eca report on financial assistance

8/15/2019 ECA Report on Financial Assistance

1/120

Special Report Financial assistanceprovided to countriesin difficulties

EN 2015 NO

EUROPEANCOURTOF AUDITORS


2/120

EUROPEAN COURT OF AUDITORS12, rue Alcide De Gasperi1615 LuxembourgLUXEMBOURG

Tel. +352 4398-1

Email: [email protected] Internet: http://eca.europa.eu

Twitter: @EUAuditorsECAYouTube: EUAuditorsECA

More information on the European Union is available on the internet ( http://europa.eu ).

Luxembourg: Publications Office of the European Union, 2015

Print ISBN 978-92-872-3173-4 ISSN 1831-0834 doi:10.2865/278824 QJ-AB-15-016-EN-CPDF ISBN 978-92-872-3224-3 ISSN 1977-5679 doi:10.2865/66883 QJ-AB-15-016-EN-NEPUB ISBN 978-92-872-3172-7 ISSN 1977-5679 doi:10.2865/15592 QJ-AB-15-016-EN-E

© European Union, 2015Reproduction is authorised provided the source is acknowledged.

Printed in Luxembourg

mailto:[email protected]://eca.europa.eu/http://europa.eu/http://europa.eu/http://eca.europa.eu/mailto:[email protected]


3/120

Financial assistanceprovided to countriesin difficulties

(pursuant to Article 287(4), second subparagraph, TFEU)

Special Report

EN 2015 NO


4/120

02Audit team

The ECA’s special reports set out the results of its performance and compliance audits of specic budgetary areas ormanagement topics. The ECA selects and designs these audit tasks to be of maximum impact by considering the risksto performance or compliance, the level of income or spending involved, forthcoming developments and political andpublic interest.

This performance audit was produced by Audit Chamber IV — headed by ECA Member Milan Martin Cvikl — which spe-cialises in auditing revenue, research and internal policies, and European Union’s institutions and bodies. The audit wasled by ECA Member Baudilio Tomé Muguruza, supported by the head of his private office, Daniel Costa de Magalhães andIgnacio García de Parada, attaché; Zacharias Kolias, director; Josef Jindra, team leader; Giuseppe Diana, Marco Fians,Daniela Hristova, Didier Lebrun, Ioanna Metaxopoulou, Adrian Savin, Kristian Sniter, auditors.

From left to right: D. Lebrun, I. García de Parada, K. Sniter, J. Jindra, G. Diana,B. Tomé Muguruza, A. Savin, D. Hristova, Z. Kolias, M. Fians, D. Costa de Magalhães.


5/120

03Contents

Paragraph

Abbreviations

I-XIX Executive summary

1-26 Introduction

1-7 The sovereign debt crisis in the EU Member States

8-18 Design and implementation

8-10 Programme design

11-15 Conditions

16-18 Implementation

19-26 Financial resources

19-23 Programme nancing

24-26 Borrowing on the markets

27-32 Audit approach

33-47 Part I — The Commission was unprepared for the magnitude of the crisis thatbroke out

33-44 The Commission underestimated the extent of the fiscal imbalances in the run-up to the crisis

35-40 The Commission estimated the public budgets to be stronger than they actually were

41-44 Assessments underestimated the potential scal risks

45-47 As a result the Commission was not ready for the programme management


6/120

04Contents

48-114 Part II — Processes were generally weak

50-96 The Commission’s processes did not always incorporate sufficient checks and balances

52-54 The Commission was mostly thorough in obtaining the necessary information for monitoring purposes

55-70 Programme forecasts provide for broad checks of consistency in programme design, but with many caveatswhich are not veried

71-75 Programme features were generally soundly based, but important records were sometimes missing

76-79 The Commission occasionally departed from Council decisions

80-92 Countries were not treated in the same way

93-96 Programme management weaknesses due to weak controls

97-107 Programme documents had improved with time but still displayed shortcomings

97-99 Improvements in the quality of programme documents

100-107 Weaknesses in compliance monitoring and reporting

108-114 Cooperation with the other partners was informal only

115-136 Part III — Borrowing met nancing needs even though circumstances initiallymade it difficult to always abide by best practice

117-122 Bond issues were well subscribed

117-118 Cost of debt in line with peer levels

119-122 Pricing levels at issuance sometimes higher than initial guidance from banks

123-136 Circumstances initially made it difficult to always abide by best practices

137-178 Part IV — Programmes met their objectives

138-143 No need to increase overall funding

138-141 Funding envelopes proved sufficient, if tight in two Member States

142-143 Certain Member States received more advantageous nancial terms than provided for the initial design


7/120

05Contents

144-154 Updated deficit targets mostly but not always sustainably met

144-148 With some exceptions, revised annual decit targets were met

149-152 All countries improved their scal stance, but to varying degrees

153-154 Countries used one-off measures

155-164 Compliance with the conditions was high

155-158 Member States eventually complied with at least 80 % of conditions

159-164 Programmes prompted reforms, but occasionally the desired effects did not materialise

165-178 How well did countries increase their competitiveness?

165-172 Competitiveness gains, but to varying degrees

173-174 Internal devaluation worked through the adjustment of the wages of the newly hired

175-178 Foreign exchange loans remained a challenge

179-204 Conclusions and recommendations

179-180 Pre-crisis surveillance

181-192 Programme management processes

193-194 Borrowing operations

195-204 Programme achievements

Annex I — Reasons motivating the request for assistance

Annex II — The lending facilities for the ve countries

Annex III — Audit methodology

Reply of the Commission


8/120

06Abbreviations

Council : Council of the European Union — the text uses this term as a shortcut for any Member Statedecision-making body

ECB: European Central Bank

EFC: Economic and Financial Committee

EFSF: European Financial Stability Facility

EFSM: European Financial Stabilisation Mechanism

EMTNs: Euro medium-term notes

ESA: European System of Accounts

EWG: Eurogroup Working Group

IMF: International Monetary Fund

MoU : Memorandum of understanding

SGP : Stability and Growth Pact


9/120

07Executivesummary

ISeven years ago, Europe faced a financial crisis whichturned into a sovereign debt crisis. The sovereign debtcrisis was a consequence of various factors, includingweak banking supervision, poor fiscal policies and thedifficulties experienced by large financial institutions(and the consequent bailout costs borne by the gen-eral public). The crisis swept across EU Member Statesin two waves, first affecting the non-euro area coun-tries in 2008-2009 and later spreading to the euro areaitself. In total, eight EU Member States were forced toseek macrofinancial assistance.

IIBut was this turmoil well resolved? We analyse theCommission’s management of the financial assistanceprovided to five Member States (Hungary, Latvia,Romania, Ireland and Portugal).

Some warning signs went unnoticed

IIIIt is important to remember that before the crisis therewas already a framework geared towards monitoringthe Member States’ budgets. Warning the Councilabout the mounting fiscal imbalances was the Com-mission’s responsibility. The Commission was notprepared for the first requests for financial assistance.

IVWe found that the European Commission estimatedthe countries’ public budgets to be stronger than theyactually turned out to be. An important weakness ofthe Commission’s assessments prior to 2009 was thelack of reporting on the build-up of contingent public-sector liabilities, which often became real liabilitiesduring the crisis. Nor did the Commission pay suf-ficient attention to the link between large foreignfinancial flows, the health status of the banks and,ultimately, government finances.

VReforms to the Stability and Growth Pact in 2011,2013 and 2014 sought to address the weaknesses ofthe pre-crisis period by introducing greater macro-economic surveillance. However, back in 2008 theCommission found itself unprepared to manage thefinancial assistance when countries started presentingtheir requests.

Standing up to the challenge

Playing catch-up

VI The Commission managed to take on its new pro-gramme management duties, which included hold-ing talks with the national authorities, preparingprogramme forecasts, financing gap estimates andidentifying the necessary reforms. Given the initialtime constraints and limited relevant experience, thiswas an achievement.


10/120

Executive summary 08

VII The Commission was mostly thorough in obtaining theinformation it needed. It has increasingly marshalledinternal expertise and engaged with a wide range ofstakeholders in the countries which appealed for help.

Complex tools

VIII The production of macroeconomic and deficit fore -casts was not a new activity. The Commission used anexisting and rather cumbersome spreadsheet-basedforecasting tool. Quality control was limited mainly toreconciling the various parts of the forecasts, with noenquiry by management into the reasoning behindthe forecasting assumptions. It was very difficult toassess the plausibility of key assumptions such as fiscalmultipliers, not only in any subsequent review butalso, by management, during the actual production offorecasts.

Teething problems

IX The assistance programmes were mostly soundlybased, given the prevailing EU-level policies and theeconomic knowledge at that time. But a commonweakness of the Commission process was the lack ofdocumentation. It was not geared towards going backin time to evaluate the decisions taken. We could notvalidate some of the essential information that wasforwarded to the Council, such as the initial estimatesof the financing gap for some programmes.

