to a separatist than a united futurefor BH.

A possible alternative ap proach would be asfollows:

• The claims on the authorities could bereduced to some extent by limiting the

immediate settlement of claims for frozenforeign exchange deposits, either to thosebelow a specified income level, or to a maxi-mum absolute amount. Settlement of theremainder would be contingent on a settle-ment of the corresponding claims on theFRY, which should be backed by theinternational community.

• Holders of the remaining public sectorliabilities cou ld b e issued with vouch ers in theform of tradeable financial instruments —either indexed instruments widi no interest,

or perpetuals27 — if possible guaranteed (bydonors) against political risk. This wouldallow, ho lde rs of go vernm ent liabilities toreceive compe nsation immediately (althoughprobably at much less than face value, sincethere would be substantial financial pressureon holders to sell in the early stages). Novouchers should be issued to those not hold-ing claims on the public sector.

• Potentially viable socially-owned enterp riseswhich can achieve full utilisation of their

existing capacity at relatively low cost per jobcould be reactivated immediately, using afund established with donor funds, beforetheir privatisation was considered. Otherscould be considered for privatisationimmediately.

• Within each of these categories, enterprisescould be further categorised according totheir nee d for capital and technology inputs.A 51 per cent stake in those enterprisesneeding substantial inputs could be offered

for sale to foreign investors. Either the salecould be conditional on the investor fullyfinancing the rehabilitation, this beingreflected in the price; or the investor'scontribution could be used to settle thegovernm ent's (49 per cent) share of the costsof rehabilitation.

• Th e rem ainin g 49 per cent stakes in theseenterprises, together with those enterprisesnot needing substantial capital or technol-ogy inputs (eg lower-technology industrieswhich have sustained more limited war-

damage) could be sold through the vouchersystem w hen the conditions were a pprop riate

43




Dilemma 4: How to privatise in the Federation? (continued)

in terms of income levels, confidence, etc.

This could (but need not necessarily) bedone on a national basis for those enter-prises with 51 pe r cent foreign own ership, asno ethnic group would be able to secure acontrolling interest. If a decentralisedapproach were used, it would be preferableto wait until substantial resettlement ofminorities had occu rred, or un til itwas clearthat this would not happen within areasonable time-frame. Cash paymentsshould be optional, and reflect the marketvalue of vouche rs.

The implications of this approach for theproceeds of privatisation (relative to govern-ment liabilities) are ambiguous. The sale ofstakes in som e enterp rises to foreign investors,and the delayed sale of 49 per cent stakes,would increase the prices received; and thelimitation of foreign exchange deposit liabil-ities would help. However, the indexation or

paym ent of interest on vouc hers, and the use of

part of the proceeds to finance rehabilitationwould act in the opposite direction. It istherefore possible that some liabilities wouldremain (or that a greater volume of liabilitieswould remain than under the currentproposals) after completion of the privatisationprocess. These could remain as (or, if initiallyindexed, be converted to or refinanced with)perpetual notes, to be redeem ed by the govern-ment w hen resources were available. However,this could be a significant financial burden onthe government, which would be reflected in

other components of government spending(with a direct one-to-one relation ship while thecurrency board is in operation).

It should also be noted that the skewing ofbenefits to those curren tly in a relatively s tron gfinancial position would not be fully resolved,as they would almost certainly be able topurch ase vouch ers at less than their face-valuein the short term .

However, the World Bank does not agreewith this appro ach . Th e social funds a re seen astoo passive; individual investors are seen as toosmall to take an active interest in the manage-ment of firms; and the stake to be offered tostrategic investors is seen as inadequate toprovide them with an incentive to overcome theresulting inertia. In addition, the conditionalsale of strategic stakes is seen as unnecessarily

restrictive, unenforceable, an d non-transp arent.Th e Bank wou ld like to see the privatisation

of the social funds, conso lidation of their claimsin individual enterprises, and the developmentof a more active role in enter prise man agem ent;the abandonment of conditions on strategicinvestments; and a greater role for foreigninvestment (EU/World B ank, 1996b, p8).

These differences are probably not insuper-able, since the RS position appears to benegotiable. Howeve r, the p rocess of negotiationand the passage of new legislation is likely to

delay the process somew hat.Th e pace of privatisation is again an issue. As

in the Federation (if not more so), the currentweak economic position will limit the proceeds

from a rapid privatisation. Moreover, a slowerpace of privatisation is arguably more viable inRS than in the Federation, because the problem sof state-owned enterprises arise primarily fromeconomic sanctions (which hav e now been lifted)rather than from war-damage. This suggeststhat rehab ilitation costs (which arise mainly fromlack of maintenance) should be relativelylimited, so that the financial constraints on the

government will not be such an importantlimiting factor; an d tha t the financ ial situation ofthe enterprises (and thus th e potential revenuesfrom privatisation) should improve fairlyquickly as recovery prog resse s.

Th ere is also a significant political d imen sionto this issue. Th e exclusive focus of the privatisa-tion programme within RS at the Entity levelsuggests a consolidation of RS as a self-contain-ed economic unit. With an immediate 100 percent privatisation, ownership of each e nterp risecan be expected to extend , to a greater or lesser

extent, throughout RS, without crossing theboundary into the Federation, or therefore tonon-Serb investors. Once again, this implies a steptowards separation rathe r than reunification.

44



Economic policies

This effect could be moderated by a phased(51 per cent-49 per cent) approach, to becompleted on a BH-wide basis if and when thepolitical circumstances permit. The govern-

me nt's sha re of rehab ilitation costs could readilybe financed from the procee ds of the 51 per centsell-off, since (according to the autho rities) thereare no competing government liabilities whichneed to be settled.

However, there are two critical obstacles tothis option:

• Since the auth oritie s' objective is separationfrom the Fed eration, they canno t be expectedto ado pt a policy aimed explicitly at reunifica-tion.

• In the absence of outstanding liabilities, theproceeds from the 15 per cent stake sold onthe market or to strategic investors willcontribute directly to government spending.In view of the high expenditure needs, therelatively weak revenue base, and the tightfiscal constraint imposed by the currencyboard, the authorities might well prefer toreceive lower proceeds now rather thanhigher proceed s in the future. This incentiveproblem is much more acute because the

second instalm ent would be conditional on adegree of national reconciliation which couldalmost certainly not occur until the currentRS autho rities had lost power.

D.4 Private sector development

The long-term development strategy envisagedfor Bosnia is based on development of theprivate sector. This entails three broad compon-ents:

• privatisation of state-owned enterprises (asdiscussed in Section D3);

• prom otion of small and med ium en terprises(SMEs);

• rehab ilitation and privatisation of theban kin g system (as discussed in Section D5).

The encouragement of SMEs is partly direct,through the provision by donors of lines ofcredit (World Bank and USAID) and equityfunds (EBRD) in the Federation. In addition,the intentio n is to provide a favourable env iron-

ment for private sector development throughderegulation, particularly of the labour market(as discussed in Section D6). Rehabilitation andprivatisation of the ban king sector is also a pa rt

of this process, intended to allow the develop-ment of domestic sources of financing over themedium term.

SME development is seen as a major s ource of

employment generation, together with therehabilitation and privatisation of public enter-prises. Th e p rom otion of SMEs is probably wise:while many of the management and financialskills required for a thriving SME sector arecurren tly lacking, the re was a significant level ofsmall-scale private s ector activity before the war;and in Sarajevo, at least, the SME sector hasalready recovered significantly. The manage-men t skills requ ired shou ld be developed in thefuture with the help of donor-supportedtraining courses, altho ugh the process may take

some time.

How ever, ther e ar e som e caveats to the likelybenefits of SME development. Firstly, thereappears to be a strong sectoral bias in SMEdevelopment so far, at least in the Federation,towards trading, retailing, and other servicessuch as restaurants and transport, with muchless activity in the productive sectors —especially the production of internationallytradeable goods. This may chan ge over time, assome service-providers are reportedly

diversifying into the prod uctio n of goods; but itis also likely to b e a slow p rocess.

An SME sector based too heavily on non-tradeab le services raises two imp ortan t problem s:

• It will be critically de pe nd en t on the state ofthe domestic economy; and if declining aidflows, coupled with the currency boardsystem, lead to a severe recession w ithin thre eto five years (as sugge sted in Section D l) , thesector could be disp rop ortio nate ly affected,with knock-on effects on employment.

• It implies a substantial negative impact ofenterprise development on the balance ofpayments: within the Federation, firmsimport 60 per cent of their inpu ts, but ex portonly 3 per cent of their output28

(EU/EBRD/World Bank, 1996, p81).

Secondly, at least some of the SME fundsseem likely to be unduly biased in favour of thebetter-off. The requ ireme nt of the World Bankfund that an enterprise should have beenop eratin g for at least two years before receivingcredit limits its use to those who have the

resources needed to sustain an enterprise fortwo years witho ut credit.29

Thirdly, there would appe ar to be a dange r ofa regional and ethnic bias in the support of

45




SMEs. This could arise both directly andindirectly. The focus of the SME sup po rt fundsat present is strongly, if not exclusively, on theFederation; and some funds have an explicit

regional focus even within the Federation, forexample, the USAID programme in the(Bosniac) Tuzla area. The World Bank's two-year rule limits access to funds in areas wherethe private sector was worst affected during thewar, most notably the enclaves and areas ofconflict, where most enterprises are likely tohave been established more recently; and inareas where the economy remains mostdepressed, so that few people have their ownresources to sustain enterp rises.

Overall, while SME development is approp-riate as an element of employment genera tion,greater consideration should be given to whattype of enterprises are being su ppo rted, in w hatareas, and who owns them. It is also importantto avoid over-optimism: the potential of thissector to absorb labour is likely to remainrelatively limited, at least over the mediumterm; and there is a real risk that many of thejobs created will be lost later if the sector isseverely affected by the macroeconomic effectsof declining aid flows, especially if this coincid es

with the period when credits are due to berepaid.

D.5 Financial sectorrehabilitation

The banking system faces extremely acuteproblems; and this constitutes a major cons traintto the rehab ilitation of socially-owned ente rpr ises ,the development of the private sector, andeconomic recovery. An estimated 90-95 per centof the assets of the banking sector are non-performing, and the sector is divided betweenlarge socially-owned banks virtually paralysed bythis encumbrance, and small under-capitalisedprivate banks with very limited experience offinancial intermediation. Th e overall result is thatvery litde commercial lending is available, andwhat there is is short-term and at very high realinterest rates (2-6 per cent per m onth ).

In view of the private-sector orien tation of thecurrent economic strategy, this makes banking

sector rehabilitation a high p riority. This is to bebased on five main e lements:

• the rehabilitation and restructuring of thelarge state-owned banks, including relieving

them of their existing overhang of bad debts,with a view to puttin g them on a sou nd financialfooting and privatising them whe re possible, orliquidating them where the situation is

irredeemable;

• providing technical assistance and othersup port to existing small private ban ks;

• 'supporting the development and growth ofnew private banks' (EU/World Bank, 1996b,P9);

• open ing the market quickly (by end-19 97) toforeign com petition; an d

• deve loping effective me chan isms for ban kin gsupervision.

The first of these elements is clearlyappropriate. The financial situation of themajor bank s is such that the ir ability to o per ate isseriously impaired; and this represents animportant constraint on private sectordevelopm ent. In the long term, ther e is no goodreason to retain the banking system within thepublic sector, although the a pp ropr iate pace ofprivatisation is open to debate (as discussed inSection D3). The last element — improvingsupervisory mechanisms — is clearly essential.

Sup port for existing small banks is less clearly apriority, but probably desirable. The smallprivate banks are under-resource d and inexper-ienced, but they provide a useful alternativesource of financial services. The ir ro le will remainimportant at least until the larger banks arerehabilitated. However, encouraging the devel-opment of new private banks is more question-able. The banking sector is already over-populated, with a total of 49 banks op erating inearly 1996 (World Bank, 1996b, p52); and bothnew and existing private banks would probably

be too small to compete effectively in the longterm — most of the current private banks haveless than DM2.5m of net capital (EU/Wo rld Bank,1996b, p9). A shake-out of the system is likelyover the medium term , throu gh a combination ofmergers and bankruptcies. The benefits ofopening the banking sector to foreign compet-ition — or at least of doing so imm ediately — a realso de batable.

The World Bank (1996b, p58) presages adifference of view with the authorities on thenature of the financial sector. The Bank antici-pates that the authorities will opt for a universalbanking system, in which the same financialinstitutions provide the whole range of financialservices (insurance, for example, as well as

46



Economic policies

deposit-taking and lending), p artly because thisis in line with the situation in Western Europe,and may thus be seen as contributing to BH'sultimate accession to the EU. However, the

Bank argues that this option is inap prop riate toBH, because it 'may create major regulatorydifficulties'. There seems little reason toquestion this judg em ent.

Dilemma 5: Free Entry for Foreign Banks?

It is proposed that the banking sector shouldbe opened to foreign competition as quickly aspossible, as a means both of increasing theefficiency of the sector (through competition)and of imp rovin g access to credit a nd financialservices for the pro duc tive sec tors.

