Addressing the net balances problem as a prerequisite for EU budget reform: A proposal *

Transcription

1 Addressng the net balances problem as a prerequste for EU budget reform: A proposal * Angel de la Fuente (Insttuto de Análss Económco, CSIC) Rafael Doménech (Ofcna Económca del Presdente and Unversdad de Valenca) Vasja Rant (Faculty of Economcs, Unversty of Ljubljana) February 2008 Abstract Conflct among member states regardng the dstrbuton of net fnancal burdens has been allowed to contamnate the entre desgn of the EU budget wth very negatve consequences n terms of equty, effcency and transparency. To get around ths problem and pave the way for a substantve budget reform, we propose to decouple dstrbutonal negotatons from the rest of the budget process by lnkng member state net balances n a rgd manner to relatve prosperty. Ths would be acheved through the ntroducton of a system of compensatng horzontal transfers that would take to ts logcal concluson the Commssons proposal for a generalzed compensaton mechansm. We dscuss the mpact of the proposed scheme on member states ncentves and llustrate ts fnancal mplcatons usng revenue and expendture projectons for 2013 that are based on the current Fnancal Perspectves and Own Resources Decson. Keywords: EU budget JEL Classfcaton: H87 * Ths paper s part of a research project co-fnanced by the ERDF and Fundacón Caxa Galca. Addtonal fnancal support from the Spansh Mnstry of Educaton under research grants SEJ and SEJ s also gratefully acknowledged. The vews expressed n the paper are exclusvely those of ts authors and do not n any way represent the offcal poston of the Spansh government.

2 1. Introducton In ts 2005 Brussels meetng the European Councl unanmously approved to undertake a comprehensve revew of the EU budget that would cover all expendture programmes as well as the Unons fnancng system. Ths agreement can be seen as a manfestaton of a wdespread dssatsfacton wth the structure and outcomes of the European budget process that s shared both by the governments of member states and by ndependent commentators. There seems to be an emergng consensus that the EU budget, as t stands, does not adequately meet Europes needs and challenges. Whle there s certanly dsagreement about specfcs, many polcymakers and analysts do agree that way too much money s beng spent on certan thngs and not nearly enough on others and have questoned the value added of dfferent expendture programmes. 1 In addton, the EU budget has often been crtczed for ts lack of transparency for European ctzens and even member governments and for the questonable dstrbuton of net fnancal burdens that t acheves. General rules on the dstrbuton of expendtures and fnancal contrbutons across member states are often volated by ad-hoc exceptons that ntroduce a consderable degree of complcaton and arbtrarness. Partly as a result of these exceptons and partly due to the nature of EU expendture programmes, member states wth smlar ncome levels often end up wth very dfferent net fnancal postons. But f ths s wdely accepted, why has there not been a serous budget reform for the last twenty years? The answer to ths queston brngs us to a serous structural problem that needs to be solved before we can hope to make any progress on more substantve ssues. Its essence s that conflct among member states over the dstrbuton of net fnancal burdens has been allowed to condton the desgn of the entre European budget. The frst part of ths note brefly dscusses the nature of what we wll call for short the net balances problem and explores ts mplcatons for the EU budget process. Buldng on de la Fuente and Doménech (2001), we then advance a proposal for a reform of the Unon s fnances that should allevate ths problem and help brng about a more equtable and effcent budgetary polcy. 2. The net balances problem and ts mplcatons The basc desgn of the European budget, on both ts expendture and revenue sdes, s decded at an ntergovernmental meetng, the European Councl, where all actors have veto power. Experence has shown that member states behavour n budget negotatons has often been drven by concerns over ther net fnancal returns, understood n a very narrow sense as the dfference between ther respectve contrbutons to the fnancng of the Unon and the return flow of expendtures n ther terrtory that are fnanced by the European budget. As a result, Councl meetngs have been domnated by dstrbutonal ssues and member states have not 1 Among many others, see Begg (2005), Sapr et al (2003), Boldrn and Canova (2001) and DEFRA (2005). The frst two references nclude general analyses of the defcences of the EU budget and the last two are very crtcal dscussons of the two core EU polces: coheson and agrculture. 1