X This can be partly explained by the crisis contex t, theinitial time pressure and the novelty of programmemanagement to the Commission. The availability ofrecords improved with time, but even for the mostrecent programmes some key documents weremissing.

XI The conditions in memoranda of understanding weremostly justified by specific reference to the Councildecision. However, conditions were not always suffi-ciently focused on the general economic policy condi-tions set by the Council.

Different approaches

XIIWe found several examples of countries not beingtreated in the same way in a comparable situation. Theconditions for assistance were managed differentlyin each programme. In some programmes they wereless stringent overall, which made compliance easier.When comparing countries with similar structuralweaknesses, it was found that the required reformswere not always in proportion to the problems facedor that they pursued widely different paths. Some, butnot all, countries’ deficit targets were relaxed morethan the economic situation would appear to justify.

Limited quality control

XIIIOne of the reasons for these weaknesses was that theprogramme design and monitoring were largely in thehands of the Commission’s programme teams. Beforereaching the Council or the Commission, the key docu-ments resulting from a team’s work were subject toreview, but this was insufficient in several respects. The underlying calculations were not reviewed byanyone outside the team, the work of the experts wasnot thoroughly scrutinised and the review process wasnot well documented.


11/120


Room for improvement

XIVOn the positive side, the programme documents,which provide the basis for the decisions that theCouncil or the Commission take, have improvedsignificantly since the first request for financial assist-ance. This was due to the allocation of additional staffto programme management, more experience gainedand better preparation. However, even the morerecent programme documents lacked some essentialinformation.

XVFor monitoring, the Commission used accrual-baseddeficit targets. This ensures consistency with theexcessive deficit procedure, but it also means that,when a decision on programme continuation is to betaken, the Commission cannot report with certaintywhether the beneficiary Member State has compliedwith the deficit target, as accrual-based deficits canonly be observed after a certain time has elapsed. Themanner of reporting on compliance with conditionswas not systematic. Many different terms were used toconvey non-compliance, leading to confusion. Someconditions were not reported on. A few conditionswere reported to have been met when this was not infact the case.

Programme met their objectives

XVI The revised deficit targets were mostly met , with someexceptions. As economic activity declined in 2009,countries suffered losses in revenue, which wiped outany gains from new revenue measures. Tax-neutral ortax-offsetting reforms caused additional short-termfiscal costs and some countries took additional taxmeasures to offset the declining tax-to-GDP ratios.Structural deficits improved but at a varying pace. Partof the fiscal adjustment took place in a non-lastingway. Countries used one-off measures to meet thetargets.

XVIIMember States complied with most conditions set outin their programmes, albeit with some delays. Thesewere mostly incurred for reasons beyond the controlof the Commission. Occasionally, however, the Com-mission set unrealistically tight deadlines for wide-ranging reforms. High compliance does not mean thatall important conditions were complied with. In addi-tion, we observed that Member States tended to leavecompliance with the more important conditions untiltowards the end of the programme period.

XVIII The programmes were successful in promptingreforms. Countries mostly continued with the reformsthat were sparked by the programme conditions.Reversals of the reforms were rare at the time ofour audit. They were offset by alternative reforms,which were often not equivalent in terms of potentialimpact.

XIXIn four of the five countries, the current accountadjusted faster than expected. This is largely explainedby the unexpected improvement in income balance,and to a lesser extent by the unexpec ted improve-ment in trade balance.


12/120


Our recommendations

(a) The Commission should establish an institution-wide framework allowing the rapid mobilisation ofthe Commission’s staff and expertise if a financialassistance programme emerges. The Commissionshould also develop procedures in the context ofthe ‘two-pack’ regulations.

(b) The forecasting process should be subject to moresystematic quality control.

(c) To ensure the factors underlying programme deci-sions are internally transparent, the Commissionshould enhance record-keeping and pay attentionto it during quality reviews.

(d) The Commission should ensure that properprocedures are in place for the quality review ofprogramme management and of the content ofprogramme documents.

(e) For budget monitoring purposes, the Commissionshould include, in memoranda of understanding,variables that it can collect with short time lags.

(f) The Commission should distinguish conditionsby importance and target the truly importantreforms.

(g) For any future programmes, the Commissionshould attempt to formalise interinstitutionalcooperation with other programme partners.

(h) The debt management process should be moretransparent.

(i) The Commission should further analyse the keyaspects of the countries’ adjustment.


13/120

11Introduction

The sovereign debt crisisin the EU Member States

01 The financial crisis that hit Europe in2008 triggered the European sover-eign debt crisis. The sovereign debtcrisis had many causes, including weakbanking supervision, poor fiscal pol-icies and the difficulties experiencedby large financial institutions (thebailout costs of which were borne bythe general public). The effec ts of thecrisis are still being felt as this reportis published, and the resulting loanprogrammes have since run into thehundreds of billions, bringing onerousmanagement responsibilities for allconcerned.

02 The crisis affected the EU MemberStates in two waves. In the first in-stance, the contagion of the financialcrisis spread through foreign creditand trade channels to a number ofnon-euro area Member States. Thesharp increase in foreign capitaloutflows and decrease in exportsrevealed significant external imbal-ances for some countries, three ofwhich (Hungary, Latvia and Romania)sought financial assistance from theEU’s balance-of-payments mechanism 1,the International Monetary Fund (IMF) 2 and other sources 3.

03 The second wave of the crisis con-cerned euro area Member States.Although monetary union initiallyprovided some liquidity cushioning,credit agencies started to downgradeseveral Member States’ sovereignbond ratings in 2009. Sovereign debtmarket interest rates increased sharplyin some countries 4 but decreased inothers 5. As a result of the decline inprivate funding and escalating borrow-ing costs, the governments of Irelandand Portugal, among others 6, appliedfor external financial assistance.

04 The purpose of this f inancial assistancewas to help countries repay or financetheir maturing debt and deficits. Itprovided a cushion to ease implemen-tation of the adjustment programmesthat were necessary in each country tocorrect underlying problems. In broadterms, the mechanisms addressed theneed to safeguard the stability of theeuro area or the EU as a whole, to limitthe risk of contagion and to preventa sudden shock to the economies ofbeneficiary Member States. The objec-tives of each programme differed indetail, but the overall aims of financialassistance were to return MemberStates to sound macroeconomic orfinancial health and restore theircapacity to meet their public-sector(euro area) or balance-of-payments(non-euro area) obligations. Figure 1 provides a timeline of the financial as-sistance provided to the five MemberStates covered in our audit.

1 Based on Article 143 of the Treaty on the Functioning ofthe European Union.

2 Poland received access to IMFresources in May 2009through a exible credit line,but it did not require EUsupport.

3 Depending on eachprogramme, other sources ofnancing were to be providedby the World Bank, theEuropean Bank forReconstruction andDevelopment, the EuropeanInvestment Bank, the CzechRepublic, Denmark, Finland,Poland, Norway and Sweden.

4 Ireland, Greece, Spain, Italyand Portugal.

5 Germany and France.

6 In the euro area, Greece, Spainand Cyprus also requestednancial assistance.


14/120

12Introduction

05As the crisis followed a differentcourse in each country, assistancewas requested for different reasons. Annex I provides an overview of thosereasons.

06A number of legal instruments weredevised for providing financial assist-ance. Non-euro area Member Statescould take advantage of the existingbalance-of-payments mechanism.Ireland and Portugal were helped bythe newly created European FinancialStabilisation Mechanism (EFSM) andthe European Financial Stability Facil-ity (EFSF). Only the balance-of-pay-ments mechanism and the EFSM were

established as EU instruments on thebasis of the Treaty on the Functioningof the European Union, as the EFSF isan intergovernmental vehicle outsidethe EU framework. Annex I I providesan overview of the legal basis of thethree instruments. The five countriesalso received financial assistance fromthe IMF.

07Assistance to non-euro area countrieswas managed by the Commission andthe IMF. The European Central Bank(ECB) — the third member of the troikaof creditors — was involved in themanagement of financial assist ance toIreland and Portugal. This text refers tothese institutions as ‘the partners’.

F i g u r e

1 Timeline of financial assistance

Note: The start date of each programme is the date on which the memorandum of understanding between the Member State and theCommission was signed.

Source: ECA.

Euro area

Non-euro area

2008 2009 2010 2011 2012 2013 2014

HU (Nov ’08 – Nov ’10)

LV(Jan ’09 – Jan ’12)

RO (Jun ’09 – Jun ’11) RO 2 (Jun ’11–Mar ’13)

IE (Dec ’10 – Dec ’13)

PT (May ’11 – May ’14)


15/120

13Introduction

Design andimplementation

Programme design

08Requests for financial assistance fromthe EU are initiated by Member Stateauthorities. Requests must be submit-ted to the Commission and to the Eco-nomic and Financial Committee (EFC) 7 together with a draft economic adjust-ment programme 8. Commission staff 9 prepare the macroeconomic forecastsunderpinning each programme andconduct negotiations with the nationalauthorities. The mandate for thosenegotiations is set out in advance ina policy brief 10. The Commission willinformally contact other stakehold-ers (the IMF, the EFC and, in the caseof euro area countries, the ECB) toobtain a coordinated position on theprogramme conditions that will ac-company the financial assistance (seeparagraph 11).