At present, it seems very unlikely that the

domestic banking sector would be able to copewith international com petition — particularlyif new entrants were to cross-subsidise theiropera tions initially as a me ans of break ing in tothe market. Foreign competition during therehabilitation and restru cturin g of the bankingsystem could seriously weaken the domesticfinancial sector, to the detriment of develop-men t over the longer term .

From the point of view of the bank ing sector,a more considered pace would seempreferable. This would provide a breathing-

space for the large banks to re-establishthemselves post-restructuring, and for thesmaller banks to strengthen their operations(with donor support) and consolidate theirmarket position (through mergers). Exposingbanks to the full rigo urs of foreign com petition

before these steps were accomplished wouldraise the risk of widespread bankruptcies —especially among the small private banks —leading to a largely foreign-owned financialsector in the future. This is unnecessary, andless than ideal from a long-term economicperspective. The collapse of much of the small

bank sector would also largely negate thebenefits of don or sup po rt to small banks.

There is an argum ent for a substantial delay,of perhaps three to five years, in opening thebanking sector. However, this would reducethe availability and increase the cost of creditand financial services to productive sectorsdur ing the critical reconstruction phase , whichin turn would have some negative effect onemployment.

Th ere is no obvious answer to this problem.It is necessary to make a jud gm en t between theshort-term economic and social benefits of animmediate opening of the market (whichwould have longer-term knock-on effectsthrough health, education, and politicalstability), and the more direct long-termbenefits of deferment.

D.6 Trade policyIt is proposed that all administrative restrictionson imports should be abolished, and tha t imp orttariffs sh ould be levied at a stand ard rate of8 percent. Apart from the usual arguments aboutremoving economic distortions, the p roposal isjustified by the intention to join the WTO; andthe objective of ultimate accession to theEuropean Union. A potentially significantconstraint on trade policy is the recent free-trade agreement with Croatia, which limits the

scope for p rotectionist p olicies.Th e removal of administrative restrictions onimports is unproblematic (assuming that thebasic requirements of health and safety are

observed). However, the immediate reductionof tariffs to a standard 8 per cent raises import-ant political issues. It is also less clear-cut ineconomic terms (because of the direct effect ongovernment expenditure under the currencyboard system), and is arguably based as much onideology and presupposition as on the realitiesof the Bosnian situation.

The political dimension arises because thecu rre nt level of tariffs ap pe ars to be significantlyhigher in RS than in the Federation. In RS, theaverage rate is 15 per cent, with substantially

higher rates for imports classified as 'luxurygoods' (RS, 1996, plOO); in the Fed eratio n, ratesvary between 0 per cent and 20 per cent(World B ank, 1996b, p25). If this is the case, the

47



Th e IM F, the World Bank and econom ic policy in Bosnia

unification of rates at a lower level would implya g rea ter fiscal cost to RS than to the Federation.

In any event, the RS authorities appear tohave little desire to redu ce tariffs, in view of the

cur ren t tight fiscal constraints. They also ap pe arto envisage at least the possibility of intro du cin gnew trade barriers in some circumstances:

Restrictive measures and import limits, as well as

customs tariffs will not b e introduced without a

detailed assessment of domestic costs an d resources

an d effective protection levels.

(RS, 1996, pi 19)

The operation of different trade regimesbetween the two Entities would raise substantial

problems. If there is free movement of goods,there will be a strong incentive for importers inRS to import goods through the Federation, inorder to benefit from the lower tariffs; and

unless distrust o r national solidarity w ere suffic-ient to overcome this, the RS authorities wouldface a substantial loss of revenue as a result.They may be forced to match the Federation's

tariff levels; but, to the extent that they are ableto , they may prefer the alternatives of obstruct-ing the free movement of goods and fosteringdistrust and anti-Federation solidarity.

The immediate needs for low and uniformtariffs are relatively limited. The potentialbenefits of WTO m emb ership are sm all; and, asdiscussed above, the prospect of even startingthe laborious process of EU accession remainsdistant. There could also be some economicbenefits from maintaining (temporarily) am od erate level of protection, on a selective basis,in the form of higher impo rt tariffs — altho ughthere are some potentially im porta nt trade-offsinvolved.

Dilemma 6: Trade policy

Virtually the entire BH economy is at presentmade up of 'infant industries' and 'conval-escent industries'. This would suggest that th eremay be a case for the ma intenan ce of a m ode r-ate level of protection through tariffs on non-essential consumer goods for a strictly limitedperiod, to help the development of small andmedium private enterprises in import-substituting sectors, and the rehabilitation ofsocially-owned and newly-privatised ente rprises,by offering them a breathing-sp ace before the yface the full rigou rs of foreign com petition .

Such protection would also provide an

additional source of government revenue,helping to ease the very tight constraint onpublic sector spending and allowing a fasterreduction in wage-based taxes and contribu-tions. It could also help to ease a source ofpolitical dispute between the two Entities, bybringing Federation policies into line withthose of RS. Th e potential for economicdistortions could be limited by phasing out thetariffs over time, according to a rigidly-imposed schedule established at the b eginningof the process, reflecting the expected timing

of a retu rn to full capacity-utilisation.At first sight, this might ap pe ar to com pou nd

the problem of exchange-rate adjustment

under the currency board, as discussed inSection Dl. Trade liberalisation has an effectsimilar to a real appreciation of the exchangerate (hence the need for accelerated devalua-tion during trade liberalisation in countriesfacing serious balance-of-payments constraints);and a real appreciation is desirable, but notachievable, during the reconstruction phase.Conversely, to liberalise imports as aid flowsare reduced over the medium term, when areal devaluation is desirable, would a pp ear tobe at odds with real excha nge r ate objectives.

In practice, however, the distinction doesnot appear to be so clear-cut. Significant realappreciation of the exchange rate would onlyoccur during the reconstruction phase if thecapacity constraint were reached, whichappears unlikely. (See Section El.) Even then,maintaining or increasing the level of protec-tion would merely enable a faster rate ofaccumulation of foreign exchang e reserves, orslow down the accumulation of external debt.30

Reducing tariffs over time would imply a(perverse) real appreciation of the exchange

rate, as aid flows decline, and a realdepreciation is req uire d. H owev er, this effect ismitigated by thre e factors:

48



Economic policies

Dilemma 6: Trade policy (continued)

• the level of the real exchange rate wouldthroughout be lower (and thus the balanceof payments stronger) than under theimmediate liberalisation option, ending upat the same level at the e nd of the process;

• to the extent that prod ucer s benefited fromthe breathing-space offered by temporaryprotection, and thus achieved a hig her levelof production of import substitutes (or werein a stronger position to move into exportproduction), the balance-of-payments posi-tion would be strengthe ned for a given realexchange rate, and a higher real exchangerate would thus be ap prop riate; and

• the reserves accumu lated dur ing the earlierpart of the process would provide anadditional cushion, to finance a current-account deficit; and any reduction in theaccumulation of debt would improve thecurrent-accoun t deficit by lowering interestpayments.

After the six-year period for which thecurrency board is mandated, if the politicalcircumstances were conducive, it would bepossible to adjust the exchange rate if necess-ary. The m itigating factors noted above shouldat least delay the poin t at which o ver-valuationof the real exchange rate occurs, reducing thelikelihood of a financial crisis in the inter im .

A second caveat is the argument thatprotection would give rise to a distorted set ofeconomic incentives, encouraging productionof goods which could be imported more

cheaply, and discriminating against exportproduction. However, while the economy isoperating far below full capacity, over-produc-tion of import substitutes would do little eco-nomic harm, and could bring substantial

benefits in the form of employment. Thenegative effect on exports generally associatedwith protection should be limited, as it ariseslargely from upward pressure on the nominalexchange rate, which cannot be reflected inactual rates under the currency board system.If there were thought to be a significant neg-ative effect o n ex po rts, this cou ld be co un teredby a system of export subsidies at a rate nohighe r than imp ort tariff levels. Since the trad ebalance is likely to remain substantially in

deficit for the foreseeable future, the net effecton the public finances should rem ain positive.To limit the distortionary effect, it would be

critical to ensure that higher tariffs weregenuinely temp orary, and the schedule for theirreduction strictly adh ere d to. This often provesdifficult; an d in BH could be a particular pro b-lem if the benefits of protection were signifi-cantly skewed in ethnic terms ." The re appearsto be a need for the sch edule of tariff reductionto be embodied in entrenched (ie unchange-able) legislation. However, the general aspira-

tion to membership of the European Unionwould represent an increasingly strongincentive for tariff re duc tion, as the prospect ofaccession cam e (at least slightly) closer.

Th ere is a critically im po rtant caveat to anyautomatic process of import liberalisationthrough entrenched legislation. Whatever themerits of this particular case, allowing theauthorities to pass entrenched legislationwould create a preceden t; and could allow thepresent nationalist leaderships to tie the h andsof future moderate authorities. Even with therequirement of consensus, this could createserious obstacles to a long-term politicalsolution, unless there were some form ofexternal control or constraint.

49



Th e IMF, the W orld Bank and economic policy in Bosnia

D.7 Labour market deregulation

As noted above, part of the strategy for thedevelopment of the private sector and thecreation of employment is the liberalisation ofthe labour market, so as to rem ove an obstacle toSME viability and to enc ou rage en terp rise s of allsizes to take on more employees. Liberalisationis also intended to discourage the diversion ofecono mic activity into the informal sector, whichwould both remove workers from even the m ostbasic protection, such as health and safetyprovisions, as well as reduc ing the tax base.

This policy encompasses three mainelements:

• the removal of restrictive regulations onemployment contracts (eg those making itdifficult and/or expensive to fire workers,requiring very favourable annual leavearrangem ents, etc);

• lowering minimum wage rates, or remo vingthem altogether; and

• lowering tax rates on wages (including em p-loyer and employee contributions for statepensions, unemployment benefits and healthservices).

It seems fairly clear that some movement inthis direction is necessary. Even before the war,

the form er Yugoslavia was widely seen (not leastby some Bosnians) as 'living beyond its means'.Bosnia's means are now more limited than ever;and it is not realistic to expect the high wage-levels and long holidays which existed beforethe war to be reinstated . N either is it realistic, inthe current state of economic uncertainty andchange, to expect employers to take on staffwhom they cannot later shed (or only at aprohibitive cost); and, with the currentexceptionally high level of unem ploym ent, anartificially high minimum wage would merelyadd to the problem, making any kind of socialsafety-net un viable.

Lowering tax rates is probably also necessary.

At present, well over 50 per cent of the gross w ageis taken up by taxes and social contribution s; an dthis is likely to be a significant impediment toemployment. It is argued by some that areduction in rates could actually have a positiveoverall effect on the public-sector finances, byencouraging more formal-sector employment(and thus broadening the tax base) and reducin gthe need for expenditure on unemploymentbenefits and other social safety-nets. However,this view is probably exaggerated.

In all thre e aspects, the question is how far to go;

an d the details of these policies are as yet too vagueto assess whethe r they a re likely to go too far.

Dilemma 7: L abour market deregulation

A case could be made for m aking the reductionof minimum wage levels and deregulation of

contracts temporary, ma ndating a progressiveincrease in minimum wages (though to a lowerlevel than pre-war) and a reintroduction of(less restrictive) regulations on dismissal,annual leave, etc over time. This would offerthe prospect of a gradual restoration of livingstand ards, subject to the inevitable con straintsof political stability and economic perform-ance.

To the extent that the two Entities (and thetwo parts of the Federation) continue tooperate as largely separate, competing

economic units, gradual re-regulation couldlimit the extent to which competition tookplace on the basis of bidding-down real wages

and working conditions. In the short term,while unemployment remains exceptionally

high, this would not be appro priate: the effectwould simply be to increase unem ploym ent inthe poorest (ie RS and Bosniac) areas. Lowwages will be necessary to off-set the costs toemployers of the worse economic conditions.As unemploy me nt falls to m ore n orma l levels,however, the risk of bidding-down would beincreased; and there would be considerablescope for resentment in Croat areas at losingjobs to low-wage workers elsewhere in BH.

Ultimately, if Bosnia were to accede to theEU, it would need to accept the provisions of

the Social Chapter. An approp riately-designedlabour-market re-regulation could lead up tothis gradually.

50



Economic policies

Dilemma 7: Labour market deregulation (continued)

However, there are some impo rtant caveats:

Potential investors (foreign and domestic)could be dete rred by the prosp ect of havingto increase wages and improve contractualterms in the future. This risk could bereduced by limiting changes to newly-contracted staff, which would also provide adisincentive to shedding staff unnecessarily;but investors would still anticipate anincrease in production costs, (albeit a more

gradual one). The investors discouragedwould be those who migh t otherwise make along-term commitment to production inBH, rather than foot-loose industries readyto move on once real labou r costs increased— although the latter may also be pu t off ifthey thought their window of opportunitywould be reduced so far as to makeinvestment unviable.

If BH is to perform well in an increasinglyglobalised world economy, it will have to

retain its international competitiveness overthe long term. It will be im po rtant to ensurethat any re-regulation of the labour marketdoes not compromise this objective, eitherthrough its nature or through its rate ofimplementation.