3 hestated to threaten to block budget agreements n order to protect ther perceved natonal nterests. Needless to say, ths has made negotatons rather complcated. More mportantly, the ongong dstrbutonal conflct among member states has serously dstorted budget outcomes n ways that nvolve large losses n the effcency, equty and transparency of EU fnances. A key obstacle to substantve budget reform, Perhaps the most mportant cost of the net balances problem s that t has been a very serous obstacle to substantve budget reform. A clear llustraton can be found n the negotatons leadng to the approval of the last two mult-annual EU budgetary frameworks (the so-called fnancal perspectves). In both cases there was wdespread agreement on the need to devote addtonal resources to areas of common European nterest n order to respond to the economc and poltcal challenges posed by a changng nternatonal envronment. In both cases what were n fact rather tmd proposals by the Commsson for ncreased fundng n key areas had to be severely curtaled. The requred money smply could not be found because net contrbutors blocked any ncrease n the overall budget celng for fear of havng ther defcts ncreased, and the man benefcares of exstng expendture programmes or fnancng prvleges strongly ressted any attempt to curtal ther fundng. As a result, the fnal outcome looked very much lke the status quo wth only margnal changes n the structure of the budget. wth a heavy cost n terms of smplcty and transparency A second mportant cost s that the sde payments necessary to secure the approval of a fnancal package have resulted n an ncreasngly complex and untransparent budget, rddled wth preferental clauses and ad-hoc exemptons to general rules that would be very dffcult to explan to the average European ctzen. As the stuaton stands now, there s a rather bg and questonable excepton on the revenue sde of the budget -- the rebate to the UK of two thrds of ts ex-ante defct-- whch nvolves addtonal exceptons through the dscounts granted to Austra, Germany, Sweden and the Netherlands n the fnancng of the rebate. In addton, these four countres enjoy reduced rates of payment for the part of ther natonal contrbutons that s lnked to the VAT base and two of them have been granted specal reductons n ther GNI-based contrbutons. On the expendture sde, the stuaton s not very dfferent, partcularly n connecton wth the allocaton of coheson fundng. The current fnancal perspectves are full of specal provsons n favour of specfc countres or regons that volate the general allocaton rules set out n the same document (see CEU 2005). These specal provsons affect at least 16 of the member states of the EU. whch does not guarantee an equtable dstbuton of net fnancal burdens Such messness mght be acceptable f t served to acheve an equtable dstrbuton of net fnancal burdens among member states. Lookng at the numbers, however, t s hard to argue that ths has been the case. 2

4 Fgure 1 shows the relatonshp between member states relatve prosperty and ther net fnancal postons n As an ndcator of relatve prosperty we use ncome per capta n normalzed PPS unts 2 (roughly speakng, n euros of average purchasng power) measured n percentage devatons from the EU average. Fnancal postons are measured by what we call per capta relatve real balances. To construct ths varable, we frst compute each countrys per capta net balance by subtractng ts contrbuton to the EU from EU expendtures allocated to t, wth both varables measured n normalzed PPS unts per capta. From ths we subtract the average per capta net balance of the entre EU, whch wll generally be negatve because certan types of expendture (lke foregn ad) cannot be allocated to any member country. Fnally, the result of these calculatons s normalzed by average EU ncome per capta n normalzed PPS unts. Fgure 1: Per capta relatve real balances vs. relatve real ncome per capta n Key: At = Austra; Be = Belgum; Cy = Cyprus; Cz = Czech Republc; De = Germany; Dk = Denmark; Ee = Estona; El = Greece; Es = Span; F = Fnland; Fr = France; Hu = Hungary; Ie = Ireland; It = Italy; Lt = Lthuana; Lu = Luxembourg; Lv = Latva; Mt = Malta; Nl = Netherlands; Pl = Poland; Pt = Portugal; Se = Sweden; S = Slovena; Sk = Slovaka; UK = Unted Kngdom. Appled to the case at hand, commonly accepted notons of equty would requre that net fnancal burdens should be dstrbuted n proporton to ablty to pay. Hence, rcher member states should pay more than poor ones, and countres wth the same level of real ncome should have smlar fnancal postons. The fgure shows that the EU budget roughly meets the frst of these crtera but certanly not the second. The negatve slope of the regresson lne we have ftted to these data (weghtng countres by populaton) ndcates that the net effect of the EU budget s redstrbutve, as relatve balances 2 Throughout the paper, we have renormalzed member state relatve prce ndces so that the prce level of the EU as a whole s equal to 1 n each year. We apply the same correcton to ncome levels and budget flows expressed n PPS unts to convert them to normalzed PPS unts. 3