09 The negotiation visit 11 to the countryculminates in a set of key documents,listed below.

(a) A proposal for a Council decisionon the financial assistance, set-ting out the general economicpolicy conditions attached to theassistance and the broad financialterms of the loan (financial enve-lope, instalment arrangements andmaximum average maturity 12).

(b) An explanatory note to the aboveproposal, presenting the Commis-sion’s assessment of the MemberState’s planned economic adjust-ment programme. The assessmentis based on internal Commissionforecasts, financial and macroeco-nomic models, Eurostat data andinformation obtained from thenegotiation visit.

(c) A memorandum of understanding,which represents a more detailedcommitment by the Member Stateand is typically signed by theMember State authorities and theCommission. The memorandumof understanding sets preciseeconomic conditions for the dis-bursement of each instalment andestablishes the Member State’sreporting requirements.

(d) A loan agreement, containing thetechnical and legal details of theloan (such as the average ma-turity calculation, interest rates,disbursement and repaymentarrangements 13).

7 The EFC was set up topromote coordination ineconomic and nancialpolicies among the MemberStates. The committee iscomposed of senior officialsfrom national administrationsand central banks, the ECB andthe Commission.

8 Article 3(2) of CouncilRegulation (EC) No 332/2002of 18 February 2002establishing a facilityproviding medium-termnancial assistance forMember States’ balances ofpayments (OJ L 53, 23.2.2002,p. 1) and Article 3(1) of CouncilRegulation (EU) No 407/2010of 11 May 2010 establishinga European FinancialStabilisation Mechanism(OJ L 118, 12.5.2010, p. 1). The economic adjustmentprogramme is a set ofadjustment policies that thecountry aims to implement tocorrect underlying problemsand re-establish itsmacroeconomic and nancialstanding.

9 The programme team, thehead of the unit responsiblefor the programme country,the programme head and thedirector responsible for theprogramme country.

10 The policy brief analyses thenancial and macroeconomicsituation of the Member Stateand the potential impact ofany proposed measures on itseconomy. It also sets out theframework for negotiationswith the Member State. Thepolicy brief is approved by theCommissioner responsibleafter having received thecomments from thepresidents of the EFC andEurogroup Working Group(EWG).

11 The visits are conducted bysenior staff from theCommission and the IMF, plusthe ECB in the case of euroarea countries. In general,these visits last between 2 and3 weeks.

12 Each disbursement of thenancial assistance can havea different maturity. Theweighted average of thesematurities must not exceed

the maximum averagematurity set in the Councildecision.


16/120

14Introduction

10As the IMF and the EU 14 remain respon-sible for their own lending decisions,two separate sets of programmedocuments are prepared, one by eachof the two institutions. In the EU, theultimate decision on granting financialassistance rests with the Council.

Conditions

11Programme conditions explicitly linkthe approval or continuation of finan-cing to the implementation of a rangeof measures, generally those whichare important to programme object-ives. The programme conditions mayalso include key data targets that willsound warning bells if policies veer offtrack (e.g. the inflation target) 15.

12Conditions are either quantitative (e.g.a limit on fiscal deficit) or structural(e.g. greater enforcement of competi-tion law). They also differ by intendedfunction:

(a) to correct key imbalances (e.g. fis-cal deficit reduction);

(b) to prevent imbalances from recur-ring in the future (e.g. adoption offiscal responsibility laws); and

(c) to detect the scale of imbalances(e.g. a comprehensive assessmentof public–private partnerships(PPPs)).

13 The conditions set effectively build onthe Member State’s economic adjust-ment programme and the partners’analysis. Thus, each programme willhave a different set of conditions fo-cusing on fiscal consolidation, struc-tural reforms to promote economicgrowth and/or financial sector reformsor support (see Table 1 ).

14All the assistance programmes tar-geted fiscal consolidation through theadoption of fiscal policies to reducegovernment deficits and make debtsustainable. Apart from this primaryobjective, fiscal consolidation is im-portant in the programme contextbecause it can, inter alia:

(a) facilitate internal adjustmentthrough various channels (suchas a direct reduction in aggregatedemand, or the impact of public-sector wages and employment onprivate-sector labour costs 16);

(b) facilitate external adjustment byreducing an excessive current-account deficit;

(c) help tackle inflation by affectingprices through a range of channels(such as its impact on aggregatedemand or on demand for money);

(d) free up resources for other sec-tors if the economy is financiallyconstrained; and

(e) inspire confidence — underspecific conditions — within theeconomy and thus have an expan-sionary effect 17.

13 A borrowing contract is alsosigned for each instalment.

14 EFSM funding to Ireland andPortugal was provided inconjunction with EFSFfunding, and identicaleconomic conditions appliedto both lending sources.

15 Moon, S. and Bulíř, A., DoIMF-supported programs helpmake scal adjustment moredurable? , IMF working paper,2003.

16 Hernandéz de Cos, P. andEnrique Moral-Benito, The roleof public wages in scalconsolidation process,workshop ‘Government wagebill: determinants, interactionsand effects’ organised byDirectorate-General forEconomic and Financial Affairs, 2014.

17 Daniel, J., Davis, J., Fouad, M.and Van Rijckeghem, C.,’Fiscaladjustment for stability andgrowth’, Pamphlet Series No 55, 2006, International

Monetary Fund, Washington,D.C.


17/120

15Introduction

15 The role of struc tural conditions ismainly to implement structural re-forms aimed at enhancing the long-term sustainability of macroeconomicadjustments. Financial sector condi-tions aim to stabilise the financial sec-tor, which often plays a key role at thebeginning of a programme, when bankrestructuring typically takes place.

Implementation

16 The Commission must verif y, at regularintervals, whether the economicpolicy of the beneficiary MemberState complies with its adjustmentprogramme and with the programmeconditions. The programme conditionsare monitored by the Commission 18 in coordination with the IMF. To thatend, the beneficiary Member Statesreport regularly on the progress ofprogramme implementation 19, and thepartners undertake monitoring visits.Both the Commission and the IMFprepare their own review reports aftertheir visit.

T a

b l e 1 Focus of programme conditions

Programme Fiscal consolidation Structural reforms Financial sector

Hungary Yes No signicant reforms — Strengthening of regulation and supervision

Latvia Yes— Business environment— Active labour market policies— Management of EU funds

— Strengthening of regulation and supervision— Bank resolution or consolidation plans

Romania I Yes — Business environment

— Management of EU funds— Strengthening of regulation and supervision

Romania II Yes — Labour market reforms— Competitiveness within the economy No signicant reforms

Ireland Yes — Labour market reforms— Competitiveness within the economy

— Deleveraging— Strengthening of regulation and supervision— Bank resolution or consolidation plans

Portugal Yes

— Labour market reforms— Competitiveness within the economy and outside— Education— Health sector— Business environment— Judicial reforms

— Housing market

— Deleveraging— Strengthening of regulation and supervision— Bank resolution or consolidation plans

Source: ECA based on relevant Council decisions and initial programme documents.

18 In the case of the EFSM, theECB is also involved inmonitoring.

19 A memorandum ofunderstanding generallyincludes an annex with theindicators and data that thecountry must provide ona regular basis.


18/120

16Introduction

17On the basis of its review, the Commis-sion, after consulting the EFC, decideson the release of further instalmentsand whether any programme con-ditions should be modified and/orappended to address unforeseencircumstances such as a change in themacroeconomic situation, significantdata revisions or political or institu-tional factors in the Member State. The initial memorandum of under-standing can be revised to account forthese new circumstances and secureimplementation of the adjustmentprogramme.

18If the review results in significantamendments to the conditions 20, theCouncil decision must be amendedas well. The procedure for amendingthe Council decision is largely similarto the initial negotiation process (seeparagraphs 8 to 10) 21.

Financial resources

Programme financing

19 The amount lent under each assistanceprogramme covers the Member State’s‘financing gap’ — its financial needsuntil it is able to finance its govern-ment accounts (maturing debt andnew deficits) or its external economicposition (balance of payments) undernormal conditions.

20 The external f inancing of euro areaeconomies goes through the Eurosys-tem 22 . The financing gaps of euro areacountries are therefore limited to theneeds of their governments. Out-side the euro area, the financing gapincludes all external financing needsas presented in a country’s balance ofpayments.