It would be very difficult to make a realisticestimate ex-ante of the rate at which mini-mum wages could realistically be increased,or contractual regulations reimposed. Sincean automatic re-regulation would probablyneed to take the form of entrenched legisla-

tion passed at the o utset (to overcome politi-cal inertia), over-optimism could be verydamaging, unduly increasing unemploy-ment.

Since responsibility for labour-marketregulation is held at the Entity level, and the

labour market prospects of the two Entities(and of the two parts of the Federation) maybe very different, the prospects of a satis-factory agreement on re-regulation beingreached at the n ational level are lim ited. It islikely that this would entail a long process ofnegotiation, which would take up scarceadministrative resources, as well as creatingconsiderable potential for dispute. It is notclear that this is necessarily preferable to laterarguments about low-wage competition.

• As in the case of trade, whatever the meritsof this particular case, establishing aprecedent for entrenched legislation wouldbe unhelpful unless such legislation weresubject to some form of extern al c ontro l.

All in all, this may be an issue worthconsidering; but it is far from clear that anautomatic labour-market re-regulation isdesirable. A more general commitment tomove gradually tow ards provisions in line with

the EU's Social Chapter might be moreappropriate; but the potential for dispute andthus for political inertia would again beconsiderable.

A third alternative would be to establish anon-entrenched schedule of changes, begin-ning in, say, three to five years' time, leadingprogressively from the (post-deregulation)status quo to reach the standards of the EU'sSocial Chap ter at the (conservatively) exp ectedtime of accession. It would be preferable toallow some scope for delaying the start of the

process and adjusting its pace after the initialgrace period, in accordance with the economiccircumstances of the time and revisedexpectations on the timing of accession.Allowing for more detailed am en dm en t wouldprobably not be desirable, as this wouldprovide much greater scope for dispu te.

51




D.8 Utilities and public services

Beyond physical reconstruction and rehabil-itation, the intention is to reunify the p ublic utilityinfrastructure between the Entities. Energy, water,and telecommunications systems were originallydeveloped for Bosnia as a whole, and wereintricately linked between the Entities prior to th ewar. They have now been effectively separatedalong the inter-Entity b oundary, to the detrimentof efficiency, service costs, and financial viability.Reintegration is highly desirable from an economicperspective, and would be a significant steptowards reunification. However, precisely becauseof the latter aspect, there may be important

political obstacles to movement in this direction.A second, and more controversial, policy inthis area is to develop and extend cost recoveryfor services such as water, sanitation, and solid-waste disposal, as well as for energy supplies(electricity, gas, and district heating). Thepro po sed timetable for cost recovery is relatively

quick: for water, sanitation, and solid waste, theobjective is to recover all operations and main-tenance costs in three years, with full costrecovery after five years. Social-pricing mechan-

isms (to limit the social impact) are regularlymentioned in Bank documents in this context,bu t only in passing. Th ere is no discussion of thetypes of mechanism which could be use d, or ofissues in their design and application.

The desire to increase cost recovery ismotivated primarily by the fiscal constraints onthe public sector — no doubt considerations ofeconomic efficiency are also at play, but theseare no t emphasised. It should be noted that costrecove ry was well develop ed pri or to the w ar, so

the perception problem which arises when feesare introduced for previously free servicesshould not be serious.

Nonetheless, there are potential concernsabo ut the ex tent of cost recovery, the pace of itsimplementation, the potential social impact,and the regional/ethnic dimensions.

Dilemma 8: Cost recoveryThere are three major concerns about therapid and general application of cost-recoverypolicies in B H .

• While cost recovery for energy is fairlyuncontroversial, for some other services thecase is less clear. This applies p articularly tosanitation and solid-waste disposal, as thereare major public-good benefits arising fromtheir use. From an economic perspective,this represe nts a stron g case for subsidy.

• Th e social impa ct of increasing cost recoveryis a matter of potential concern. Th e bread thof cost recovery envisaged, and the relativelyrapid achievement of full cost recovery,suggest a major escalation of basic living costsover the next five years. While rapideconomic growth is exp ected, at least initially,the benefits will by no means be universal,and will be critically dependent on BH'ssuccess in employment generation, and theexte nt an d effectiveness of safety- nets. Th er e

are serious concerns on both these issues, asdiscussed in Sections El and E2.

• Th er e is also a potentially impo rtant reg ion-al dimension to this issue. While the ex istingregional disparities in income persist, theability to afford services will vary consider-ably between RS, Muslim and Croat areas.Assuming similar levels of consumption,applying the same principles and targets inall areas would mean a much greate r imp acton real incomes in the poorest (generallyMuslim and RS) areas.

These concerns would suggest a morelimited and /or more g radual application of costrecovery. H owever, there would be a high costto such a policy: because of the tight fiscalconstraint, spending money on subsidisinginfrastructure services directly reduces theresources available for other uses. Thepotential benefits of subsidising infrastructureservices therefore need to be evaluated relativeto alternative uses of funds.

Social-pricing m echanisms of some kind will

be critically important, especially over dieme dium term , as a m eans of limiting the social

52



Economic policies

Dilemma 8: Cost recovery (continued)

impact. If such m echanisms a re n ot in place, itmay well be necessary to moderate the rate ofincrease of charges, and to delay cost-recoverytargets. It should be noted, how ever, that socialpricing would reduce the revenues generated— it may well be necessary to subsidise theconsumption of a large proportion of thepopulation, at least initially. The administrativecosts of social-pricing mechanisms may also besubstantial.

In view of the regional dimension, it mightseem preferable from a distributional perspec-

tive to apply more rigorous standards of costrecovery in higher-income (mostly Cro at) areas,in effect cross-subsidising services in the poor-er areas. However, this could create consider-able resentment. Social-pricing mechanisms

could ease the distributional problem withoutan explicit regional differentiation; butresentments could still arise if it becameapparent that the costs and benefits of socialpricing had a strong regional or ethn ic bias.

There are no easy answers here. Thebenefits of not imposing full cost recovery forsanitation and solid-waste disposal should becarefully evaluated, and com pared with altern-ative uses of governm ent spending. W here costrecovery is introduced, social pricing shouldbe include d as an integral p art of the policy, to

protect poorer households; and it should bedesigned in such a way as to protect all thosewho would face an excessive financial burden.However, direct cross-subsidisation betweenthe thr ee ethnic regions should be avoided.

D.9 Agricultural policies

The proposed agricultural strategy is based onpractical support for small private farms,including the provision of imported inputs,credit, livestock, and equipment. These policiesinclude cost-recovery measures, and it isstrongly recommended that subsidisation ofagriculture should be avoided. The state-ownedfarms are to be privatised, as are supportservices and the production of agriculturalinputs and processing industries. Food aid is tobe phased ou t, to avoid un derm ining th e incent-ives for domestic production and reducing

produ cers' incomes. There is also a proposa l fora land bank, to facilitate the departure offarmers from the agricultural sector, and toallow the e nlargem ent of the rem aining plots. AWorld Bank Agriculture Sectoral AdjustmentLoan is expected in the near future to supportthese policies.

These policies appea r to be somewhat at od dswith the policies envisaged by the RS authorities(RS, 1996, pp 121, 123), which include, forexample, direct state provision of veterinaryservices and credit, the possibility of subsidising

producers in relatively undeveloped areasthrou gh direct grants, and 'protective prices' forpriority products, including wheat, corn, sugarbeet, oil-bearing crops, meat and milk. These

represent a large proportion of RS's totalagricultural production.

Pursuing different agricultural policies in thetwo E ntities — particularly with regard to prices— would cause serious problems, and is not aviable option. This implies an urgent need toreconcile policies and priorities between theEntities, if this has not yet been achieved. Thiswould be a prerequisite for a SectoralAdjustment Loan.

The evolution of conditions affecting theagricultural sector over time will be criticallyimp ortant, and needs to be taken carefully intoaccount in policy design. In the short and

med ium term, importa nt constraints arise fromlandmines, and ownership disputes arisingfrom population movements, as well as the lackof equip me nt and limited access to othe r in puts.Major concerns over this period will be to re-absorb as much as possible of the rural labourforce into the agricultural sector at adequateincome levels (to limit rural unem ploym ent andpoverty and excessive pressure for rural-urbanmigration); and to encourage an expansion offood production at affordable prices, to fill thegap left by declining food aid receipts. In the

longer term , it will be necessary to impro ve farmincomes; to increase efficiency and intern ation alcompetitiveness; and to prepa re the grou nd forultima te accession to the EU.

53




These aims require very different policies insome res pects ; and it seems possible that policiesdirected towards longer-term objectives arebeing started too soon. For example, support

for farm mechanisation (financing for tractors,for example), and the land bank will tend toincrease the size and capital-intensity of farms,which implies a reduc tion in labour use. In someareas, population shifts from rural to urbanareas have already occu rred an d are unlikely tobe reversed in the near future. However, theremay be other areas where the shedding of

labour could contribute to rural unemployment,declining real wage levels, and rural-urbanmigration (adding to urban un employm ent rates).

A long-term agricultural strategy is needed.

However, uncertainties relating both to BH'spolitical future and to the prospects for EUaccession make devising a strategy a compli-cated process. Creating the conditions for EUme mb ership will be an imp ortant objective; butit will be important to be realistic abou t the likelytime-frame of accession, and to ensure theviability of the sector in the interim .

54



£ . Social dim ensions

£.1 Em ployment and theEmergency Social Fund

The need for employment creation is wellrecognised in World Bank documents. Unemp-loyment is described as 'an overwhelmingeconom ic and social problem' (EU/EBRD/WorldBan k, 1996, p94), and job creation as 'absolutelyessen tial' (EU/World Bank , 1996a, p9) , 'no t only

an economic necessity but...essential for peace'(World Bank, 1996b, p xvii), 'the most importantchallenge for the economic recovery program-me' (World Bank, 1996b, p i 1), and 'an absolutepriority' (EU/World Bank, 1996b, plO).

At present, unemploym ent is estimated at 50 -60 per cent of the population of working age,although it is substantially lower in the Croatarea of the Federation. It is estimated that thereconstruction programme created some 250,000jobs at its peak in 1996 (EU/EBRD/World Bank,1996, p94). This is equivalent to half of totalemployment in BH during that year. If theaverage level of employment generated byreconstruction during 1996 were half of thislevel, this would imply an unemployment ratewithout the reconstruction programme in theorder of 60-70 per cent. Adjusting further forthe effects of demobilisation du ring 1996 -7 couldincrease the un derlying rate post-demobilisationfurther to between 85 per cent and 95 per cent.32

As the reconstruction programme iscom plete d, these job s will disap pea r (hopefully

to be replaced by jobs in running the rehabilit-ated infrastructure and services). At the sametime, unemployment may be further increasedby the return of refugees, and possibly byfurther demobilisation (as well as shifts in thegeograph ical distribution of emp loym ent du e tothe r etu rn of displaced peop le).

The potential effects of population move-men ts are enormo us: there are between 1.0 and1.4 million refugees; and a greater proportionof these people are likely to be of working agethan in the settled population. If all refugees

were to return, the demand for employmentcould increase by more than half. Excludingreconstruction-related employment, this would

imply an unemployment rate of about 85-90pe r cent. Reducing unem ploymen t to a sustain-able level (say 10 per cent), while allowing areturn of the refugees, would require anincrease in non-reconstruction employment inthe order of 400-650 per cent from its currentlevel. Even if no refugees were to return, non-reconstruction employment would need to bedoubled or trebled.

The relationship between population move-ments and employment is a two-way one. Anecessary (th ough not a sufficient) co ndition forthe return of refugees and displaced people totheir areas of origin is their expectation thatthey will have an adequate income. Withpensions and social benefits likely to remain atminimal levels over the medium term, this willreq uire the creation of adequa te employ men t atreasonable wage rates to reabsorb the return eesinto the economy. Unless and until this isachieved, displaced people are likely to remain

where they are, and refugees will be returnedonly by compulsory repatriation. Moreover,compulsory repatriation will have a seriousecono mic an d social impact, not only leaving th ereturnees themselves (or those they displacefrom jobs) without a viable income, but alsoincreasing the strain on the budget and poten t-ially reducing overall real wages by adding tothe excess supply of labou r.

This suggests that the compulsory repatria-tion of Bosnian refugees from 'host' countrieswill be seriously d etrime ntal (except in the caseof people with substantial financial or valuablehuman capital). It will be economically damag-ing; it will have a substantial social impact; and,in cons equ enc e, it could be a significant obstacleto a viable and peaceful future. Donors shouldboth avoid this option themselves, and putpressure on others to do so. They might alsoconsider mechanisms for sharing the financialburden of refugees, to reduce the pressure forrepatriation on the most important hostcountries.

As discussed earlier, the strategy for emp loy-ment creation is based on private-sector develop-ment, through privatisation and the promotion

55




of small and me dium enterprises. Ho wever, it isrecog nised that this will not be achieved quickly;and that domestic social safety-nets will beinadequate to fill the gap in incomes for theimmediate future. Thus the World Bank(1996b, pi5) acknowledges that:

A significant share of the population will continue

to depend on a ssistance in the medium

term...Temporary programs are required until

pension, unemploym ent and social assistance

systems are restarted.