5 are on average postve n poor countres and negatve n rch ones. The slope of ths lne, whch we wll call the redstrbuton coeffcent, s Gven the defnton of our varables, ths parameter ndcates that, for a representatve European ctzen, the net effect of the EU budget s equvalent to a flat tax of 1.75 leved on the dfference between hs ncome (adjusted for purchasng power dfferences) and the EU average -- or to a subsdy of the same magntude f hs ncome s below the average. On the other hand, countres of smlar wealth often end up wth very dfferent fnancal postons. Greece, Ireland and Luxembourg are extraordnarly well treated gven ther ncome levels. France and Germany have approxmately the same ncome per capta, but the latters defct s roughly twce the sze of the formers. The same s true of the UK and Sweden or Austra and Denmark. On the opposte end of the ncome scale, the per capta surpluses of Greece and Malta are sx tmes larger than those of Slovena and the Czech Republc. 3. The way forward: changng ncentves by solatng redstrbutonal ssues The European Commsson and many analysts have argued repeatedly that net fscal balances are a rsky accountng exercse that, at best, captures n a rather mperfect way only a small part of the benefts of membershp n the European Unon. 3 Whle ths s certanly true, we argue that these balances cannot be gnored. However mperfectly, they do measure the most vsble and easly quantfable aspect of such benefts. As a result, the vews of European ctzens and member state governments on the overall farness of the system are lkely to depend crtcally on the extent to whch the dstrbuton of such balances s perceved to be reasonable. And however msguded ths concern may be n some cases, t s unlkely that we can make t go away just by preachng about the need to take a broader vew of natonal nterests. A more realstc approach to the problem s to take member states concern wth net balances as gven and ask what can be done to mnmze ts undesrable sde effects. A possble way out of the current deadlock would be to change the structure of the EU budgetary system n such a way that the unavodable conflct over dstrbutonal ssues can be solated and does not spll over nto the rest of the budget dscusson. Ths requres the ntroducton of a new budgetary nstrument that can be used to neutralze the undesred dstrbutonal consequences of core EU polces. The logcal choce would seem to be a system of horzontal transfers across member states. In our opnon, such a system should be desgned to lnk member country net balances rgdly to ther levels of relatve prosperty. We beleve the adopton of such a system would help mprove the equty and transparency of the EU budget whle openng the way for sgnfcant effcency gans by changng member government ncentves n a manner that would make substantve reform possble. A smple rule lnkng net balances to per capta ncome levels would greatly smplfy the European budget and ncrease ts transparency by elmnatng the need to dstort expendture polces n order to acheve an acceptable dstrbuton of net fnancal contrbutons. It would also ensure that ths 3 See for nstance CEC (1998 and 2004). 4

6 dstrbuton wll be n strct accordance wth a smple equty prncple that s easy to explan and can be supported by a large majorty of European ctzens. More mportantly for our purposes here, the ntroducton of such a rule would also radcally change member governments ncentves n budget negotatons. Under the current stuaton EU expendtures are perceved as havng an almost zero margnal cost to natonal treasures. Hence, member state representatves enter Councl meetngs wth an ncentve to fght for every possble ncrease n spendng favorng ther country -- and ths ncludes programs they would probably not be wllng to fnance wth ther own budgets. If net balances are set n advance, so that expendture gans n any gven program wll have to be fnanced at the margn by natonal budgets, ths perverse ncentve dsappears. Ths has two mportant mplcatons. The frst one s that fundng for horzontal programmes can be allocated more effcently because the Councl or the Commsson need not worry about ther mpact on member state net balances. The second s that natonal governments are lkely to become more selectve n ther support for EU actvtes. Although we beleve that ths would be a postve development n general terms, there s some rsk that t may lead to the curtalment of those programs that are valued more by the Commsson than by member states. To the extent that the Commsson defends truly European nterests, rather than natonal ones, ths would be an undesrable outcome. 4. A specfc proposal and ts fnancal mplcatons In ts 2004 report on the fnancng of the Unon (CEC 2004), the European Commsson has called for the ntroducton of a generalzed compensaton mechansm for member countres experencng excessve defcts. The Commsson s central proposal essentally nvolves extendng a streamlned verson of the UK rebate to all net contrbutors n a smlar stuaton. Member countres experencng defcts over 0.35 of GNI would be granted a rebate of two thrds of the excess over ths threshold. These rebates would be fnanced by all member states n proporton to ther natonal ncome. The total amount of compensatory payments would be capped at 7.5 bllon euros. If ths maxmum amount were to be exceeded, rebate rates would be reduced n the proporton needed to respect the cap. Whle the Commsson s proposal would be a step n the rght drecton, n our opnon t does not go far enough. It tackles only one sde of the problem (excessve defcts but not excessve surpluses), t does so only partally and wthout takng nto account the relatve prosperty of net contrbutors and ntroduces an unnecessary dscontnuty n the form of a fxed defct threshold below whch no correctve acton would be taken. As a result of all ths, ts contrbuton to budget equty wll be qute lmted, as t wll mtgate but not elmnate dfferences of treatment across member states only at one tal of the ncome dstrbuton. Its effects on ncentves wll also be reduced by ts lmted applcablty to a small subset of member countres, by the cappng mechansm and by the threshold and partal rebate features. Our proposal nvolves takng the prncple behnd the Commsson s proposal to ts logcal concluson. Ths prncple, whch was already establshed n the 1984 Fontanebleau Councl, 5