21During programme negotiation, thepartners estimate the financing gapusing macroeconomic and fiscal projections and a set of assumptions, suchas the rollover rates of debt 23 . The finallevel of programme financing is basedon this estimate, but it is also subjectto negotiation between the partnersand the Member State. Each lenderthen takes a share in the programmefinancing 24. The Member State’sinternal funding sources may also beconsidered.

22For an overview of the key elementsof programme financing, the rela-tive share of each key lender and theamounts of financial assistance dis-bursed, see Figure 2 .

20 For example if decit targetsare changed.

21 The Commission carried out36 formal reviews of the sixaudited programmes. In 16cases, the review was followedby an amendment to thecorresponding Councildecision.

22 The Eurosystem is the centralbanking system of the euroarea. It consists of theEuropean Central Bank andnational central banks.

23 When a country’s debt isabout to mature it needs to be‘rolled over’ into a new debt,which will cover repayment ofthe old. Rollover brings the riskthat the debt may berenanced at a higher interestrate, thus increasing interestexpenditure, or that theborrowing country will notraise enough funds to coverrepayment of the old debt infull.

24 Although the legal framework

allows for the provision ofloans by the EU alone, inrecent practice assistance hasusually been granted incooperation with the IMF andother international institutionsor countries.


19/120

17Introduction

F i g u r e

2 Programme financing

Source: ECA, data provided by the European Commission and the IMF.

… as a percentage of each country’s GDP

19 % 40 % 17 % 43 % 46 %

Hungary20 billion

Latvia7.5 billion

Romania20 billion

+ 5.4 billion1

Ireland67.5 billion

Portugal78 billion

33 %6.5 billion

41 %3.1 billion

25 %5 billion

33 % (60 % with EFSF)22.5 (40.2) billion

33 % (67%)26 (52) billion

Financing committed by the European Union

Overall programme nancing in euro

9 4 %

6 5 % 1

0 0 %

9 2 %

8 5 %

7 2 % 9

2 %

1 0 0 %

9 7 %

1 0 0 %

1 0 0 %

9 7 %

Disbursementsas a percentage of committed nancing

PortugalIrelandRomaniaLatviaHungary

Balance of payments/EFSM EFSF IMF Other lenders

1 Precautionary assistance. Not included in the other charts.


20/120

18Introduction

23 The five Member States used 95 % ofthe financing committed by the Euro-pean Union, and 13 % of that amounthad already been repaid by 31 March2015 (see Figure 3 ).

F i g u r e

3 Financial assistance: amounts disbursed and repayments outstanding

Source: ECA, data provided by the European Commission.

63 billion euro

Assistance to be providedby the European Union

Amounts borrowedand disbursed

60 billion euro

Disbursedby country

Repaymentsoutstandingby country1

Total outstanding52 billion euro

5.0 3.5

23.9 23.9

2.9 0.7

22.5 22.5

5.5 1.5

Romania

Portugal

Latvia

Ireland

Hungary

70 %

100 %

24 %

100 %

27 %

1 Data as of 31 March 2015.


21/120

19Introduction

Borrowing on the markets

24 To provide its share of total f inancialassistance, the European Union bor-rowed on the capital markets and lentthe money to the Member States indifficulty on the same terms, i.e. forthe same amount and with the samecoupon and maturity date. By the endof 2014, the EU had made a total of22 bond issues with a nominal valueof 60.1 billion euro. The Commissionmanaged the borrowing on behalf ofthe EU.

25Borrowing generally took the formof euro medium-term notes (EMTNs),which are issued on the capital marketsand listed on the Luxembourg stockexchange. EMTNs are redeemable infull at maturity and bear fixed annualinterest over the duration of their ex-istence. European investors representaround 80 % of all investors in EMTNs.

26 The debt servicing of EMTNs remainsan obligation of the European Union,which must ensure that all paymentsare made in a timely manner if thedebtor Member State fails to reim-burse the capital and/or interest due.


22/120

20Audit approach

27 The audit examined whether the Com-mission’s management of financial as-sistance programmes was appropriate. This was addressed by the followingsub-questions.

(i) Were the growing fiscal risksdetected on time?

(ii) Were processes suff iciently welldesigned to make a compre-hensive input into programmedecisions?

(iii) Did the Commission borrow atthe best possible rates and in ac-cordance with best debt issuancepractices?

(iv) Did the financial assistanceprogrammes meet their mainobjectives?

28 The audit covered the Commission’smanagement of financial assistanceprovided under the balance-of-pay-ments facility and the EFSM, for whichthe Commission borrowed on thecapital markets using the EU budget asguarantee. The audit encompassed thefinancial assistance paid to Hungary,Latvia, Romania (the first two pro-grammes), Ireland and Portugal, withan emphasis on the Commission’s rolein these programmes. We also exam-ined the Commission’s cooperationwith its partners (the ECB and the IMF),but did not audit the partners.

29We did not audit the decisions takenat the EU’s political level and welimited the scope of the audit in sev-eral aspects. We did not consider thecounterfactual scenario of no financialassistance or the feasibility of resolv-ing the crises by other means (e.g. themutualisation of sovereign debt). Nordid we assess debt sustainability or thelikelihood that the loans will be repaid.We also did not evaluate if the Coun-cil had chosen the most appropriatedeficit targets or structural conditionsto resolve the crisis. When auditing theCommission’s cooperation with otherpartners, we did not assess whethertheir involvement was justified.

30 The audit criteria were derived fromthe following sources:

(i) regulatory requirements (Councilregulations, Council decisions,memoranda of understandingbetween the Commission and theMember States);

(ii) the Commission’s internal rulesand procedures (internal controlstandards and guidelines);

(iii) good management practices andevaluation criteria derived fromevaluations of IMF programmes,ex post evaluations of the EU’sbalance-of-payments assistanceto non-EU countries, previousCourt reports on macrofinancialassistance, guidelines and otherpublications from independentinstitutes and think tanks, andacademic research; and

(iv) best practices suggested in a num-ber of guidelines on public debtmanagement, investor relationsand transparency in debt issuance.


23/120

21Audit approach

31 Annex I II provides further details ofthe audit methodology.

32 The audit evidence was collected onthe basis of:

(i) a detailed review of documenta-tion relating to the EU’s financialassistance programmes (pro-gramme documents, internalCommission analyses, spreadsheetforecasts, evaluations and studiesby other organisations);

(ii) interviews with the staff of theCommission’s Directorate-Generalfor Economic and Financial Affairs;

(iii) interviews with the staff of nation-al authorities, such as ministries offinance, central banks, programmecoordination units and sectorministries;

(iv) a scorecard analysis of programmeconditions 25 to classify them bycharacteristics and compliance;and

(v) interviews with the staff of the IMFand the ECB.

25 The sample included360 unique conditions.


24/120

22Part IThe Commission was unprepared for themagnitude of the crisis that broke out

The Commissionunderestimated theextent of the fiscalimbalances in the run-upto the crisis

33Warning the Council about mountingfiscal imbalances in the Member Statesis arguably one of the Commission’smost important activities in the areaof economic and financial affairs. Priorto the crisis, there was already a sur-veillance framework geared towardsmonitoring the fiscal situation in theMember States 26 .

34Although the financial assistanceprogrammes’ macroeconomic object-ives went beyond fiscal adjustment,in all five programmes budget sup-port 27 turned out to be a key part ofhow assistance was used. This raisesa question as to whether the mount-ing fiscal risks were detected on timeand flagged under the surveillanceprocedures.

The Commission estimatedthe public budgets to bestronger than they actuallywere

35 There are two key types of f iscalindicator. The first is the actual fiscalbalance, or the difference betweengovernment revenue and expenditure.However, this balance is sensitive to

temporary factors, such as economicgrowth, asset prices and one-of f fiscalmeasures. The second is other fiscalindicators (e.g. cyclically adjusted bal-ance, structural balance) which aim tocorrect these temporary factors andthus provide a more accurate view ofa country’s underlying fiscal position. These indicators cannot be directlyobserved but are the result of complexcalculations.

36During the pre-crisis period, actualfiscal balances were generally correctlyestimated no later than at the end ofthe year in question. Nevertheless, theemergence of high budget deficits for2008 in Ireland, Latvia, Portugal andRomania was poorly predicted in theCommission’s deficit forecasts. Forexample, speaking of Ireland in March2008, the Commission stated that ‘therisks attached to the budgetary projec-tions are broadly neutral for 2008’ 28 .Yet at the end of 2008 the actual fiscalbalance was 7.2 percentage points ofGDP lower than forecast.

37In its cyclically adjusted balance calcu-lations during 2005-2008, the Commis-sion systematically overestimated thestrength of countries’ public finances 29. This error was mainly caused bychanges in estimated potential GDPgrowth rates. All five countries wereexperiencing an economic boom, butthe Commission initially estimated thescale of overheating to be more mod-est than, with hindsight, we know it tohave been.