If the World Bank's projections for outputgrowth (World Bank, 1996b, Table 5.3) provecorrect, and employment grows in line with

outpu t, unem ploy m ent would fall to somewherebetween 6 per cent and 25 per cent in 2000(assuming that no refugees r etu rn, bu t allowingfor 0.6 per cent pa pop ulation g rowth). In otherwords, the re would be n o scope for reabsorbingrefugees into the labour mark et without addingto unemployment until at least 2001. At thehigher starting-point for unemployment, fullemploym ent of the cu rrent population would bereached only in 2004. If the projected growthrate for 2001-5 were continued thereafter,refugees could then be fully reabsorbed into the

labour market by 2006 on the most optimisticscenario, or 2011 on the most pessimistic.

However, these projections are fairly sensi-tive to the assum ptions on which they are based.Moreover, as noted above, the World Bankprojections appear very optimistic; and extra-polating the projected 2001-5 growth rate (8.4per cent pa) into the indefinite future is almostcertainly not realistic. It also seems qu estionablethat a 1 per cent increase in output would leaddirectly to a 1 per cent increase in employment.Apart from anything else, real wages are

currently very low, and capital very scarce (atleast in RS and Bosniac areas); and a one-to-onerelationship between output growth andemployment growth would imply avoiding anyincrease in either th e real wage rate or the shareof capital in total income. Apart from thequestion about its realism, the assumption ofcon stant real wages would imply very high levelsof poverty (and therefore a considerable needfor social support) for employed workers andtheir families. (It should also be borne in mindthat the jobs associated with externally-

sup po rted reconstruc tion will be lost by 2000 asthe reconstruction programme winds down;and that demobilisation will further reduceemployment from the 1996 level.)

A fairly limited adjustment to take account ofthese considerations would have a substantialeffect on th e results. If the grow th ra te of emp loy-men t were one-third lower than the W orld Bankprojections for output growth (by no means apessimistic assumption), unemployment wouldremain between 24 per cent and 39 per cent in2000; and full em ploym ent of the curren t popula-tion would be achieved only between 2006 and2010. It would take until 2 019-27 to reabsorb allof the refugees into the labo ur m arket.

At present, emergency food-aid is an impor-tant form of social sup po rt. How ever, it is largelyan overhang from the war, and is being phasedout by donors. The proportion of the population

receiving food aid is officially estimated to havereached 80 per cent in 1995, although someinformed observers believe the figure to bemuch lower. The Federation authorities expectfood-aid recipients to decline to 40 pe r cent of thepop ulatio n in 1997, 20 pe r cent in 1998 and lessthan 5 per cent in 1999 (World Bank, 1996b, pi 5) .

Apart from food aid, there are two maintemporary programmes for income support:

• The Emergency Public Works (EPW) pro-gramme is intended to increase employment

opportunities until private-sector employmentgeneration takes over, by providing funds forsmall, labour-intensive rehabilitation projects.

• Th e Em ergency Social Fund (ESF) is intendedmainly to provide a minimum level of socialsupp ort to the poores t househo lds, in the formof monthly cash pa ym ents, until the pensions,unem ploym ent benefit, and othersocial supportsystems are once again fully operational.

However, there are serious doubts about thesize of these programmes relative to the scale of

the problem; and their duration appears to fallfar short of the likely time-frame for the resto ra-tion either of full employment or of benefitpayments. Both of these problems are com-pounded by the major shortfalls in cofinancingfor these two programmes, which will furtherreduce their scale and duration.

The EPW programme is, in principle, veryhelpful, although its scale is very limited. Thetotal amount originally envisaged was $45mover two years, implying the creation of about3,800 jobs for the two years of its operation33

—enough to reduce unemployment by less than0.4 per cent. If the cu rre nt financing gap for thisproject is not filled, this will be reduce d to abou t0.1 p ercen t .

56



Social dimensions

T h e ESF is also very welcome, bu t limited interms of its coverage, the level of paymentspro vide d, and its likely duration. C overage re p-resents less than half of the number of people

unemployed (and is likely to include a largeproportion of recipients not classified as unem-ployed, such as those above working age anddependent on pensions, orphans, and house-holds with some employment but includingm em be rs with war-related disabilities).

The ESF is supposed to provide monthlypayments to households totalling DM5m(US$3m ), spread between 150,000 poo r house -holds across the country, and 100,000 orphansand war disabled in four municipalities (WorldBank, 1996c, pp 9, 93). The average monthlypay me nt per recipient household would thus beDM20. Allowing for an average household sizeof three, this implies payments of about $0.13per person per day, a fraction of the absolutepoverty line ($1 per person per day atpurchasing-power parity exchange rates).

The planned duration of the ESF is onlytwelve months; and this could be furthersho rten ed by the substantial financing gap . Thefunding gap has so far been filled partly by theBank accelerating its own disbursements, and

partly by the authorities diverting some of the irbalance-of-payments support to fill the gap.However, disbursements nonetheless seem tohave been slowed down: as of mid-November1996, eight months after the EmergencyRecovery Project (of which the ESF forms a part)became effective, disbursements totalled onlyDM9.4m (World Bank, 1996c, p4). Thisrepre sen ts a monthly rate less than o ne-q uarte rof that assumed in the last paragraph.Mo reover, the Bank resources available for th eESF will soon ru n out; an d, in view of the strict

budget constraints imposed by the currencyboard, the diversion of balance-of-paymentssupport puts additional pressure on other areasof governm ent spending.

It seems clear that there is a considerable g apbetween employment and social needs and the

current scale of assistance aimed at meetingthem; and that the time-frame for reducingdependence on social safety nets, and for thereturn of refugees, is considerably longer than

the World Bank appears to envisage. Thestatem ent qu oted ab ove, that 'a significant sh areof the population will continue to depend onassistance in the medium term' does not comeclose to reflecting the gravity of the situation.Coupled with the very short time-frame of theESF an d EPW programm es, and the rapid phasingout of food aid, this implies a choice betweenwidespread and severe poverty, or an intolerableburden on the public finances to support safety-net pro gram me s for the next 20-30 years.

There are also potentially serious implicationsin the shorter term. If, as seems quite possible,unem ploym ent remains at 25-40 per cent in theyear 2000, and specific employment creationand e xtern ally-su ppo rted social safety-nets havebeen phased out, the prospects for reconcilia-tion or a shift towards political moderationcould be seriously jeopardised. This could becompounded if there were a contraction in theeconomy du e to the reduction of aid flows andthe exchange rate rigidities associated with thecurrency board arrangement in 1998-9. Sincethe W orld Bank's growth projections (and thusthe decline in un employm ent) rest heavily on areturn to pragmatism and moderation in thepolitical sphere, there is a real risk that thewhole political and economic future of BHcould be irrevocably comprom ised.

It is not realistic to assume tha t the don ors willcontin ue to fund social programm es on the scalerequ ired over the nex t 20-30 years. However, itwould seem essential that programmes such asthe ESF and the EPW project are both substan-tially enlarged and extended over at least the

next five years. Only if the decline inunemployment and poverty can be greatlyaccelerated can there be any hope of creatingthe political conditions for long-term economicgrowth. This raises some important questionsabou t the economic strategy being p ursu ed.

Dilemma 9: Employment creation and the economic strategy

A major part of the rationale for the currenteconomic strategy is the rapid creation ofemployment opportunities. However, this isnot being achieved at an adequate rate, and it

is by no means obvious that it is likely to beachieved within a politically viable time-frame.The question is whether a fundamentallydifferent economic strategy would be more

57




Dilemma 9: Employment creation and the economic strategy (continued)

effective in reaching this objective, or whetherpotentially beneficial changes are limited tomino r modifications to the curre nt strategy.

It would be possible to envisage a ratherdifferent strategy, modelled in part on theEmergency Public Works programme. Thiswould increase the emphasis on (and donorsupport for) labour-intensive reconstructionand rehabilitation work. The re is clearly a veryconsiderable amount of such work, only afraction of which has, as yet, been done. Thespending of those employed in such projectswould generate additional demand in theeconomy, strengthening small and mediumenterprises and creating more (hopefullysustainable) job s.

However, there are some important limita-tions to this altern ative:

• Increa sing su pp or t for labour-intensivepublic works could only be achieved byincreasing total aid (which would be difficultto justify), or by reducing sup por t to otherprogrammes. If other components of the

reconstruction programme were substan-tially reduced (for example, institution-building, infrastructure rehabilitation, orsup por t to th e social sectors), the re would bea dang er of comprom ising long-term sustain-ability and oth er aspects of basic needs in theinterests of short-term employm ent gains.

• The cost of employment generation underthe EP W program me to date is very high - inthe order of $5,500 per person-year. Arough estimate suggests that this may be

only slightly lower than the equivalent figurefor the overall aid programme.34 Based onthis estimate, transferring the entirefinancing for the reconstruction program mein 1996 would have led to a net reduc tion inunem ploym ent of only about 3 per cent.

• The potential for labour-intensive publicworks is much greater in the Federation(where there is substantial war damage)than in RS (where the main damage hasbeen don e by econom ic sanctions). An EPW-based strategy would therefore have morelimited scope in RS, so that shifting theemphasis towards this type of programmecould inhibit any correction of the regionalimbalance in a id.

• Th e jobs created directly by EPW schemeswould be tem pora ry, lasting only as long asthere was app ropriate work to be done andadequate donor funding to pay for it. Thelonger-term benefits would be limited to thesecondary effects on employment, throughincreased demand.

In practice, an approach centred on labour-intensive public works is therefore unlikely tobe viable as a basis for sustainable long-term

grow th. To have any significant positive effect,it would be essential to re du ce th e cost of EPWprojects substantially. At present the cost perjob seems unnecessarily high — 130 per centmore th an the averag e net wage in Croat areas,and nearly five times that in Bosniac areas.How ever, a detailed assessment of how the costcould be reduced is well beyond the scope ofthis paper.

Even at a much lower cost, the argum ent forincreasing suppo rt for EPW is essentially politi-cal: as long as unemployment and poverty

remain at their current very high levels, thiswill be a serious imp edim ent to the creation ofa political climate co ndu cive to peace , stability,and economic growth. It is possible — but onlypossible — that greater employment creationcould shift the political process in the direc tionof moderation and conciliation; and that thiswould substitute for t he effects of those projectswhich would need to be forgone to finance amajor EPW p rogram me.

A m ore limited shift in this direction (coupledwith a substantial cost reduction) would be mo re

viable. This would en tail a detailed appraisal ofthe whole aid programme, to identify thosecomponents which are, on the whole, lessbeneficial than sup po rt to EPW.

However, employment creation based onpublic works can be no mo re than a tem porarypalliative; it is no substitute for an economicpolicy programme designed to generateemployment. The alternative approaches toprivatisation and trade policy outlined inDilemmas 4 and 6 might help to strengthenand accelerate job creation und er the curren t

private-sector-led strategy.

58



Social dimensions

£.2 Pensions and benefits

In time, it is hoped that domestically-financedsocial safety-nets, such as pensions and unem-

ploy me nt benefits, will take over from the ESF.Ho wev er, this is a very lo ng way off, and mov e-ment in this direction will be very slow.

Pensions have fallen dramatically since thebeg innin g of the war, and at times have not beenpaid at all in the Federation, reflecting theextrem e p ressures on public sector finances.

35 In1996, average pensions in Croat areas wereDM65 (US$38) per month, slightly above theabsolute poverty line. In Bosniac areas, however,they were only one-fifth of this level; an d in RS,while the authorities claim that there are no

payments arrears, they were 'at a very minimallevel'. As a result, many pensioners were (andstill are) living in 'conditions of extrem e pov erty'(World B ank, 1996b, p48).

A major contributory factor is the exception-ally high dependen cy ratio. Because of the dispro-portionate number of people of working agewho left BH as refugees, coupled with the veryhigh ra te of unem ploym ent, there are nearly asmany pensioners as there a re people in employ-me nt to pay for their pensions. Moreover, the reis a strong regional pattern, in that pensioners

represent a substantially larger proportion ofthe po pulatio n in Bosniac areas and RS than inCroat areas (Table 6). This is a particular problembecause unemployment is also substantiallyhigh er in these areas: the ratio of pensioners tothose in work is roughly three times as high inRS an d Bosniac areas as in Croat areas.

Taking account of the high proportion ofworkers in the informal sector, who are likelynot to be paying wage contributions, m akes thesituation considerably worse. In RS, only 30,000

people were employed in the formal sector in1995, one pe r 5.5 pensioners. While figures arenot available for the Federation, the situation islikely to have been broadly similar in Bosniacareas, but substantially better in Croat areas,where the formal sector represents a muchgreater proportion of total employment.

The problem is of a similar order of magni-tude in the case of unem ploym ent benefits: withunemployment at 50-60 per cent, there are atleast as many, and up to 50 per cent more,

people ou t of work than in work, overall. Again,the ratio of unemployed people to those informal employment is considerably worse:assuming a 60 per cent unemployment rate,there would have been 7.5 people unemployedfor every person in formal employment in RS in1995. This implies that, for a pension/benefitsystem fully paid for from formal sector wages,each wag e-earner would have had to pay for 5.5pensions and 7.5 unemployment benefits.Again, the situation is likely to be considerablybetter in Croat (though not Bosniac) areas.