7 requres that no member state should sustan a budgetary burden that s excessve n relaton to ts relatve prosperty. The smplest and most straghtforward way to acheve ths s to make sure that budgetary burdens, measured n net terms, are proportonal to relatve prosperty. Hence, we would advocate the establshment of a budgetary rule that would make relatve real balances per capta, as defned above, nversely proportonal to member states relatve ncome per capta n real terms (n PPS unts) measured n devatons from the EU average. In terms of Fgure 1, our proposal would amount to makng sure that all countres are on the estmated regresson lne, or on some other agreed upon lne gong through the orgn. Ths can be acheved through a system of horzontal transfers across member states that would fully compensate any devatons from the desred allocaton rule that would result from the standard fnancng mechansms of the Unon and from ts expendture polces. These compensatng transfers would be calculated and pad followng the same procedures that are now used n the case of the UK rebate. A frst estmate of the compensatory transfer correspondng to year t would be pad at t+1 and fnal settlement would come at t+4, wth the possblty of ntermedate correctons as the fnal data requred for the calculaton become avalable. Under ths scheme, all member states wth real ncomes below the EU average would receve a net transfer from the rcher members of the Unon. Such transfers should, however, be subject to strct condtonalty and addtonalty requrements to ensure that they are used to fnance growth-enhancng nvestments that wll help mtgate ncome dspartes. To acheve ths, allocatons under coheson polces should be at least equal to the desred fnal net balances for all countres wth below average ncomes. In addton, these countres should not be allowed to use compensatory transfers as a way to free up coheson funds for consumpton purposes. To prevent ths, compensatng transfers should be calculated on the bass of expected rather than actual coheson polcy dsbursements. Adoptng the procedure we have just outlned would effectvely splt budget negotatons nto two separate parts. On the one hand, member countres wll have to bargan over the desred level of fnancal soldarty, as measured by the redstrbuton parameter that lnks net balances to relatve ncome (that s, by the slope of the regresson lne shown n Fgure 1). On the other, expendture programmes and the standard fnancng mechansm of the Unon wll have to be dscussed. The man advantage of the setup s that member countres wll be able to thnk about how much money the Unon needs and how t should be spent wthout the dstracton of worryng about how such decsons wll affect ther own fnances. As shown n the Appendx, the fnal contrbuton of each member state to the EU budget (net of compensatory transfers) under the proposed scheme wll be approxmately equal to the sum of three components. 4 The frst two amount to full repayment of all EU expendture allocated to the country and the equal per capta sharng n real terms of overhead or general-nterest EU expendtures (ncludng external acton and admnstraton). The last component wll be a 4 The equalty s not exact because compensatory transfers, whch are estmated n real terms, have to be adjusted so that they add up to zero when measured n current euros. See the Appendx. 6