26 This was the ‘preventive arm’of the Stability and GrowthPact established by CouncilRegulation (EC) No 1466/97 of7 July 1997 on thestrengthening of thesurveillance of budgetarypositions and the surveillanceand coordination of economicpolicies (OJ L 209, 2.8.1997,p. 1). The surveillanceframework was last reformedin 2005 by Council Regulation(EC) No 1055/2005 amendingRegulation (EC) No 1466/97(OJ L 174, 7.7.2005, p. 1).

27 The nancing of running scaldecits and the repayment ofgovernment debt.

28 European Commission, Ireland— Macro scal assessment— an analysis of the December2007 update of the stability programme , p. 36.

29 Compare the rst estimatemade after the year’s end (i.e.in spring forecast) witha recent estimate(i.e. October 2013).


25/120

23Part I — The Commission was unprepared for themagnitude of the crisis that broke out

38 The estimate of the scale of overheat-ing (the size of a positive output gap 30)was in line with the strong increase inthe indicators of the overall produc-tion capacity of the economy. How-ever, this increase was generated bythe housing booms, and investmentsin housing do not have the same effecton increasing the economy’s potentialas other more productive investments.

39 The Commission had focused onassessing the output gap in the realeconomy, without considering sectorimbalances such as in the financialor construction sectors. However, itbecame clear during the crisis that in-flation and output monitoring are notenough to guarantee macroeconomicstability and sustainable fiscal bal-ances. The surveillance framework hassince been significantly overhauled totake greater account of financial-sectorand macroeconomic risks.

40 The Commission identified some sec-tor booms after the event; for instance,a construction boom was noted inIreland, Latvia and Romania and pri-vate credit booms in all five countries.However, the algorithm used to calcu-late the output gap did not make anyallowance for such imbalances.

Assessments underestimatedthe potential fiscal risks

41 The Stability and Growth Pac t re -quires Member States to report recentfiscal developments and forecastsin their stability and convergenceprogrammes. These programmes areassessed by the Commission, whichchecks the plausibility of the trajec-tories forecast for GDP and the publicbudget.

42One important drawback of the Com-mission’s assessments prior to 2009was the lack of reporting on the build-up of contingent public-sector liabil-ities 31. Fiscal risks related to financialstability, state-owned enterprises orother public guarantees (e.g. PPPs)were not identified.

43 The Commission’s surveillance did notadequately assess the implicationsof the large foreign financial inflows,especially in the form of debt, whichwere fuelling construction and de-mand booms and had the potential todisrupt the banking sector in a subse-quent bust phase. The Commission didnot note this link prior to its analysisof Member State requests for financialassistance.

30 A positive output gap is theamount by which the actualoutput of an economyexceeds its potential output.

31 For example, the long-termliabilities from public pensionsor signicant future scalexpenditure resulting fromthe need to take measures inthe event of a bailout. With theexception of the publicpension liabilities, theCommission did not report onother contingent liabilities.


26/120

24Part I — The Commission was unprepared for themagnitude of the crisis that broke out

44Neither did the Commission’s as-sessments examine the sensitivityof budget balances to a sudden hikein the interest rates applied to gov-ernment debt, or to equally suddenshocks to economic growth or ex-change rates. Yet these risks too hadbeen recognised or warned againstin various research papers publishedprior to the crisis 32.

As a result theCommission was notready for the programmemanagement

45Prior to 2008, the balance-of-paymentsfacility had last been used in 1993 33. By2008 any experience acquired in thefirst half of 1990s was lost. So, provid-ing countries with financial assistancewas in effect a new activity for theCommission.

46Having underestimated the fiscal im-balances and their implications for theMember States, the Commission wasunprepared for the task of programmemanagement. The initial programmesin particular had to be set up under ex-treme time pressure. For instance, theproposal for the first programme (Hun-gary) had to be produced in just a fewdays. The Commission possessed verylittle in-house experience of designingand managing financial assistance pro-grammes, and that experience couldnot be acquired at such short notice.

47During the initial phase, IMF teamshelped the Commission to prepareparts of the programme analysis. Thepractices used by the Commissionwere also developed on the basis ofIMF practices (see paragraph 109).

32 See, for instance, ‘Fiscalpolicies and nancial markets’in ECB Monthly Bulletin02/2006, p. 71.

33 Council Decision 93/67/EEC of18 January 1993 concerninga Community loan in favour ofthe Italian Republic (OJ L 22,30.1.1993, p. 121).


27/120

25Part IIProcesses were generally weak

48 The regulations entrust the EuropeanCommission with the operationalmanagement of financial assistance.A large part of this is about arriving atkey programme decisions concerningthe initial granting of assistance and(at each review) the continuation ofthe programme. Decision-making re-quires a sequence of steps and coord-ination with the Council and the otherpartners (see Figure 4 ).

49We divided the analysis of programmemanagement processes into the fol-lowing three parts.

(i) Processes within the Commis-sion, in particular within theDirectorate-General for Economicand Financial Affairs. Did theCommission keep control of theprogramme’s development?

(ii) Process of providing informationto the Council and the Commis-sioner in charge. Did the Counciland the Commissioner in chargereceive programme documents ofgood quality?

(iii) The process of interaction withother partners. Did the interactionwith other partners work?


28/120

26Part II — Processes were generally weak

F i g u r e

4 Programme decisions

Programmenegotiation

1streview

2ndreview

nthreview

Decisions ongranting assistance

Decisions onprogramme continuation

Decisions onprogramme continuation

What decisions are taken?

Each decisionconsists of three

separate acts

Setting of general

programmeparameters

Drawingup of a detailedreform agenda

Releaseof payment

Commission Decision

Council decision¹

Overall amount and maturityDecit targetsBroad areas of structural reforms

B

Programme timeline

Typically taken by the Vice President on the Commission's behalf

1 In case of the Irish and Portuguese programmes, the general programme parameters were set both by the Council and the EFSF board of directors, i.e. the Eurogroup.

The EFSF decision-making process is separate from the EU decision-making procedure.

A

Not all decisions are needed at eachreview. For instance if no changes arenecessary to the general parametersof assistance, the Council decisiondoes not have to be amended.

Set out in amemorandum ofunderstanding signed by theMember State and theEuropean Commission


29/120


F i g u r e

4

Source: ECA.

Sequence of stepsC

+

+

2

3

1

Audit areas1. Process within the Directorate-General

for Economic and Financial Affairs2. Information provided to the Council

and the Commissioner responsible3. Interaction with other partners

2 EFC – Economic and Financial Committee,EWG – Eurogroup Working GroupMoU – memorandum of understanding

Consultation in the Economic and Financial Committeeand Eurogroup Working Group

If requested by the Member State

Payment is made only if a new MoUhas been signed by the Member State

Programme documents

Payment

Commissiondecision

Signatureof MoU

Councildecision

Decision by theCollege of

Commissioners

Internal review by the Commission

If new programmeor Council conditionsare to be changed

If changes to reformagenda are necessary

Draft Councildecision

Draft MoU

Explanatorynote or review

report

Councilposition

Carried out in parallelwith IMF and ECB staff

Denes the mandatefor the negotiation team

Negotiationof review visit

Policy brief

Other partners

Interaction with the other partners to align their positionsto reach a common agreement

EFC and EWG chair

Commissionernegotiation mandate

update to the

statement by Council or Eurogroup if new programme or important amendment discussion by EFC and EWG 2

indications from chair of EFC and EWG

Depending on importance, can take different forms:


30/120


The Commission’sprocesses did not alwaysincorporate sufficientchecks and balances

50 The Commission’s processes shouldensure — through sufficient checksand balances — that the initial pro-gramme design and any updates meetthe following criteria:

(i) comprehensiveness — relevantinformation, both in relation to theMember State and from economicresearch, should be taken onboard;

(ii) consistency — different elementsof the programme should be mu-tually consistent, with a minimumof conflict between the dif ferentprogramme objectives;

(iii) documentation — due recordshould be kept of the reasoningbehind the decisions made on vari-ous elements of the programme 34 ;

(iv) conformity — the memoranda ofunderstanding should be in linewith the Council decisions;

(v) even-handedness — as pro-gramme conditions are mostlydrawn up by Commission’s ‘pro-gramme teams’, countries shouldbe treated consistently by the dif-ferent teams, applying similar con-ditions in similar circumstances 35.

51 The work of programme teams shouldbe checked and challenged to ensurethat teams respect these criteria.

The Commission was mostlythorough in obtaining thenecessary information formonitoring purposes

The Commission extensivelyrelied on internal expertise aswell as liaising with a number ofactors in Member States

52 The Commission relied on the internalexpertise available in the Directorate-General for Economic and FinancialAffairs. Over time, it also increasinglyconsulted specialists from other direct-orates-general. These other expertsprovided limited input for the initialprogrammes, mainly through inter-departmental (‘inter-service’) consul-tation. They were more extensivelyinvolved in the two euro area pro-grammes. The use of wider expertiseallowed for a more thorough analysisof countries’ problems, especially inthe area of structural reforms. TheCommission’s human resources policyalso did not allow greater use of ex-ternal expertise for specific narrowlydefined areas, a practice common inthe IMF.