The ratios for Bosniac areas and to a lesserextent RS are likely to have improved signifi-cantly since 1995, to the extent that formal-sector employ men t has increased. How ever, thesituation remains serious. Even on theoptimistic assum ption that half of the w orkers inthe informal sector in 1995 have been

Table 6: Estima ted

Number of pensioners

as % of: population

labour force

total employment

formal employment

number of pensioners relative toand labour ind icators

Bosniac Croatareas areas

200,000 52,000

12 8

50 25

95 35

n/a n/a

population

Rep.Srpska

170,000

14

45

115

550

Sources: Numbers of pensioners and of formal sector workers in RS (in 1995) are taken from World Bank (1996b, pp 24, 26, 28).

Population and labour force are the author's estimates: overall population is based on estimates of the total pre-war population

(4.4m) and numbers of refugees (l-1.4m); and the overall labour-force (1.1m) is taken directly from EU/EBRD/World Bank, 1996,

p93. These figures are combined with the pre-war ethnic composition of the population (from Crnobmja, 19 96, p22, with self-identi-fied Yugoslavs and other ethnic groups divided pro rata between the three areas). Total employment is based on the resulting labour-

force figure s, combined with rough estimates of regional unemployment ra tes (Bosniac area s and RS 60 per cent; Croat areas 2 0

per cent). It should be noted that all the percentage figures are very approximate.

59




reabsorbed into the formal sector, increasingthe formal sector's share of total employmentfrom around 20 per cent to 60 per cent, eachformal-sector worker in RS would be paying for

1.5-2 pensions and 2.5 unemployment benefits.The situation in the Federation is somewhat

better overall, because of the more favourablecircumstances in the Croat areas. On similarassumptions (60 per cent formal-sector employ-ment, with unemployment of 60 per cent in

Bosniac, and 20 per cent in Croat areas), eachformal-sector worker would have to pay for

around 1-1.5 pensions and 1.5-2 unemploy-ment benefits. However, if the benefits systemwere operated on a uniform Entity-wide basis,

the inequality between Bosniac and Croat areaswould imply a considerable transfer of resour-ces between the two areas: taking account also ofwage differentials, Croat areas would makeabout 60 per cent of total contributions, whilereceiving only abo ut 20 per cent of pensions and12-15 per cent of une mp loym ent benefits. Thiscould raise serious political problems.

In view of the very low wage-levels, especiallyin Bosniac areas and RS, the high dependencyratio and unemployment rate means that pay-

roll contributions are very high, even at the

current very low levels of pensions and benefits— 17.5 per cent of wages in Croat areas and 24

per cent in Bosniac areas for pensions alone.Th e curren t situation also makes for a very acutetrade-off between formal-sector incomes and

pensions/benefits: increasing pensions by DM 1would require an increase in the cost to formal-sector workers of DM1.50-2.00 in RS and

DM 1.00-1.50 in the Federation, while a similarincrease in unemployment benefit would costaround DM2.50 per worker in RS and

DM 1.50-2.00 in the Federation. Since the

average net wage in Bosniac areas is DM1 60 pm

(about US$3 per day), any household with threeor more members and only one (average) wageis alread y on or below the absolute poverty line.This means that poverty among pensioners or

those dependen t on other domestically-financedbenefits can only be reduced by increasingpoverty among people in employment.

The problem is not likely to be resolved in the

near future. As noted in Section El, unemploy-ment could well remain at 25-40 per cent in

2000, even without any ret ur n of refugees. Evenif the share of the formal sector in totalemployment could be increased to 80 per centover this period, this would reduce the ratio of

pensioners and the unemployed to those in

formal employment to perhaps 1-1.25 in the

Federation and 1.75-2 in RS. This would be a

substantial improvem ent, but the ratio remainsvery high: in the ord er of 2.5-3 times that in the

UK for the Federation, and 4.5-5 times for RS.

It would also be greatly increa sed by any signifi-cant return of refugees; if all the refugees wereto return during this period, the ratios for

unemployment benefits would be at leastdoubled.

Th e prob lem of inequality between Croatand

Bosniac areas would also be moderated in thisscenario, but would remain serious even withoutthe ret urn of any refugees. On the same assump-tions, and allowing for a halving of income

differentials between the two areas (from 50 percent to 25 per cent), Croat areas would makeabout 40 per cent of contributions, whilereceiving around 20 per cent of pensions and

perh aps 15 per cent of unem ploym ent benefits.

The donors recognise the existence of thisproblem. The EU/EBRD/World Bank (1996)paper observes with reference to 1997 that'there are still significant gaps between the

needs for protecting vu lnerable groups and the

revenues that can possibly be generated by the

domestic economy' (op . cit. pl45 ). However,the

donors appear to under-estimate (or at leastthey under-state) the exte nt of the problem; and

they give little consideration either to its likelydurat ion or to its regional dimension.

In the short term, the gap is to be filled by

donors. However, the level of donor supportenvisaged app ears to be grossly inade quate evento bring p ensioners and the unemployed in RS

and Bosniac areas up to the absolute povertyline. For pensioners alone, this would costsomething in the order of $80m; and the

amou nt needed for the unemployed might be of

a similar order of magnitude. By comparison,the am oun t deem ed necessary in the EU/EBRD/World Bank paper is $20m in 1997. This is

roughly $20-25 per unemployed or pensionedperson, m aking no allowance for other potentialrecipients. This would represent an averageadditiona l income equivalent to about6 per centof the absolute per capita poverty income.Taking account of dependents would reducethis figure still fu rther.

In the medium term, the intention is to shifttowards a self-financing, pay-as-you-go pen sionand benefit system, wholly financed frompayroll contributions. Pensions would be basedon a three-tiered model, with a universalminimum pension, supplemented by earnings-

60



Social dimensions

related pensions (operated at the cantonal orinter-cantonal level in the Federation) andprivate pension schemes. The minimumpensio n would be base d initially on a fixed

contribution per employee, and supplementedby donor funding (World Bank, 1996b, p49),then shifted to a proportional rate, which wouldbe increased in 1997—8, as donor support isphased out.

A ceiling of 15 per cent is envisaged for payrollcontributions to the minimum pension. By 2000,even with unem ploym ent at 25-40 per cent, thisshould be sufficient to provide pensions abovethe absolute poverty line in the Federation. InRS, however, this may well not be the case: withunemp loyment at 25 -40 per cent, even if80 percent of employment were in the formal sector,pensions would reach the absolute poverty lineonly if real wages were increased by 150-200 percent from their current level (Table 7).36 Suchchanges in unem ploym ent and real wages wouldalmost certainly require considerably fastergrowth even than the very high rates currentlyenvisaged by the W orld Bank.

Four further points should be noted in thiscontext:

• The initial use of a fixed contribution per

employee is wholly inappropriate. This ishighly regressive, in that low-wage earnerswould pay a mu ch g reater p ropo rtion of theirincome than those on higher incomes. Theabsolute amounts are very large — about$3-40 0 per em ployee p er year for pensions atthe absolute poverty level — and even theaverage net wage in Bosniac areas is alreadyon the absolute poverty line for a householdof three .

• The proposed rate of 15 per cent is not far

below the curr en t rate of 17.5 per cent in theFed eration , which is described by the B ank as'extraordinarily high ' (World Ban k, 1996b, p48), suggesting that there may be strongpressure to keep wage contributions wellbelow this level in both Entities. This would

imply a commensurate reduction in pensionsbelow th e levels shown in Tab le 7.

• It seems unlikely that pensions and benefits at

absolute poverty levels would be politicallyviable for m ore than a very short p eriod.

• Once again, it is important to note that auniform Federation-wide system wouldimply a substantial net transfer of resourcesfrom Cro at to Bosniac areas, which could be asource of political tension. The alternative ofseparate arrangements for the two areasmight well imply pensions below the povertyline or wage contribu tions in excess of 15 percent in Bosniac areas.

In the long term, a system based on individ-ual accounts is envisaged. However, it seemsunlikely that this will be feasible for somedecades. The transition from a pay-as-you-gosystem to a system base d o n individual accountsrequires pension payments to continue, whilecontribution s are set aside for future paym ents.The resources for pensions thus need to comefrom somewhere else; and it will be a very longtime before additional resources are availableon the necessary scale. There seems little pointin even considering this option at this stage.

No explicit ceiling has been proposed forcontributions to unemp loym ent benefits. However,these will need to remain very high until sub-stantial improvements in wages and employ-me nt have been achieved (Table 8). In 2000, withunem ploym ent at 2 5 ^ 0 p er cent, contributionsin the Federation would remain at around10-25 p er cen t even if the re w ere a 50 per centincrease in real wages and 80 per cent ofem ploym ent w ere in the formal sector.37 In RS,the situation is again much worse: this scenariowould leave contributions at around 35-70 per

cent of wages, which would clearly be at oddswith the objective of reducing wage deductions.Only if unemploym ent w ere reduced to about 10per cent and real wages doubled couldcontributions be reduced below about 10 percent o f wages in RS.

61




Table 7: Pensions at alterna tive levels of unemploym entand real wages (US$ pm)

a. Federation

wages:

real %

change

from

1996

0

25

50

100

150

200

Unemployment (%)

60

31

38

46

61

76

92

50

38

48

58

76

95

114

40

47

58

68

92

114

138

25

58

71

85

114

143

172

10

69

85

103

138

172

206

0

76

95

114

153

191

229

b. Republika Srpska

wages:

real %

change

from

1996

0

25

50

100

150

200

Unemployment (%)

60

8

10

12

15

20

23

50

10

12

14

20

24

29

40

12

14

17

23

29

35

25

14

18

22

29

36

44

10

17

22

26

35

44

52

0

20

24

29

39

49

58

Notes: Author's estimates, assuming 1.1m work-force (one-third in RS), 0.42m pensioners (0.17m in RS), wage con-

tributions at 15 per cent, and administrative costs equal to 10 per cent of the total amount collected. The shaded

area represents pensions less than $1 per day. Base-line wages are estimated on the basis of the net monthly

wage figure given in Table 5, assuming current taxes and wage contributions of 50 per cent.

62



Social dimensions

Table 8: W ag e contributions required for minimum unem ployme ntbenefit coverage (% of wages)

a. Federation

wages:real %changefrom

1996

0

25

50

100

150

200

b. Republika Srpska


1996

0

25

50

100

150

200

Unemployment (%)

60

78

62

5 1

39

3 1

26

50

5 1

4 1

34

26

2 1

17

40

34

28

23

17

13

1 1

25

17

13

1 1

98

74

6

10

6

4

3

3

2

2

0

0

0

0

0

0

0

Unemployment (%)

60

230

183

153

112

92

77

50

153

122

102

77

6 1

5 1

40

102

8 1

68

5 1

4 1

34

25

5 1

4 1

34

26

20

17

10

17

13

1 1

9

7

6

0

0

0

0

0

0

0

Notes: Author's estimates, assuming unemployment benefit of $3 per day ($1 per person per day for a three-person

household). Other assumptions are as in Table 7.

63




£.3 Health and education

Th e health strategy for BH is based on a shift awayfrom the top-heavy pre-war structure, and towards

a system based more heavily on primary heathcare (PHC). Within the F ederation, the inten tionis to shift towards a three-tier, insurance-basedsystem, comparable to that for pensions. Th e firsttier would be an Entity-wide package of uniformbasic services, and the second a supplementarypackage specified by each C anton, both financedfrom payroll taxes and administered at thecantonal level. Th e third tier would be voluntaryprivate health insurance. It remains unclearwhether co-payments (ie additional paymen ts madeat the p oint of service delivery) are envisaged for

either or both of the first wo tiers. Healdi financingarrangements within RS remain unclear.

The shift towards PHC is wholly ap prop riate,and the insurance approach should be viable inthe Bosnian context, provided appropriatearrangements are made for non-wage-earners.However, co-payments should be avoided forthe foreseeable future, as they would riskexcluding a substantial proportion of thepop ulation from access to health services, whileleaving them to pay part of the cost of services

for the better-off through wage contribu tions.Some caution is also needed with respect toprivate health insurance, as there is a risk of unde r-paid health professionals being lured away fromthe public sector by much higher incomes inprivate practice. In the long term , this problemshould be eased to some extent by increasingsalary levels in the public sector, providedsufficient financing is available for recurrentexp end iture. Th e return of health professionalswho have left as refugees would also helpconsiderably; but it remains far from clear that

they will in fact ret ur n on a sub stantial scale.The administration of health services at the

cantonal level could cause significant regionaldisparities in health sectors, with more limitedservices, fewer or lower-quality staff, moresupply shortages, and possibly higher co-payments in poorer areas. The need for 'somemechanism...to provide for cross-cantonalsubsidies as required' is recognised by donors(EU/World Bank, 1996b, pi2). While thisinevitably raises the issue of inter-ethnicresource transfers, such a mechanism would

seem to be essential. (See Dilemma 3.)Financial constraints, particularly on recur-

rent expenditure, are a critical factor in thehealth sector in both Entities. In Croat areas of

the Federation, health services are currentlyfunded primarily through an insurance fund,financed throug h wage contributions at a rate of12.5 pe r cent, which pays mainly for sa laries. In

Bosniac areas and RS, similar insurance fundswere effectively merged with the generalgovernment finances during the war. In allareas, the bulk of recu rrent exp end iture goes tosalary payments; but even so, salaries felldrastically in real terms, and substantial arre arswere accumulated.