8 redstrbutve payment whose per capta amount wll be a functon of the dfference between the country s real ncome per capta and the EU average. Hence, the margnal cost to a member state of allocated expendture wll be equal to one, unlke n the present system where t s consderably below one snce all countres bascally fnance all EU expendtures n proporton to ther share n nomnal aggregate ncome. As a result, member countres wll have no ncentve to approve expendtures they would not be wllng to fnance wth ther own budgets and wll have the correct ncentves to channel through the European budget only those programmes where the EU can provde some added value. Fnancal mplcatons: a tentatve estmate To llustrate the fnancal mplcatons of our proposal, we wll use the estmates of budget flows across EU member states n the fnal year of the current Fnancal Perspectves that we have constructed n a companon paper (de la Fuente, Rant and Doménech, DRD 2008) workng wth the country allocatons establshed n those Perspectves (CEU, 2005), the Commssons growth forecasts contaned n the Agng Report (ECP, 2006) and our own projectons of the evoluton of relatve prces n EU member states. Needless to say, the exercse s extremely rsky but t wll gve us an dea of the order of magntude of the requred transfers and of the expected drecton n whch they wll flow. Fgure 2: Per capta relatve real balances vs. relatve real ncome per capta n 2013 relatve real balance per capta Ro Bg Pl Hu Lt/Ee Sk Lv Pt Mt Cz El Sl Cy Es It Fr F Dk Nl/At Se De UK Be Ie Lu relatve real ncome per capta - Key: Bg = Bulgara; Ro = Romana. See Fgure 1 for the rest. - Source: de la Fuente, Rant and Doménech (2008). Fgure 2 shows the relatonshp between member states expected relatve per capta real balances and expected real ncome per capta n 2013 along wth the ftted regresson lne that captures the average degree of redstrbuton through the EU budget. The value of the redstrbuton coeffcent for ths year (2.68) s relatvely low by hstorcal standards (see 7

9 DRD 2008) but consderable hgher than the extremely low values observed n the transton years followng Eastern enlargement. Table 1: Estmated excess balances and compensatng transfers, 2013 keepng constant the observed level of redstrbuton (at 2.68) Excess balances Compensatng transfers. 1. per capta as a of EU avge. GNIpc 2. per capta n PPS unts 6. adjusted total n meuros 3. total n mllon PPS 4. relatve prce ndex 5. total n meuros Luxembourg Ireland , ,007-2,099 Estona Hungary , ,643-1,718 Czech Rep , ,721-1,799 Lthuana Latva Slovena Slovaka Greece , ,097-1,147 UK , ,482-3,641 Austra Denmark Poland , ,041-1,089 Netherlands Portugal Fnland Sweden Bulgara Span Belgum France , ,298 2,194 Germany , ,308 5,066 Malta Cyprus Italy , ,127 5,848 Romana , ,478 1,411 total extra contrbutons due: 16,964 14,905 15,584 total postve transfers due: 16,964 16,327 15,584 transfers - contrbutons 1,423 adjustment factor 4.55 new entrants coheson 4-4,825-2,914 rest of EU15 7,739 - Note: Columns 5 and 6 n mllons of 2004 euros. As a startng pont, we wll take as gven the exstng degree of redstrbuton through the European budget and calculate the amount of the compensatng transfers that would be requred to move all countres to the estmated redstrbuton lne. Table 1 shows the results of ths calculaton. Column 1 shows member countres excess balances measured as a fracton of 8

10 average EU GNI per capta n normalzed PPS unts. Excess balances are defned as the dfference between observed and desred relatve balances. In graphcal terms, they correspond to the vertcal dstance between the ftted regresson lne and the dot representng each country s poston n Fgure 2. Column 2 converts these amounts to normalzed PPS unts per capta. Multplyng these fgures by each country s populaton, we obtan ts total excess balance n PPS unts, whch s shown n column 3. Next, we multply ths amount by the ndex of relatve country prces shown n column 4 and change ts sgn to obtan a prelmnary estmate of the compensatng transfer due to each country expressed n nomnal terms (n euros of 2004). Hence, countres wth postve excess balances would be assgned negatve transfers, that s, would have to pay addtonal contrbutons to elmnate ther excess balances. A complcaton that arses at ths pont s that whle excess balances n normalzed PPS unts add up to zero by constructon (because they are the resduals of a regresson), compensatng transfers n current euros do not necessarly do so. The dfference between the frst round estmates of total postve net transfers and total extra contrbutons under the compensaton scheme s calculated n the lower part of the table and amounts to 1.4 bllon euros. To get thngs to balance, we reduce all transfers and ncrease all contrbutons n the unform proporton requred for net payments to add up to zero. The value of the adjustment coeffcent that wll do the trck s Adjusted compensatng transfers are shown n column 6. The bottom part of ths column shows the mplcatons of the compensatng transfer scheme for three groups of countres. Those countres that joned the Unon n 2004 and 2007 would collectvely lose 4.8 bllon euros whle the tradtonal coheson countres or C4 (Span, Ireland, Portugal and Greece) would have to pay addtonal contrbutons amountng to 2.9 bllon. These payments would fnance a net transfer of 7.7 bllon to the remanng members of the EU15. There would be both wnners and losers n each group of countres. Wthn the new entrants, Romana, Malta and Cyprus would receve postve compensatng transfers, as would Span wthn the second group. As for the group of rcher countres, Luxembourg, the UK, Austra, Denmark, the Netherlands, Fnland and Sweden would have to ncrease ther net contrbutons, whle Germany, France, Italy and Belgum would be enttled to sgnfcant refunds. Our results ndcate that the wealther member states would be the man benefcares of the mplementaton of the compensaton scheme f the expected degree of redstrbuton n the year of reference were to be held constant. If ths outcome s deemed undesrable, the redstrbuton coeffcent should be rased. Table 2 llustrates the senstvty of the net postons of the three groups of member states mentoned above to changes n the redstrbuton coeffcent and shows the total cost of the scheme as measured by the total amount of postve (or negatve) compensatng transfers. An ncrease n the redstrbuton coeffcent to 3.0 would reduce the 5 Let T be total (postve) transfers due and C total extra contrbutons (negatve transfers). Snce the requred transfers exceed contrbutons, we have (1) T C = X > 0. We seek the value of the adjustment factor a such that (2) (1-a)T (1+a) C = 0. Substtutng (1) nto (2) and solvng for a, we have a = X/(T+C). 9