34 Commission internal controlstandards Nos 8 and 11.

35 Several studies conrm thatthis risk is real. For instance, anoverview of relevant academicresearch can be found inHakan Gunaydin, Compliancewith the IMF conditionality:Selection bias and conditions onsocial policy — Prepared for thePEIO Annual Meeting, 2014. TheIndependent Evaluation Officeof the IMF also provided anaccount of lack of even-handedness that it noted in itsevaluations. For more details,see IEO, Recurring issues froma decade of evaluation lessonsfor the IMF , 2014.


31/120


53 The regular country visits made by theCommission and the other partnersprovided the main opportunity forobtaining the necessary informationfor monitoring. During its negotiationand review visits, the Commission con-sulted a number of ministries, othergovernment departments and keyindependent institutions (such as cen-tral banks, fiscal councils and supremeaudit institutions). It also held meet-ings with other bodies, including em-ployers’ associations, trade unions andopposition political parties. In threeof the five countries 36 the Commissionmaintained a resident representative.All of this enabled the Commission tostay informed and to identify reformsfor inclusion in the programme.

The Commission collecteda wide range of information

54In addition to the information gath-ered from country visits, each memo-randum of understanding specifiedmonitoring reports that the countryin question was required to provide tothe Commission. The information tobe collected in these reports was verybroad and touched on many areas. TheCommission requested similar moni-toring information in all programmes.However, in a few cases key informa-tion was not requested or only a lim-ited amount of key information wascollected.

Programme forecastsprovide for broad checks ofconsistency in programmedesign, but with manycaveats which are notverified

55Good forecasting is essential for pro-gramme consistency. However, fore-casts are unlikely to be 100 % ac curateall the time, as they involve manyimponderables which cannot be pre-dicted with any precision. Instead offocusing on the accuracy of forecasts,we assessed the quality of the fore-casting processes and tools appliedby the Commission in the programmecontext. In practice, the Commissionprepares two distinct but closely re-lated forecasts (see Figure 5 ):

(i) a programme forecast (compris-ing macroeconomic and deficitforecasts); and

(ii) a financing gap estimate.

36 Latvia, Ireland and Portugal.


32/120


F i g u r e

5 Links between the programme forecast and financing gap estimate

Source: ECA.

Current account

Key:

If part of forecast

Covering all nancingneeds of the economyOnly for countries outsidethe euro area

Cash needs due to nancialsector rescue

One-off expenditure due tonancial sector rescue

Cash-basedgovernmentdecit

Accrual-basedgovernmentdecit

Net repayments ofall government debt

Net repayments

ofexternalgovernmentdebt

Other balance-

of-paymentaccountsLinked through

several accountingidentities

External nancing gapestimateMacroeconomic forecast

Link between two closely related variables

Note: Only the key links are depicted.

Accounting identity

Government sectornancing gap estimateDecit forecast

Programme forecast Financing gap estimate


33/120


34/120


60 To manage these risks, quality controlof the forecasting results was done atseveral stages during the forecastingprocess. Some consistency checks 38 arebuilt into the spreadsheet. The fore-casters responsible for, respectively,the fiscal and the macroeconomicsides of the forecast discussed theirresults among themselves and, whenrequired, applied additional judge-ment in order to ensure consistency ofthe estimates 39.

61Essentially, however, quality controlof the forecasters’ work took the formof reconciliations, with no inquiry intothe forecasting process within theprogramme team. The spreadsheet ap-proach and the scarcity of documenta-tion made it hard to assess the plausi-bility of the assumptions and implicitparameters used in the forecasting.Discussion between partners was theonly control stage at which the plausi-bility of the assumptions and implicitparameters was checked.

F i g u r e

6 Macroeconomic and deficit forecasts at a glance

Source: ECA.

Ministry of nance Statistical office

Historical scal data

Eurostat Statistical office

Historical macro data

Judgement onbudget projection

New scal measures Ministry of nance

Effects of macroon budget

Effects of budgeton macro

Judgement onmacro variables

Decit forecast Macro forecast

38 For instance, accountingidentity or ratio-basedplausibility checks.

39 In autumn 2013 theCommission introduced a newtool for the automaticdetection of numerical andeconomic inconsistencies, butthe tool was not used for theprogramme forecasts.


35/120


62Analysis of the forecasting sheets alsorevealed examples of errors and incon-sistencies that were present despitethe consistency checks (see Box 1 ).

Difficulties in assessing theplausibility of scal multipliers

63Fiscal variables were recorded in thespreadsheets in the form of the ex-pected volume of revenue or spend-ing, which was increased or decreasedby the net budgetary yield of anynew fiscal measures during the year. The impact of new measures on theeconomy and the knock-on effec t onthe public budget were not explicitlystated or recorded in the spreadsheets.Instead, the forecast values of ad-ditional budgetary yield already tookaccount of the knock-on effec t.

64 The forecasts of net budgetary yieldwere taken from forecasts generatedby the national authorities, which esti-mated them internally as part of theirplanning procedures. The nationalauthorities discussed their estimateswith the partners’ staf f, whereuponthey might be revised before beingformally accepted and included in theprogramme forecasts.

Examples of errors and inconsistencies in forecasts(i) The amount of VAT levied on private consumption was not aligned between government revenue and

private-sector payments, leading to a difference of 0.7 % in GDP in 1 year.

(ii) Owing to manual formula adjustments, the amount of direct taxes levied on labour was not alignedbetween the government and household sectors. The difference exceeded 0.5 % of GDP.

(iii) A tax measure applicable from the middle of the year was applied to the forecast on a full-year basis. The corresponding deficit projection was used in the programme document supporting the decision torelease an instalment and communicated to the EFC in June 2009.

(iv) The impact of property prices was not factored consistently into projections of the same price index usedin different forecasts for the same year.

B o x

1


36/120


65 The impact of different f iscal measureson the real economy was accountedfor implicitly and outside the spread-sheets by way of several heteroge-neous procedures. The spreadsheetsused did not allow the source of par-ticular assumptions, or the reasons forthem, to be traced. The country desksthemselves had no real knowledgeof the models applied by nationalauthorities to estimate the parametersfor assessing the impact of differentfiscal measures on the economy. Therewas a lack of transparency and know-ledge about the implicit value of thefiscal multipliers 40 . Only in 2013 did theCommission carry out an exercise toestimate the implicit multipliers.

66 The only checks made of net yieldstook the form of a judgemental plausi-bility assessment by the par tners’ staff.While this approach is highly likely todetect any substantial over- or under-estimates, it is less amenable to iden-tifying relatively small errors whichmay cumulatively still be of significantmagnitude.

67 The programme teams kept recordsof some information, but to varyingdegrees because there was no formalrequirement to do so. As a result, wewere unable to identify and reviewthe assumptions and parameter valuesthat were used. If a forecast provedto be wrong, (for instance if it under-estimated the negative effect onGDP growth of a reduction in publicexpenditure), it was impossible or ex-tremely difficult to identify the sourceof the error.

Financing gap estimates wereincomplete and inaccurate

68 The financing gap estimate is ofparamount importance at the negotia-tion stage, because this is when theamount of financial assistance is set,typically to remain unchanged. If newfunding needs emerge during theprogramme, they have to be accom-modated within the overall financialenvelope. Unlike the preparation ofprogramme forecasts, estimating thefinancing gap was a new activity forthe Commission.

69For the euro area countries, the financ-ing gap corresponds only to the needsof their governments. Outside the euroarea, the financial assistance mustcover all external financing needs aspresented in a country’s balance ofpayments. The Commission always justified its decisions to provideassistance by submitting the initialfinancing gap estimate to the Coun-cil. The underlying calculations wereprepared in spreadsheets. However,the shortage of documentation greatlylimited the possibility of review. Whenrequested, the Commission was onlyable to provide provisional docu-ments 41, which differed from the finalversions submitted to the Council,could provide no information onimportant elements of the estimate 42 or could not provide any backgroundfiles43.

40 The scal multiplier measuresthe magnitude of impact ofdifferent scal consolidationmeasures on key macro-economic (consumption,output) and scal (decit)variables.

41 Latvian and Romanianprogrammes.

42 Irish programme.

43 Hungarian and Portugueseprogrammes.


37/120


70Our review of the provisional docu-ments revealed significant forecast-ing errors, such as double countingof some variables, use of incorrectamounts in foreign reserves or unreal-istically high assumptions of debtrollover rates. The errors in the spread-sheets were sometimes substantial 44 and had not been detected by internalreview. Discovery of the errors wouldnot necessarily have led to an increasein the amount of financial assistance,as this also depends on the lender’swillingness and ability to lend, but itcould have resulted in a different set ofpolicies to close the financing gap.