The programme of donor support to thehealth sector, designed by the World Bank andWHO, includes as one of its six objectives toprovide support for recurrent expenditure, tobe phased out over three years. However, only

$5m was firmly committed for this purpose in1996 (equivalent to about $1.40 per capita),partly reflecting the 43 per cent shortfall inoverall commitments to the health sector (EU/EBRD/World Bank, 1996, pp 132-3). This is avery small amount relative to the shortfall inresources for recurrent spending on h ealth; andif the planned reduction in support goes aheadfrom this level, su pp ort will become increasinglyinadequate.

As well as the overall level of do no r sup po rt,

its composition is also a problem. Despite thearguments put forward by the World Bank forsalary support, in practice aid has been mainlyfor essential drugs and other medical supplies,while 'no significant su pp ort has been prov idedfor complementing health workers salaries'(EU/EBRD/World Bank, 1996, pi34).

The government has tried to fill the gap(overall and for salaries) from gen eral ba lance ofpayments support; but this again increasespressu re elsewhere in the bu dg et. While this hasper m itted some increase in salary levels, it is still

the case that 'most health workers are paidinsufficient salaries' (EU/EBRD/World Bank,1996, p 140). More g enera lly, it is acknowledgedthat 'budge tary resources for the opera ting costsof health facilities will most likely remaininsufficient in 1997' (EU/EBRD/World Bank,1996, pi37). However, this again appears to bea serious under-statement of the extent of theproblem.

Yugoslavia's pre-war public expenditure onhealth was arou nd $120 per p erson pe r year (in1987, based on data from UN DP, 1992). A

return to this level of funding through pay-rolltaxes would give rise to very h igh rates of wagecontributions: 3 0 ^ 0 p er cent in the Federationand 100-120 per centin RS at prese nt, declining

64



Social dimensions

only to 15-20 per cent and 40-50 per cent in2000. In the long term, assuming real wages100-200 p er cent above cu rre nt levels and 0-10per cent unemployment,the rate would become

sustainable in the Federation, at 5-10 per cent.In RS, however, the rate would remain in theorder of 15-25 per cent, which is clearly unsus-tainable (Ta ble 9).

These figures are probably somewhatexaggerated. Costs should be reduced to someextent by the proposed shift towards primaryhealth care; and (in the short and medium term

only) by low salary levels. Nonetheless, evenallowing for a halving of per capita healthspending from the pre-war Yugoslav level,together with the costs of minimum pensions

Table 9: Wage contributions required to restore pre-war perexpenditure (% of wage s)

a. Federation

wages:real %change

from1996

0

25

50

100

150

200

capita health

Unemployment (%)

60

40

32

27

20

16

13

50

32

26

22

16

13

1 1

40

27

22

18

13

1 1

9

25

22

17

14

1 1

9

7

10

18

14

12

9

7

6

0

16

13

1 1

8

6

5

b. Republika Srpska


1996

0

25

50

100

150

200

Unemployment (%)

60

119

96

80

60

48

40

50

96

76

64

48

38

32

40

80

64

53

40

32

27

25

64

5 1

42

32

25

2 1

10

53

42

35

27

2 1

18

Notes: Author's estimates, based on 1987 health expenditure of $120 per person per year,

as in Table 7 (excluding administrative expenditure, which is included in the $120 figure).

0

48

38

32

24

19

16

Other assumptions are

65




and unemployment benefits, this would leavetax and contribution rates on wages at very highlevels, in marked contrast with the objective of

reducing the rates of wage taxes and

contrib utions. (See Table 10.)

Despite the very conservative assumptionsused in Table 10 (including the absence of an yform of taxation on wages for any other purpose),the short-term levels of wage contributions are

clearly unrealistic in both Entities — around 100

per cent of gross wages in the Federation, and 3-40 0 per cent in RS. In the Federation, a moder-ate reduction from current levels might be

possible in the medium term (ie by about 20 00),a sustainable level of 10-20 per cent being reached

only in the long term (ie when unemployment isreduced to 10 per cent or less and average realwages at least doubled from their current levels.In RS, the situation remains clearly impossibleuntil well after 2000, and even the very opti-mistic assumptions of the long-term scenarioleave wage co ntributions for these three purposesalone at 31-54 per cent, suggesting that a

significant red uc tion is likely to be impossible.

In the education sector, financial supp ort has

been concentrated on school reconstruction(more than 80 per cent of funding), withsmaller-scale support for institution-building,teacher training, educational materials, and

support for students whose education was

disrupted by the war, traumatised children, and

children with special needs. The future direc-tion of education policy remains uncle ar, p artlyas a result of uncertainty about the roles of

different levels of government and limitedcapacity for policy-making at the canton al level.

As in the health sector, teachers salaries are a

serious problem; and this is 'a fundamentalthreat to education in the future' (EU/EBRD/

World Bank, 1996, pi22). This is again partlydue to uncertainty about government roles,largely as a result of inadequate administrativecapacity (and in most cases financial resources)to plan and manage education effectively at the

cantonal level. However, in view of the import-ance attached to this issue, it is far from clear thatthis is an adequate reason for the completeabsence of donor suppo rt to teachers' salaries.

Table 10: Wage contributions required for minimum pensions,unemployment benefits and health expend iture (%)

Federation:

Rep Srpska:

1996:

2000:

long-term:

1996:

2000:

long-term:

Pensions

15

11-13

4 -7

45-56

20-27

15

UnemploymentBenefit

51-78

11-23

0-3

153-230

34-68

0-9

Health

16-20

7-9

5-9

96-119

42-53

16-27

Total

82-113

29-45

9-19

294-405

96-148

31-54

Notes: Author's es timates, based on Tables 7-9, assuming unemployment of 50-60 per cent in the short-term, 25-40 per cent in the medium-term and 0-10 per cent in the long-term; and real wage increases of 0 per cent, 50 per

cent and 100-200 per cent respectively. Pension contributions are assumed to be 15 per cent, subject to a mini-

mum pension of $30 pm and a maximum of $60 pm. Unemployment benefit is set at $90 pm throughout. Annual

health expenditure is set at $60 pc, half of the 1987 Yugoslav lev el.

66



Social dimensions

Th ere is a proposal to develop cost recoveryfor higher education. There is likely to be lessresistance to this in principle than in somecou ntries, as cost recovery was well established

prior to the war. However, there is a risk ofskewing the benefits of higher educationtowards those better able to pay; and in theBosnian context, this would almost certainlyentail an ethnic bias towards the Croats andbetter-off returnees, at least over the mediumterm, due to current income differentials. Thiswould tend to compound and entrench inter-ethnic inequalities, and could also generate anethnic imbalance among decision-makers overtime. It is also impo rtan t to avoid any su bstantialreduction in overall attend anc e of highe r educa-

tion, as the disproportionate number of grad-uates and professionals among refugees is likelyto leave an increasing shortfall in the supply ofhum an capital as the economy recovers, unlessthere is a large-scale return (which is notexpected in the near future).

If the option of cost recovery for highereducation is to be pursued, it should at least bedelayed until the bulk of the population canafford to pay, and then be p hased in gradually,in line with ability to pay among the poorer

ethnic groups. Donors should fill most of thegap between the costs of higher education andthe available government resources and feerevenues. They should also finance ascholarship fund for those wh o are academicallyable to continue with their education, butunable to bear the financial cost (includingforgone income) of doing so. Explicit ethnictargeting should, however, be avoided .

The World Bank also appears to bepromoting the idea of private provision ofeducation and health services at the primarylevel. A recent Bank publication on Bosnia

noted that 'a combination of direct provisionand regulated private provision has generallybeen adop ted' elsew here, althoug h it stops shortof recom me ndin g this as a policy (W orld Bank,1996b, p47). Th is option should be ap proac hedwith great caution .

• In the health sector, there is a risk thatcompetition among private providers (with atleast partly public funding) could bid up thecosts of health inputs — particularly thesalaries of health professionals — relative to a

system based on a public sector monopoly ornear-monopoly. This could substantiallyincrease the cost of publicly-funded healthservices in the longer term. There is also adanger of perverse effects on incentives,reducin g the quality of health services.

• The education sector is critically important tothe developm ent of the political en vironm entover the lo ng te rm . If, as seems likely, educa-tion were provided by organisations moti-vated by nationalist agen das, in the absen ce ofextremely rigorous monitoring and regula-

tion, this could seriously jeop ardis e progresstowards national reconciliation.

It seems unlikely that the re would be substantiallong-term benefits in the private provision ofhealth services. In the case of education, this couldbe positively da ng ero us from a political perspe c-tive. It would be better to avoid any movementtowards private provision in eithe r case.

67



F. Conclusion

There are no easy options in Bosnia andHerzegovina. Th e situation is both very difficulteconomically and extremely complex sociallyand politically. Economic policy has to be form-ulated and implemented within three interlock-ing sets of very tight constrain ts, po litical, social/ethnic, and economic, which seriously limit theavailable options. Where alternative appro ache sto those currently proposed can be envisaged,

they often raise as many problems as theyresolve.

In consequence, this paper has focused onexploring dilemmas, and the options theysuggest, rather than recommending specificpolicy changes. Nine such dilemmas have beendiscussed:

1 Political cond itionality is necessary to achieveany progress against the background ofextreme political inertia in BH. However, itraises the risk of skewing the allocation of a id,

jeopardising economic recovery in RS, andcompromising the donors' neutrality. Acarefully designed, graduated approach issuggested, avoiding conditionality on factorsbeyond the control of the authorities.

2 The currency board arrangem ent is probablynecessary in a highly decentralised politicalsystem with a high level of political conflictand inertia. However, it seriously constrainseconomic policy, and raises the risk of a seriousfinancial crisis after 1998-9. Tw o opt ions are

proposed for further consideration:

• a self-financing system of im po rt tariffs a ndexport subsidies, to achieve effectiveexchange rate adjustment within thecurrency board system; and

• a tempo rary sup por t facility for th eCentral Bank, to be activated in the eventof a crisis beyond the aut horitie s' co ntrol.

3 Regional inequality is a particular problembecause of the political obstacles to inter-regional resource transfers. It is suggested

that part of the anticipated balance-of-

paym ents supp ort could be allocated betw eenEntities and Cantons on the basis of anexplicit formula measuring needs, for use insupport of recurrent exp enditure.

4 Privatisation in the F edera tion raises anumber of issues relating to the role offoreign investment, the pace of privatisation,and the proposed level of decentralisation. A

possible alternative approach is outlined,which would be more gradual and moreclosely related to the needs and circum-stances of individual enterprises, with agreater role for foreign investment wherethere is a substantial need for capital inputsand/or imported technology.

5 Free entry for foreign banks could help toimprove the availability and reduce the costof financial services, but, in the currentcircumstances could give rise to a heavily

foreign-dominated financial sector in thefuture. Delaying this process could reducethis risk, but could also have a negative effecton economic recovery in the short term.

6 The tra de policy envisaged for BH is based ona rapid reduction of tariffs to low, uniformlevels. A case could be made, on both politicaland economic grounds, for delaying thisprocess, to allow a breathing space for thedevelopment of 'infant' and 'convalescent'industries. This w ould also ease financial con-

straints on government spending. However,further consideration is needed of theinteraction of trade policy with the constraintson exchange-rate policies imposed by thecurrency board system.

7 Labour market deregulation may be helpfulto employment creation in the short term , buta gradual and flexible restoration of employ-ment standards (for example, to bringstandards into line with the EU's SocialChapter) may be both feasible and desirable

over the long term.

68



Conclusion

8 Rapid movem ent towards full cost recoveryfor public services and utilities fails to takeaccount of public goods, and could give riseboth to a substantial social impact and an

increase in regional disparities. This suggeststhe need for a more selective and gradualapproach, and for social-pricing mechan-isms. However, this would take resourcesaway from other uses and entail potentiallyproblematic inter-ethnic resource transfers.

9 Employment creation is recognised by donorsas a fundamental objective, essential to socialand political stability. However, the

reconstruction process to date has had a

limited impact. There might be a case for a

limited and selective shift of exte rnal resou rcestowards job-creation throu gh public worksprojects, particularly if their cost could be

reduced from the current very high level.More generally, the implications for

employment should be a key factor taken intoaccount in assessing the alternativeapproaches to privatisation and trade policyoutlined in Dilemmas 4 and 6.

A further serious problem arises from the cost

of minimum pensions, unem ployme nt benefits,and health services. These cannot be adequatelyfinanced from wage contributions in the shortterm, or, in RS, in the medium term. Even witha reduction of unemp loyment to 10 per cent anda doubling of real wage levels, the wagecontributions required to sustain pensions and

unemployment benefits on the absolute povertyline and half the pre-war level of healthexpenditure would be more than 50 per cent in

RS. This is clearly at odds with the objective of

reducing wage-based social contributions and

taxes.Greater donor support for safety-nets and

recurrent expenditure on health and educationis essential in the short term, extending wellbeyond 2000 in RS. Such support has beentotally inadequate so far, and will becomeprogressively more so if the donors' plan to

phase it out over three years is implemented.