11 net cost to new entrants to 2.2 bllon. A further ncrease to 3.5 would make the scheme neutral for the rcher group of countres as a whole and would result n a transfer of 1.7 bllon from the C4 (mostly from Ireland) to the Eastern European jonng states. Increasng the redstrbuton coeffcent would also rase the total cost of the scheme but only modestly, at least wthn the range of values contemplated n Table 2, because ncreasng transfers to the poorer member states tends to be relatvely cheap n nomnal terms due to ther low prce levels. Table 2: Senstvty of results to changes n the redstrbuton coeffcent redstr. coeffcent = total transfers to groups of member countres: new entrants -4,825-2,238 1,832 5,990 coheson 4-2,914-2,452-1,739-1,031 rest of EU15 7,739 4, ,959 total volume of transfers 15,584 15,697 17,504 19,507 In our opnon, a good case can be made for rasng the redstrbuton coeffcent to a level around 3.5. As noted, a lower redstrbuton coeffcent would mply that compensatng transfers would flow on balance from poor to rch countres, whch may be dffcult to justfy. In addton, such a change would stll leave us far below the levels of redstrbutve effort that prevaled n the years mmedately precedng Eastern Enlargement, whch ranged between 5 and 6. Table 3 shows the detaled mplcatons of the compensatng scheme wth a 3.5 redstrbuton coeffcent. One of the man changes relatve to the scenaro descrbed n Table 1 s that compensatng transfers to Poland, Bulgara and Portugal change sgn and become postve. The estmated volume of compensatng transfers s consderable n both scenaros. Under the assumptons underlyng Table 3 the total transfer of resources across member states would amount to 17.5 bllon euros at 2004 prces, whch s more than twce the maxmum cost contemplated by the Commsson n ts proposal for a generalzed compensaton mechansm. Column 6 shows the mpact of estmated compensatng transfer on each state s total contrbutons to the Unon. Some of the requred transfers are also qute large when measured n ths manner. For nstance, the UK s 6.9 bllon contrbuton to the compensatng scheme would amount to a 40 ncrease n ts total payments nto the EU budget, whle Poland s 1.2 bllon refund would reduce ts net contrbuton by over 35. Snce the fgures are even hgher for some other member states, t would be advsable to phase n compensatory transfers over a relatvely long perod. We would argue for a 10-year transtory perod durng whch the call rate on the theoretcal value of the compensatng payments would rse n even steps from 10 to 100. We would expect that changes n expendture and revenue polces durng ths perod would reduce the volume of compensatng transfers to values sgnfcantly below those shown n our llustratve tables. 10