Programme features weregenerally soundly based,but important records weresometimes missing

Programme features weresupported by analysis

71In most cases, the Commission couldpoint to a body of analysis, basedon EU policy priorities and currentdata, that was used when settingprogramme conditions. In general,the depth of analysis had increasedover time. This was mostly on accountof improved expertise and a greaternumber of Commission staff assignedto programme management.

No trace of some key records

72Record-keeping is an important partof any process. Several of the Com-mission’s internal control standardsrefer to good practices in this regard 45.Careful record-keeping can improve anorganisation’s internal control by help-ing to show why certain decisions weremade. Records are central to monitor-ing and evaluation and a prerequisitefor good accountability.

73Because the assistance programmeswere unexpected, procedures werestill being fine-tuned and formalisedafter initial operations had begun.Operations were also complicated bytime constraints, an uncertain macro-economic environment and the com-plexity of programme management.Emphasis was therefore placed onsound operational procedures at thecost of good record-keeping.

74In a number of instances, the infor-mation used to support decision-making had been lost by the time ofthe audit, or collecting it proved tobe a mammoth task. Certain criticalactivities were not documented at all.Manuals of procedures, which typicallyanchor any process, were not well de-veloped. Box 2 illustrates the obstaclesencountered during the audit 46 .

44 In assessing the Romaniannancing gap, the intra-grouploans of corporates werecounted under both foreigndirect investments and newdebt ows under otherinvestments. This doublecounting led to a nancingneeds projection that was1.8 billion euro too low.Furthermore, the initialamount of foreign reservesused in the initial projectiondiffers by 0.7 billion euro fromthe gure reported now forthe end of 2008 by theRomania’s central bank. In theLatvian programme, theCommission projected that inthe rst year of theprogramme both thecorporate and banking sectorswould be able to obtain morelong and medium-termnancing than they needed toamortise. Such an assumptionin the spreadsheet was notonly unlikely given the severenancial pressures, but wasalso in contradiction to theinformation provided in theexplanatory note.

45 For instance, internal controlstandard No 8 requires that‘the DG’s main processes andprocedures are adequatelydocumented, particularlythose associated with criticalrisks’, and standard No 11 that‘appropriate processes andprocedures are in place toensure that the DG’sdocument management is […]efficient (in particular asregards retrieving appropriateinformation)’.

46 The problem of obtaining therelevant information was notunique to this audit. ThePortuguese Court of Auditorsnoted similar constraints in itsreport on the monitoring ofthe nancial assistance toPortugal, Relatório de auditorian.º 28/2013-2.ª S., Acompanhamento dosmecanismos de assistênciananceira a Portugal,Dezembro de 2013 .


38/120


75It is understandable that recordsshould be lacking for the initial phasein 2008 and early 2009, which bore allthe hallmarks of a crisis managementsituation. Over time, record-keepingclearly improved. Even so, some of theabove examples of missing recordsrelate to more recent years.

The Commission occasionallydeparted from Councildecisions

Cases of non-compliance withthe Council decisions

76We examined the consistency betweenthe memoranda of understanding andthe corresponding Council decisionsand found that most conditions were justified by reference to the decisions.However, in a few instances the condi-tions departed from the terms of theCouncil decision, or the memorandumof understanding had not been up-dated at the correct time (see Box 3 ).

Examples of difficulties in obtaining key information

Emails were used to report upwards on the progress of a visit and to seek approval for any departure from thenegotiation mandate. They were also an important means of communication among the partners and withnational authorities. Generally, however, these key records were not available because emails had either beenlost or could not be retrieved.

Documentation or data submitted by national authorities during the programme was not kept by the Com-mission or could not be consulted in its entirety because of the extensive resources required to reassemble it.

Some spreadsheets for Romania with macroeconomic programme projections were not made available to usduring the audit.

The spreadsheets with final f inancing gap calculations had been lost or had not been collected from the IMF. The calculations underlying some key amounts were sometimes unobtainable.

Fiscal monitoring by Commission staff was not documented at all, or only very basic spreadsheets were pro-vided in support of the Commission’s analysis of budget implementation.

B o x

2


39/120


Conditions were not sufficientlyfocused

77 The purpose of the conditions in thememorandum of understanding isto establish an operational road mapwith a view to achieving satisfactorycompliance with the general economicpolicy conditions set by the Council.

78For most conditions, the link withgeneral economic policy was relativelyclear because they either requiredpolicy change or were a steppingstone to major reforms. However, themanagement 47 did not sufficientlychallenge the proposals made to as-certain whether all the conditions werereally necessary for the achievement ofgeneral economic policy goals.

79Programmes included conditionswhich were either of minor impor-tance, given the programme goals,or whose potential effects would berealised long after the programmeperiod — conditions such as theimplementation of specific IT systems,the allocation of budgetary funds toa small government entity, the train-ing of public officials and research anddevelopment. However, here they hadvery remote links — if any — to theprogramme goals.

Examples of non-compliance with the rules

An informal visit to Romania in August 2009 revealed the Member State’s inability to meet the initial deficittargets. The Commission and the Romanian authorities consequently agreed to a substantial relaxation of thetargets. This revision was communicated in a note to the EFC in August 2009. However, neither the Councildecision nor the memorandum of understanding was modified accordingly. The targets were eventually up-dated in the memorandum of understanding in February 2010. The legal basis — the Council decision — wasupdated even later, in March 2010. However, the Commission postponed a release of a payment in November2009 on account of the fact that Romania’s budget did not comply with the updated target.

In Latvia, although the Council decisions did not refer to pension reforms, the memorandum of understandingincluded specific conditions in relation to the second pillar of the pension system.

In Ireland, although the Council decisions mentioned no specific conditions with regard to banking sur-veillance and sales of state assets and privatisation, these conditions appeared in the memorandum ofunderstanding.

B o x

3

47 Senior management in theCommission and thevice-president signing thememorandum ofunderstanding on behalf ofthe Commission.


40/120


Countries were not treated inthe same way

80Having analysed various aspects of theCommission’s programme manage-ment, we found several examples ofan inconsistent approach to differ-ent countries where it was difficultto justify this by specific nationalcircumstances.

Differences in the Commission’smanagement of conditions

81 The Commission managed the con-ditions for each financial assistanceprogramme differently. In some pro-grammes the conditions representeda relatively stable set of requirements,while in others they changed substan-tially with every review. For an over-view see Table 2 .

T a

b l e 2 Various models employed by the Commission for managing conditions

Hungary Latvia Romania I Ireland Portugal

How were conditions set?

Distinction betweendifferent types ofcondition

NoneNone, but importanceindicated in one of thereview reports

Prior actions and otherconditions

Permanent and otherconditions None

How were conditions monitored?

How was compliancemonitored? Mixed

1 Mostly by condition By subject matter By subject matter By condition

How were conditions updated?

How were conditionschanged?

Existing conditionsremained valid;addenda containednew conditions only

Each addendum replacedthe previous version

Existing conditionsremained valid;addenda contained newconditions only

Each addendumreplaced the previousversion

Each addendum replaced theprevious version

Persistence ofconditions2

Conditions partiallystable over timeConditions generallyunchanged

Frequent changes toconditionsConditions were changedeven if not maturedDifficult to trace conditionsover time

Conditions always validand unchanged unlessmade obsolete by newconditions

MixedSome conditionsstable, others subject tofrequent change

Conditions stable over timeConditions valid and unchangeduntil fully complied with

1 In some reviews the Commission reported compliance for each condition, in others by subject matter only.2 For additional explanation see Annex II I .

Source: ECA.


41/120


82Some of these models had adverseimplications for programme manage-ment. In particular, frequent changesto conditions in some programmes 48 made it much harder to track the coun-try’s performance. In some instances,reforms were gradually diluted as con-ditions were replaced by less demand-ing ones or completely disappearedfrom the memoranda before they werefully satisfied (see Box 4 ).

83Conditions also sought to bring aboutdifferent structural effects. They mightvary from the publication, or prepara-tion, of a specific strategy or plan, tolegislative changes to be adopted bythe national parliament and affect-ing important parts of the economy 49.While conditions of the first typewould not bring about any meaning-ful economic changes by themselves,although they could serve as prepara-tory steps for more significant reforms.Condition of the latter type wouldprobably mean long-lasting changesto the economy.

Examples of a reform diluted over time

The initial memorandum of understanding in the Latvian programme required the ‘establishment of a singlehuman resource planning and management system for public administration institutions.’ This condition wasnot complied with and did not appear in the subsequent amendments. Instead, it was replaced by severalother conditions, the general aim being ‘the preparation of strategy on public administration’ — a conditionthat is easier to comply with.