Further consideration needs to be given to the

establishment of a sustainable system of

financing for social expenditures over the longterm.

Clearly, the economic, social, and politicalproblems of BH are intractable in the extreme;and they are compounded by the criticaltensions which remain am ong the various eth nicgroups, and particularly their political leaders.The current institutional arrangements haveprevented these tensions from turning intorenewed physical violence; but they haveneither resolved the tensions nor allowed themto be set aside so that norm al political proc essesand policy-making can be resumed.

Any progress towards resolving the problemswill require a great deal of effort and

pragmatism on all sides. This includes the

donors (and NGO advocates) as well as the

authorities and political agents within BH itself.

At present, the movement of the authorities (or

at least the Serb and Croat authorities) towardsmoderation and pragmatism remains limited.Th ere also seems to be a question-mark over the

pragmatism of the donors, in that the policiesthey are advocating for BH closely resemble the

free-market model which conforms to theirideological position. These policies are not

necessarily wrong, but nor should it be

automatically assumed that they are necessarilyright.

The need for pragmatism extends wellbeyond the realm of economic policy, in partic-ular to the long-term political future of BH.

There is a tendency on the part of the majordonors — and international NGOs — to assumethat the long-term solution to the Bosnianproblem is a reunification of the country as a

single m ulti-ethnic state. This is neither valid as

an assumption, nor defensible as an objective initself. If the ultimate objective is a peaceful and

prosperous future for the people of BH, the

country's political future must be subject to

critical and objective evaluation.

This raises the most fundamental and

intractable dilemma of all: how many countriesshould Bosnia-Herzegovina be?

69




Dilemma 10: Reintegration or partition?

The international community appears to be

firmly com mitted to the notion of a single BH ,with the ethnic integration of the pre-1991period restored. There are a number ofpossible reasons for this, some of which arebetter than others.

a. Acce pting partition would am ount to anofficial end orse me nt of the results of ethniccleansing. This would not only be verydifficult politically and raise major moralissues, but could also provide an incentivefor ethnic cleansing elsewhere.

b. More generally, the concept of ethnic parti-tion is at od ds with Western liberal prefer-ences for coexistence and m ulticulturalism.

c. Reunification and ethnic reintegration maybe seen as providing a more acceptable(and expeditious) basis for the return ofrefugees. Germany, for example, inrepatriating refugees, has accepted theprinciple that people should not at presentbe returned to areas where they would be

members of the ethnic minority. Asuccessful transition to multi-ethnicitywould remove this constraint, and allow afaster (and less politically embarrassing)repatriation. This might also be possiblewith partition; but would require a veryexplicit acceptance of the results of ethniccleansing. It could also result in disputesabout the morality or legality of forcingpeople to return to ethnically-defined areasrather than areas of origin; and it would

leave no clearly-defined destination for asubstantial proportion of the refugeepopulation (for example, people of mixedpare ntag e, mixed couples, and members ofother ethnic groups).

d. A single, unified Bosnia may be seen asmo re economically viable than two or thre eseparate states: even with the return of allthe refugees, the total population would beonly 4.5m.

e. Bosnia's economy and infrastructure arebased o n the concept of BH as a single entitywithin the Yugoslav Federation; and the

current boundary between RS and the

Federation is, from this point of view,essentially arbitrary. Partition wouldtherefore be potentially disruptive ineconomic terms.

f. Th e complex pattern of the country's ethnicdivision would make separationcomplicated: RS is virtually divided into twoparts, linked only by the Brcko corridor (andonly subject to the results of arbitration38),while the F ederation is a complex patchworkof Croat and Muslim areas. Separation

would mean a choice between non-contiguous C roatian and Bosniac territories,and at best a very tenuous link betweenWestern RS and the rest of an enlargedSerbia; or further ethnic relocation. Theformer would entail considerable adminis-trative complexity, and would require a hig hlevel of coop eration between die states (witha serious risk of renewed conflict if it failed).The ethnic relocation option would almostcertainly be unacceptable internationally.

g. In the case of RS, unification with Serbiawould raise particular problems, as thewestern half of the Entity has closereconom ic ties with Croatia th an with Serbia(or with the rest of Bosnia). Coupled withexisting political differences between thetwo halves of RS and the tenuousgeographical linkage between them, theprospect of partition could give rise toserious political divisions between the twoareas (and within the western area), whichcould undermine the stability of BH as awhole.

h. If contro l of the Croa tian p ort of Ploce wereretained by the Croatian component of apartitioned BH, the Muslim rump wouldbecome land-locked, weakening itsstrategic and economic position, andleaving it critically dependent on Croatiangoodwill.

The practical issues (e-h) carry someweight, as does the concern about endorsing

ethnic cleansing (a). Other aspects, however,are less clear-cut. The refugee issue (c), in

70



Conclusion

Dilemm a 10: Reintegration or partition? (continued)

particular, is primarily a matter of political

convenience to donor governments. There issome basis for the 'small country' argument(d). However, there are a number oflimitations to this view.

• It has been argued (by The Economist, forexample) that the process of globalisationmakes smaller countries more viable thanhas been the case historically.

• If the alternative to reunification of BHwere the accretion of RS to Serbia and ofHerceg-Bosna to Croatia, the 'small

country' argument would apply only to theremaining Muslim area.

• Economic viability depends to a great extent

on political stability and the effectiveness ofpublic institutions; and it would a t best takesome time to re-establish these conditionsin a reunified BH . It is relevant to considerwhether this could be achieved morequickly by pa rtition.

There is also some danger in seeking toimpose fashionable Western views of multi-

ethnicity and multiculturalism (b) in a verydifferent context where these concepts may beinappropriate. The fact that BH was asuccessful multi-ethnic entity before 1991does n ot mean that it necessarily can be aga in.The war, by its very nature, has had aprofound effect on the relations between theethnic groups. A retu rn to multi-ethnicity canonly be achieved to the extent that theresentments generated are resolved; but theseresentments are acute, and their resolutionwill take a considerable time.

There are some factors which couldcontribute to this process — for example amore ind epen dent m edia, the replacement ofnationalists with moderates in the politicalleadership, and a rapid and broadly-basedeconomic recovery. However, these develop-ments are not within the d onors' control, andare by no means assured.

Even if the media were more inde pen den t,they would almost certainly continue to beorganised on an e thnic basis, in terms of botheditorial control and audience. Given thedepth of the resentment, and the cumulativeeffects of nearly a decade of strident

nationalist pro paga nda, it can by no m eans be

assumed that greater independence wouldmean greater moderation.

A change in political leadership wouldrequ ire a shift in popular opinio n; and as longas the media remain nationalistic, any shifttowards moderation will at best be slow. Thecu rre nt political dominanc e of nationalists willfurther imp ede this process, as they will haveample op portunity to find or create sources ofconflict with other ethnic grou ps. Th e c urre ntnationalist authorities will remain in powerfor some years; and th e extent of movem ent inpopular opinion required to remove them isconsiderable.

As discussed earlier in this paper, a rapideconomic recovery cannot by any means betaken for granted; it could be reversed inthree to five years (ie rough ly at the time of thenext elections), as international support forrecons truction winds dow n; and it is far fromclear that employment generation and socialsafety nets will be adequate to provide a sub-stantial improvement in the standard of living

for much of the population. Improving livingstand ards are in any case only a necessary, andby no means a sufficient, condition forreconciliation.

Ultimately, reunification will only besuccessful if it is actively endorsed by anoverwhelming majority of all ethnic groups;and the factors considere d above will be majorimpediments to such endorsement. It will atbest be some time — possibly measurable indecades rather than years — beforereunification can be achieved.

If reunification can be achieved at all, it willrequire an immensely difficult balancing act.On the one hand, the institutional arrange-ments adopted must be popularly acceptablein an environment of ethnic mistrust andhostility; and be able to function effectively inan environment of ethnic separation. At thesame time, however, they must lead towardsreconciliation and moderation, rather thanpolarisation and nationalism; and fosteridentification with former ethnic enemies inthe same country rath er than with mem bers ofthe same ethnic group in neighbouringcountries.

71




Dilemma 10: Reintegration or partition? (continued)

To be politically acceptable, the structuremust allow the members of each ethnic groupto remain m ore or less autonom ous at the levelof effective power. At present, the variabledecentralisation enshrined in the Daytonarrangements secures this. To functioneffectively, however, the institutional arrange-ments require a high level of cooperationbetween the separately elected ethnic leader-ships. To ensure cooperation, the Daytonarrangements rely to a great extent on

sanctions against non-compliance — primarilythe withholding of financial support.

There is a potential contradiction here.Ethnic autonomy directly reinforces ethnicrather than national identity; and it entails analmost complete separation of politicalsystems. Rather than fostering cooperation orreconciliation, it reinforces the tendencytowards ethnic and political polarisation (asdemonstrated by the September ] 996 electionresults). This in turn weakens cooperationand thus impe des the operation of institutions

(as amply demonstrated by the currentproblems, both between RS and theFederation and within the Federation itself).

This goes to the heart of the Bosniandilemma. Successful reintegration requiresconditions which cannot realistically be expectedin a segregated society with a high level ofmutual distrust between ethnic groups (or atbetween their elected political leaders). Inothe r w ords, w hile it is possible to envisage areintegrated society which would be stable andsustainable, it may well not be possible to

engineer the transition from the current situ-ation to that ideal within a time-frame whichwould be politically acceptable to the d onor s.

Furth er pro blem s may arise from attemptsto enforce cooperation within this framework.Th e failure of cooperation requ ires the appli-cation of the available sanctions — that is, thewithdrawal of external suppo rt. How ever, bytheir nature, these sanctions impede econ-omic recovery in the recalcitrant areas; andthis will exacerbate tensions. In other words,the mechanism for achieving cooperationmay itself compound any failure of coopera-tion over the long term. However, even if

cond itionally can enforce cooperation, in theabsence of genuine political will, cooperationis likely to occu r only at the formal evel. This willmake at best a limited con tribution to the effec-tive operation of the institutional arrangeme nts.

This implies a need for political condition-ality to be closely targeted, carefully grad-uated, and designed on the basis of a clearunderstanding of the political dynamics. (SeeDilemma 1.)

At the same time, the reliance on donor

conditionality to enforce cooperation mayjeopardise the donors' (perceived) neutrality,in that it requires them to associate themselvesclearly and explicitly with the principle ofreunification — in other words, to alignthemselves closely with the Muslim objective,in opposition to those of the Serb and Croatleaderships.

Thu s it is by no means clear that the pre sentarrangements fulfil all of the conditionsnecessary to rein tegration. Nonetheless, theymay be the best that can be achieved, to the

extent that the problems noted above arelargely inhe ren t, and may well be insoluble. Inany event, reope ning the process to redesignthe institutions is not a realistic option.

It is also unrealistic to expect the d ono rs toad opt p artition as an ultimate objective. Whilethis would be an interna lly consistent position,to the extent that do nor s uppo rt would almostcertainly result in partition, it would be indirect contradiction to the principles under-lying the whole process of mediation, andwould represent a clear, though implicit,

endorsement of ethnic cleansing. This isneither likely nor desirable.

Neither is neutrality a viable option, so longas don ors a re seeking to influence policy. Thedesign of institutions and policies, and of thewhole process of reconstruction and rehabili-tation, is critically dependent on the nature ofthe future BH which they are to serve. Theyneed, to a great extent, to be designed on thebasis either of reunification or of partition . Inthe cu rre nt state of political inertia, opting outof the policy-making and institution-formingprocess would merely be a particularly messyway of bringing about partition.

72



Conclusion

Dilemm a 10: R eintegration or partition? (continued)

This leaves the international communitywith two viable alternatives:

• to design their interventions with the specificobjective of pushing BH towards reunifi-cation, and/or on the assumption that thecountry will ultimately be reunit ed; or

• to design policies in this way, whileplanning explicitly for the possibility ofultimate partition.

At present, donors appear to be followingthe former approach, with little outward signof contingency planning: to the extent thatthe re is a contingency plan, it appears to take

the form of a withdrawal of support. (Ifcontingency planning were taking place,however, it is unlikely that it would be o vert,since openly plan ning for par tition w ould risklegitimising and encouraging separatisttendencies, ultimately making partition morelikely.)

The impression created by this single-minded pursuit of reunification is of a rathersuperficial and patronising view of thedivisions generated by the war, like parents

telling squabbling children to be friendsagain. This is not obviously helpful, and coulditself alienate nationalist leaderships and/ortheir supporters.

73



Notes

1 This is a very brief an d grossly over-simplifiedaccount of the war and the factors leading upto it. Readers are referred to the substantialnumber of books on the subject, such asMalcolm (1994), Bennett (1995) andCrnobrnja(1996).