12 Table 3: Estmated excess balances and compensatng transfers, 2013 Increasng the level of redstrbuton to 3.5 Excess balances Compensatng transfers. 1. per capta as a of EU avge. GNIpc 2. per capta n PPS unts 5. adjusted total n meuros 6. mpled chge. n total contrbutons 3. total n mllon PPS 4. total n meuros Luxembourg Ireland ,117-2,473-2, Czech Rep ,905-1,377-1, Estona Hungary ,656-1,148-1, Lthuana UK ,819-7,549-6, Slovena Austra Latva Denmark Netherlands ,358-1,452-1, Greece Slovaka Sweden Fnland Belgum France , Span Portugal Germany ,147 3,293 3, Poland ,850 1,232 1, Italy ,961 6,057 6, Bulgara Cyprus Malta Romana ,133 2,989 3, total extra contrbutons due: 19,175 19,044 17,504 total postve transfers due: 19,175 16,195 17,504 transfers - contrbutons -2,849 adjustment factor new entrants 1, coheson 4-1, rest of EU Note: Columns 4 and 5 n mllons of 2004 euros. 5. How should net balances be calculated? The estmates of net budget balances we have used n our llustratve calculatons have been constructed usng essentally the same procedure the European Commsson currently follows to allocate ts revenues and expendtures to member countres. 6 We have proceeded n ths manner for smplcty and to facltate the comparson of our fgures wth other estmates of 6 See for nstance Annex IV of CEC (2007). 11

13 member country budgetary mbalances. We would argue, however, that certan aspects of ths procedure should be modfed before t can be used for the calculaton of compensatory transfers. There are two ssues that need to be addressed. The frst one has to do wth the crtera used to allocate certan expendture and revenue tems across member states, and the second wth the desred treatment of unallocated expendture. Regardng the frst ssue, the only problem that arses on the revenue sde has to do wth the treatment of tradtonal own resources (customs dutes and agrcultural leves net of the allowance for collecton costs). We have treated these tems as natonal contrbutons and assgned them to the member country that collected them, but t s clear that allocatng tarff revenue of the bass of port of entry does not adequately capture the real dstrbuton of the fscal burden on mports. 7 Hence, t would be preferable to allocate ths revenue n proporton to GNI or to consumpton. EU expendtures are allocated among member states by the Commsson s servces on the bass of where the money s physcally spent but ths s not always a good measure of who benefts from t. One mportant example has to do wth admnstratve costs: whle most EU staff lve and work n Belgum and Luxembourg, the work they carry out benefts all EU ctzens n a smlar way. As a result, t would not be reasonable to allocate ther salares only to these two countres. As the Commsson tself does when calculatng what t calls operatng balances, we have excluded admnstratve expenses from our calculatons so that member state net balances are unaffected by the geographcal locaton of EU nsttutons. A smlar adjustment would probably make sense also n the case of other budget tems that generate large external benefts outsde of the mmedate geographcal area where they are spent (as determned by the resdence of the prncpal recpent of each payment). In partcular, expendture on border control, securty, mmgraton, publc health, consumer protecton and basc research should probably be treated n the same way as admnstratve expenses and excluded from the calculaton of net balances for purposes of determnng the amount of compensatory payments. Fnally, there s the queston of how to deal wth unallocated expenses, ncludng foregn ad and other external expendture as well as admnstraton and the other tems we have just mentoned. As t s shown n the Appendx, under our proposed scheme the per capta burden of fnancng unallocated expendture (that s, expenses of general nterest and overhead costs) wll be equal n real terms for all member countres. In our opnon ths s not an unreasonable sharng rule, but a case could also be made for the fnancng of such expendtures n proporton to (real or nomnal) ncome. If ths second opton s consdered preferable, unallocated expendture should be mputed n proporton of GNI rather than gnored n the relatve balance calculatons. 7 For nstance, Holland and Belgum collect a dsproportonate share of tradtonal own resources because a large fracton of mports from outsde the EU comes through ther ports. Ths s the so-called Rotterdam or gateway effect. Ths effect may be partally offset by another one that would work n the opposte drecton. Ths reverse gateway effect arses because tax rebates on agrcultural exports to non-eu countres from Belgan and Dutch ports wll tend to be attrbuted to these countres regardless of the orgn of these products. Hence, export rebates should probably be allocated n proporton to the producton of the relevant agrcultural commodtes. 12