The Portuguese programme included a condition requiring — by the end of 2011 — the preparation of anaction plan to reform the unemployment insurance system, notably by reducing the maximum duration ofunemployment insurance benefits to no more than 18 months for newly unemployed persons. After thereform the maximum duration of unemployment benefits was reduced from 38 to 26 months and so exceed-ed the ceiling of 18 months that was initially agreed with the Portuguese authorities.

B o x

4

48 In particular Latvia, but alsoIreland and specic areas inother programmes.

49 Other examples includesignicant changes to theinstitutional landscape, suchas the establishment of a scalcouncil, local governmentreorganisation or privatisation.See Annex III .


42/120


84We found significant differencesbetween conditions in terms ofstringency. Balance-of-payment pro-gramme conditions typically requiredthe enactment of certain legislativechanges. In contrast, the euro area pro-grammes often set only the first typeof condition, and adoption by parlia-ment was seldom required. As a result,decisions on instalments were basedonly on preparatory steps and not onthe actual implementation of reform(see Box 5 ).

85Another difference lay in the numberof conditions, which grew over time.Whereas Hungary was required tocomply with fewer than 60 uniqueconditions, Portugal was faced witharound 400. This large variation didnot only reflect a need for more re-forms in certain countries. We foundinstances where the opposite hap-pened: countries that needed morereforms in a given field were asked tocomply with fewer conditions thanbetter-performing countries 0. The con-ditions also targeted widely differentreform paths despite countries beingin similar circumstances 51.

Programme targets were notrevised in a consistent manner

86Revisions of deficit targets are a natu-ral occurrence in assistance pro-grammes. They do not automaticallysignal poor implementation on thepart of the national authorities, as theymay also be motivated by objectivechanges in the economic situationwith an impact, negative or positive,on the public budget.

Example of a condition linked to preparatory steps only

In the Irish programme, the government was required ‘to introduce legislation for the enhancement of finan-cial regulation, expanding the supervisory and enforcement powers of the Central Bank 52’. The Irish authoritiescomplied with this condition with a minor delay of 1 month when the Ministry of Finance published the draftlaw. However, the enactment took an additional 2 years. There was no condition linked to the enactment andthus the enactment was not necessary for the release of instalments.

52 Referred to as ‘The Central Bank (Supervision and Enforcement) Bill’ in later conditions.

B o x

5

50 For example in the eld ofstructural funds.

51 For example in the eld ofpersonal and corporateinsolvency.


43/120


87Four of the six programmes had theirdeficit targets revised (see Table 3 ). The Commission’s records alwaysexplained downward revisions by a(significant) drop in economic activity,as was the case in Hungary, Romania,Latvia and Portugal.

88Depending on the country, a changein economic outlook can have dif-ferent consequences for the revisionof deficit targets. As part of its fiscalsurveillance, the Commission calcu-lates cyclical budget balances 53. Whenthe economic outlook changes, thecyclical balance estimate is revisedaccordingly. Hence, real-time changesin the cyclical balance estimateshould approximate to the changes indeficit targets made necessary by thechanged economic context 54 .

T a

b l e 3 Revisions of the deficit targets in ESA terms

2008 2009 2010 2011 2012 2013 2014

Hungary

MoU0 (Nov. 2008) – 3.4 % – 2.6 %

MoU1 (Mar. 2009) – 2.9 %

MoU2 (Jun. 2009) – 3.9 % – 3.8 % – 3.0 %

Latvia

MoU0 (Jan. 2009) – 5.3 % – 5.0 % – 3.0 %

MoU1 (Jul. 2009) – 10.0 % – 8.5 % – 6.0 % – 3.0 %

MoU5 (Dec. 2011) – 4.5 % – 2.5 %

Romania I

MoU0 (Jun. 2009) – 5.1 % – 4.1 % – 3.0 %

MoU1 (Feb. 2010) – 7.8 % – 6.4 % N/A – 3.0 %

MoU2 (Jul. 2010) – 7.3 % N/A – 3.0 %

MoU3 (Jan. 2011) – 5.0 % – 3.0 %

Romania II MoU0 (Jun. 2011) – 5.0 % – 3.0 %

Ireland MoU0 (Dec. 2010) – 10.6 % – 8.6 % – 7.5 %

Portugal

MoU0 (May 2011) – 5.9 % – 4.5 % – 3.0 %

MoU5 (Oct. 2012) – 5.0 % – 4.5 % – 2.5 %

MoU7 (Jun. 2013) – 5.5 % – 4.0 %

Note: N/A — not set .

Source: European Commission.

53 The budget balance can bebroken down into a cyclicaland a non-cyclicalcomponent. The cyclicalcomponent captures theeffect of uctuations in taxrevenue and spending due tothe economic cycle.

54 Assuming no signicant errorsin the forecasts of potentialGDP, which is used as an inputvariable for calculating thecyclical balance.


44/120


89Where the cyclical budget balancewas not significantly revised duringthe programme, there was no need torevise the deficit targets. This was thecase for Ireland. However, we foundthat not all revisions of targets wereconsistent with a revision of the cyclic-al balance. For instance, while the 2009deficit target for Hungary was relaxedless than suggested by the revision tothe cyclical balance, the 2012 target forPortugal was relaxed even though thecyclical balance had improved and theeconomic outlook remained the same.

90During the audit the Commissionclaimed that factors outside thegovernment’s control could havetriggered a target revision, such aschanged composition of growth orrevenue shortfalls. However, neitherthe programme documents nor theCommission’s internal methodologyshow a systematic approach to theassessment of these other factorswhen the Commission examined thefeasibility of deficit targets at eachreview.

91Another difference lay in the treatmentof bank support measures with an im-pact on deficits. The deficit targets inthe Portuguese and Irish programmeswere defined as net of such measures,whereas they were included in theLatvian programme targets.

92Programme documents also differedin the attention to the social dimen-sion. For instance, in the Portugueseprogramme, ‘tax increases (were)designed in a progressive way, withlowest-income groups retainingtheir levels of income […] minimumwages and lowest pensions remainuntouched’ 55. This contrasts with theuniversal 25 % cut in civil servants’salaries in the Romanian programme,with no reference given in programmedocuments as to how this would af-fect lower pay grades. Similarly, theoutcome of combining a decrease inflat-rate personal income tax and anincrease in VAT — which was likely toshift the burden of fiscal consolidationto the less fortunate in society — wasnot considered in the initial pro-gramme documents for Latvia.

Programme managementweaknesses due to weakcontrols

93Even though, ultimately, decisionsabout a programme rest with theCouncil and/or the Commission, it isthe Commission’s representatives thatnegotiate with the national author-ities and conduct programme reviews. The mandate for negotiation is set outin the policy brief. During a visit, theCommission’s representatives updatemanagement on the talks, typicallyonce a day, and seek approval for anydeparture from the mandate. After thevisit, the Commission’s representativesproduce the programme documents(Council decision, memorandum ofunderstanding and explanatory note). These are subject to review and furnishthe basis for decisions by the Commis-sion and the Council.

55 European Commission ‘Theeconomic adjustmentprogramme for Portugal’,European Economy , OccasionalPapers 79, 2011, p. 16.


45/120


94Preparation of the policy brief andprogramme documents is an expert-based process in which the judge-ment of forecasters and of a num-ber of policy exper ts from severaldirectorates- general plays a key role. The documents are reviewed by thedeputy director-general and the headsof the horizontal units involved in theirpreparation. The former focuses on thedocuments’ overall consistency, andthe latter are responsible for ensur-ing consistency with EU policy of thecontributions made in their respectivefields.

95Several weaknesses were found in thereview process:

(i) no person outside the programmeteam reviews the underlyingcalculations and projections tocheck the consistency of themacroeconomic projection withthe financing gap estimate or withthe macroeconomic impact of theprogramme conditions;

(ii) in a number of respects, the workof the Commission’s representa-tives was not sufficiently assessed,with too little questioning of theexperts’ forecasts (see paragraph61), financing gap estimates (seeparagraphs 69 and 70) and choiceof programme conditions (seeparagraph 79); and

(iii) internal quality control was gener-ally not documented.

96 The IMF has also set up a process ofinternal review of the key features ofprogramme design and updates. Insome respects this is similar to theCommission’s internal review process,but there are three main differences:

(i) the IMF has a dedicated reviewdepartment which provides inputto management;

(ii) more IMF off icials are involved inthe review process; and

(iii) the IMF process is documented.

Programme documentshad improved with timebut still displayedshortcomings

Improvements in the qualityof programme documents

97 There were marked differences be-tween programmes in terms of theanalytical depth of programme docu-ments. Many of these differences weredue to the Commission’s improvedanalytical capacity. In the first threeprogrammes, the programme docu-ments provided limited informationon a number of aspects, including thecomposition of the fiscal adjustment,analysis of slippages if deficit targetswere not met and social impacts. Theanalysis given in the two most recentsets of programme documents (Ire-land and Portugal) had significantlyimproved (see Box 6 ) and was, in manyrespects,

eca report on financial assistance

Documents