2 Th ere were some tensions before this, egCroat fears of Serb expansionism under KingAlexander in the 1930s, which was an

important factor motivating the Ustasas.According to Benn ett, the notion of centuriesof animosity between Serbs and Croats is a'historical myth', based on m edia prop agand arather than actual history: 'until this century,most Serbs and Croats had virtually no contactwith each oth er' (Bennett, 1995, p6). BosnianMuslims were at most collaborators with theSerbs' actual enemies (the Turks), rather thanthere being any direct conflict between them.The Second World War was not the beginningof ethnic tensions in the region; but it

represented (with the benefit of hindsight) thepoint of no retu rn.

3 The figures n this paragraph and the next aretaken from Crnobrnja (1996), pp2 2-3 .

4 Fox and Wallich (1997, p3) give very differentfigures for population: 0.7m in RS and 2.9m inthe Federation (2.5m in Bosniac and 0.4m inCroat areas). However, these figures seemdifficult to accou nt for in terms of the p re-warethnic distribution of the population. Accord-ing to Crnobrnja, pre-war census data (basedon self-identification) show 44 per cent of thepopulation (1.9m people) to have beenMuslims, 31.5 per cent (1.4m) Serbs, 17 percent (0.7m) Cro ats, leaving 7.5 per cent (0.3m)other ethnic groups and self-identifiedYugoslavs. If the Muslim share of deaths andrefugees is equivalent to their pre-war share ofthe population (which is probably a conserv-ative assum ption), the Fox and W allich figureswould be correct only if there were 1.2mrefugees from other Republics and membersof non-Muslim ethnic groups resident in

Bosniac areas tha t is, if these gro ups constituteroughly half the current population in theseareas. Since the total number of refugees

entering BH is estimated at 2-300,000, andthere were about 300,000 Yugoslavs andmem bers of other ethnic grou ps pre-war, thiswould imply that, at the very least,25 per centof the current population of Bosniac areaswould be Serbs and Croats; and, since thedifference between the pre-war Croat popu-lation and the estimated cu rrent pop ulationofCroat areas is only 300,000, at least half of

these people would be self-identified Serbs.Th ese imp lications seem implausible.

5 Following the de-collectivisation of agricul-ture in 1953, private land holdings werelimited to 10 ha .

6 GSP (gross social prod uct) is the nearestequivalent to GDP in the national accountingsystem used in most socialist countries,including the former Yugoslavia.

7 This Section is based mainly on EIU (1996a,1996b, 1997a, 1997b), supplemented byvarious articles from The Economist.

8 Th e US, in particular, has linked the FRY'sreentry to the IMF to progress on the extradi-tion of war criminals and its policy towardsKosovo. Progress on both has so far beenlimited.

9 'A Co untry with a Vision: Former YugoslavRepublic of Macedonia Sets Regional Example',World Bank News, 15 February 1996, pp 3-4.

10 As the name suggests, the final beneficiaryprinciple means that the debt is divided upaccording to which Republic, Entity, etc wasthe ultimate recipient of each individual loan.

11 The HIPC Initiative phases debt reductionover a period of six years, each phase beingconditional on a favourable track-record ofstructural adjustment policies. The othercomplication of the HIPC Initiative withregard to BH is that, while its per capitaincome is currently low enough to qualify, ifeconomic recovery proceeds at the ratecurrently hoped for, it would exceed theincome threshold before the final stage wasreached.

12 An agreement-in-prin ciple, in this context, isan agreement between the government of acountry and an appointed committee of its

74



Notes

commercial bank creditors. The agreement

cannot be finalised until the terms agreed in

principle have been accepted by creditors

holding a specified percentage of the debt -

typically abo ut 95 per cent.13 Strictly spea king , this wou ld almost certainly

be illegal without a formal waiver by the banks

of the negative pledge clauses and sharing

provisions in their loan contracts; and the FRY

would once again h ave a sufficient share of th e

debts to block such a waiver. However,

mechanisms are readily available to purchase

debts informally; and some Latin American

govern me nts are believed to have u nde rtake n

such transactions during the 1980s. While the

FRY could m ake a legal challeng e, it would be

difficult to prove the offence, if it were wellexecuted.

14 It should be noted that th e following

discussion is entirely speculative in na tur e.

15 Net pre sent value (NPV) is a measu re of deb t

which takes account of its terms as well as its

absolute amount. For a loan at the m arket rate

of interest, the NPV is the same as its face

value, but it becomes progressively smaller as

the interest rate declines. Increasing the

repayment period of a loan at a concessional

interest rate also reduces its NPV, as the

benefits of the lower interest rate are enjoyedfor longer.

16 Th is and the figures in the nex t two

paragraphs are the author 's estimates, based

on World Bank (1996b), Tables 5.2 and 5.3.

Because of the rather unhelpful presentation

of the Bank's projections, the figures given

should be regarded as very app roxim ate.

17 A conspicuous exception is Brunei, which

pledged only $2m, but committed $ 18.7m, of

which $ 16.7m had been disbursed by October

1996.

18 World Bank financing was limited to $8m

mobility aids and training for die disabled and

$2m for institutional and implementation support.

19 Compared with current salary levels. This

estimate is based on World Bank estimates of

staff requirements (World Bank, 1996a,

Annex II Appendix 2), assuming a 50-50 split

of staff between Bosniac and Croa t areas, and a

900 p er cen t increase in salaries of those from

Bosniac areas.

20 As noted above, however, that this assessment

is based on population figures very differentfrom those used elsewhere in this pa pe r (0.7m

in RS and 2.9m in the Federation, as against

1.2m and 2.3m respectively). See footnote 5.

21 This is illustrated by the dispute between

Bosniacs and Croats in Mostar following the

Sep tem ber 1996 elections: the EU 's ability to

secure a favourable outcome was seriously

compromised by the fact that virtually all itsaid had already been disbursed, so that the

Croats had noth ing to lose by intransigen ce.

22 This is prop osed by the World Bank (1997b,

pp xxv, 37, 42), although they appear to

envisage subsidies from domestic resources

rather than aid receipts. The current status of

this proposal is unclear.

23 EU/W orld Bank (1996b), p3 .

24 Th e cost of the EPW prog ramm e per perso n-

year of employment is estimated at $5,440.

(SeeSection El.)

25 The Fund for War Veterans, Invalids andFamilies of Deceased (20 per cent); the Fund

for the Protection of War Veterans (8 per

cent); the Retirement Fund (7 per cent); the

Education Fund (5 per cent); the Family

Plann ing F und (5 per cent); the Popu lation

Resettlement Fund (5 per cent); and the

Restitution F und (5 percen t) .

26 Current and former soldiers are important

holders of public sector liabilities, in the form

of salary arr ear s.

27 A bon d which is never repaid, but on which a

constan t stream of income is paid.28 This prob lem is much less serious in RS , where

69 per cent of inputs are imported, but 53 per

cent of output is exported.

29 Th e SME funds were 'almost the only sources

of cash resources for enterprises in 1996'

(EU/EBRD/Wo rld Bank, 1996, p83).

30 As discussed in Section B2, there is a risk tha t

the W orld Bank 's balance of paym ents projec-

tions will not be achieved, which could

seriously dam age BH's external deb t situation.

Strengthening the balance of payments

through changes in trade policies could help

to avert this problem .

31 The likely ethnic locus of pressure for

continued protection in unclear: the relative

strength of producers in the Croat areas

suggests that they would be best able to take

advantage of the resulting opportunities in the

short term; but equally, if this approach were

successful in promoting local production,

producers in RS and Bosniac areas might be

more d epe nde nt on the maintenance of tariffs

for their continued viability.

32 According to EU/EBRD/World Bank (1996,

p93), 'About 425,000 soldiers have been o r are

being demobilized.' If, say, 300,000 were

75




demobilised at a constant rate during the

course of 1996, the average employment

figure for the year would need to be adjusted

downwards by 150,000. This would leave a

further 125,000 jobs to be shed by further

demobilisation, so that employment would be

reduced by 275,000 (25 per cent of the labour-

force) from its 1996 level overall.

33 Up to mid-November, 38 sub-projects had

been financed at a total cost of $2.96m; and

6,530 person-months of employment had

been created. This implies an average cost of

$5,440 per person-year. The total value of the

project is $41.4m, spread over two years,

implying the creation of an average of 3,800

jobs over that period, from a labour force of1.1m. At that time, three months after the

project had come into operation, none of the

non-World Bank-managed funds had been

identified, leaving a financing gap of $31 m, or

three-quarters of its total value (World Bank,

1996a, pi5) .

34 Aid disbursements up to October 1996 were

$720m (EU/World Bank, 1996a, Table 3), an

annual rate of $864m. It is estimated that the

reconstruction programme created 250,000

jobs at its peak (EU/EBRD/World Bank, 1996,

p94). Again assuming an average level of

employment-creation for the year as a whole of

half this level, this implies an average cost per

person-year in the order of $7,000.

35 In Croat areas, no pension payments were

made in 1992-3; in Bosniac areas, payments

were six months in arrears in early 1996

(World Bank, 1996b, p48).

36 These figures are affected significantly if the

population figures in Fox and Wallich (1997)

are used. Assuming that the numbers of

pensioners in World Bank (1996b) remain

valid, this would increase the pension levels in

Table 7 by about 10 per cent in the Federation,

but reduce them by about 40 per cent in RS.37 The Fox and Wallich (1997) population

estimates would increase the cost of a given

level of wage contributions in the Federation

for a given level of unemployment benefit by

about one-eighth, because they imply a

substantially greater share of population in the

(lower-income) Bosniac areas. The RS figures

are unaffected.

38 The conclusion of the Brcko arbitration has

been deferred until February 1998.

76



References

Bennett, C. (1995) Yugoslavia's Bloody Collapse:

Causes, Course and Consequences. London: Hurs t .

Crnob rnja, M. (1996) Th e Yugoslav Drama. Second

edition. London: I .B.Tauris .

Cviic, C. (1995) Remaking the Balkans. London:

Pinter/Royal Institute of International Affairs,

Chatham House Papers .

EBRD (1996) Bosnia-Herzegovina: First PhaseReconstruction Programme. March.

EIU (1996a)

EIU (1996b)

EIU (1997a)

EIU(1997b)

Fox, W. and Wallich, C. (1997) Fiscal Federalism in

Bosnia-Herzegovina: the Dayton Ch allenge. WorldBank Policy Research Working Paper No. 1714,

January .

EU/World Bank (1996a) Bosnia and Herzegovina:

the Priority Reconstruction Program —from Emergency

to Sustainabilily. Volume 1: an Overview of Progress

an d Challenges for 1997. European Commission

and Central European Department, World Bank,

November .

EU/World Bank (1996b) Bosnia and Herzegovina:

the Priority Reconstruction Program -from Emergency toSustainability. Volume 2: The Economic Vision and

Near-Term Tasks towards Sustainable Recovery an d

Growth. European Commission and Central

Europea n D epar tment , World Bank, November .

EU/EBRD/World Bank (1996) Bosnia and

Herzegovina: the P riority Reconstruction Program —

from Emergen cy to Sustainability. Volume 3:

Reconstruction Sector Reports. European

Commission, European Bank for Reconstruction

and Development and Central European

Department, World Bank, November.

IMF (1993) International Financial Statistics, 1993.

Washington D.C.: International Monetary Fund.

Malcolm, N. (1994) Bosnia: a Short History. London:

Macmillan.

OHR/EU/UST (1997) Bosnia and Herzegovina:

Economic Policy Priorities. Office of the High

Representative, European Commission and US

Treasury, January.

UNDP (1992) Human Development Report, 1992.

New York: Oxford University Press.

Vojnic, D. (1995) 'Disparity and Disintegration:

Th e Econom ic Dimen sion of Yugoslavia's Dem ise'.

In Akhav an, P. an d How se, R. (eds.): Yugoslavia: the

Former and Future — Reflections by Scholars from the

Region. Geneva: UN Research Institute for Social

Development.

Woodward, D. (1996) 'The IMF, the World Bank

and Econom ic Policy in Rwanda: Economic, Socialand Political Implications'. Mimeo, Oxfam,

September .

World Bank (1989) Trends in Developing Economies,

1989. Washington D.C.: World Bank.

World Bank (1990a) World Debt Tables, 1990-91.

Washington D .C.: World Bank.

World Bank (1990b) World Development Report,

1990. New York: Oxford University Press.

World Bank (1992a) Trends in DevelopingEconomies, 1992. Washington D.C.: World Bank.

World Bank (1992b) World Development Report,

1992. New York: Oxford University Press.

World Bank (1993) World Debt Tables, 1993.

Washington D .C.: World Bank.

World Bank (1995) Republic of Bosnia and

Herzegovina: Reconstruction at a Glance - Clearing of

Landmines. December .

77




World Bank (1996a) Bosnia and Herzegovina: World Bank (1996c) World Debt Tables, 1996.

Priority Reconstruction Projects Update. World Bank Washington D .C : World Bank.

Resident Mission, Bosnia and Herzegovina,

Sarajevo, November.

World Bank (1996b) Bosnia and Herzegovina:

toward Economic Recovery. Washington D .C : World

Bank, Country Study.

Documents

The IMF, the World Bank, and Economic Policy in Bosnia