14 6. Concluson We advocate the ntroducton of a system of compensatory transfers that wll make the net balances of EU member states nversely proportonal to ther relatve ncome leveles and the adopton of a mult-stage budget-settng procedure consstent wth t. Under the proposed system, barganng over the desred level of redstrbuton would be decoupled from the rest of the budget negotaton n order to make t easer for member state governments to dscuss core EU revenue and expendture polces on ther own merts and not n terms of ther mpact on natonal fscal balances. Any dscrepances between the fnancal outcome of such polces and the desred dstrbuton of net balances would be elmnated by a compensaton mechansm that could take the form of a system of horzontal transfers across member states. We have also advocated a revson of the rules currently used to allocate expendtures and contrbutons to member countres n order to brng such estmates closer nto lne wth the real ncdence of certan budget tems for whch the pont of collecton or the locaton of expendture are not good ndcators of who reaps the benefts or bears the costs. The system we propose would have several mportant advantages over current practces. It would provde a transparent way of mplementng the prncple of coheson wthout sacrfcng other polcy objectves. It would also reduce the scope for real or perceved nequtes n the allocaton of budget resources by forcng member states to negotate over a sngle, easly nterpretable, parameter rather than about a host of specfc tems that may add up to apparently arbtrary allocatons. In addton, settng net balances n advance of expendture programs wll n effect force natonal governments to bear the margnal cost of EU expendture, thereby ncreasng the ncentves for a more effcent allocaton of resources. These advantages do come at a cost, however. The proposed procedure may actually ncrease the dffculty of reachng a budget consensus by sharply reducng the margn for "horse tradng" across member states, and could make t harder for the Commsson to push through proposals n areas n whch European and natonal nterests do not concde. 13

15 Appendx Ths Appendx descrbes n detal how relatve balances are calculated and dscusses some of ther propertes and the mplcatons of ther use as part of our proposed compensatng scheme. To facltate the dscusson, secton 1 goes through the analyss n nomnal terms. In secton 2 we dscuss the complcatons that arse when we take nto account dfferences n prce levels across countres and explan how we deal wth them. 1. Relatve balances n nomnal terms Let C = country s contrbutons to the EU budget AE = EU expendture allocated to country C =! C = total EU revenue from member state contrbutons I AE =! AE = total EU expendture allocated to member countres I NAE = total EU expendture not allocated to member countres N = country s populaton N =! N = total EU populaton I All expendture and revenue tems defned above wll be measured n current euros wthout takng nto account dfferences n prce levels across countres. We wll use lower case characters to denote per capta amounts, so that, for example ae = AE /N The EUs budget defct wll be gven by (1) D = C - AE - NAE. Country s net budgetary balance (NB ) wll be defned as the dfference between the expendture allocated to t and ts contrbuton to the EU budget, that s (2) NB = AE - C Notce that the sum of the net balances of all member countres wll be generally dfferent from zero even wth a zero defct because some EU expendture (e.g. foregn ad) s not allocated to any member state. In partcular, (3) NB =! NB =! AE "! C = AE " C = "(NAE + D) where we have made use of equaton (1). 14

16 Relatve budgetary balances We wll work wth net balances measured n relatve terms. A key advantage of ths procedure s that t allows budget balances to be decomposed addtvely nto a seres of partal balances that can be attrbuted to ndvdual expendture and contrbuton tems. Snce we wll not make use of ths decomposton n the present paper, we wll focus on other propertes of relatve balances. Formally, we defne country s per capta relatve balance wth the Unon as (4) rnb = nb! nb where nb = NB/N s the average per capta net balance of the entre Unon wth NB as defned n equaton (3). It s mportant to make explct how relatve balances deal wth unallocated expendture. Notce that country s total relatve net balance wth the Unon can be wrtten " (5) RNB = rnb * N = ( nb! nb)* N = AE! C # N! AE! C N & * N = AE! C ( ) + N (NAE + D) N where we have used (3). Equaton (5) shows that a country s relatve balance would be equal to ts standard net balance f non-allocated expendtures and the Unon s defct were mputed to member states n proporton to ther populaton, so that per capta allocatons would be the same for all countres. Notce that the sum of member countres relatve balance s zero: ( ) (6)! RNB =! AE " C + (NAE + D)! N N = AE " C + NAE + D = 0 Member states ncentves under the exstng fnancng system To smplfy matters, let us assume that the Unon s defct s zero and that member states contrbutons to the Unon are strctly proportonal to ther natonal ncomes so that (7) C = cy where Y s country s GNI and c a call rate that s equal for all countres. Gven the domnant weght of the GNI resource and the correctons that have been ntroduced to brng the base of the VAT resource closer to GNI, ths would be a good approxmaton n the absence of ad-hoc exceptons to general budget rules lke the UK rebate. To preserve budget balance, c must be set so that (ex-ante) contrbutons are equal to total expendtures, (8) cy = AE + NAE! c = AE + NAE Y In equlbrum country s total contrbutons wll be gven by 15

17 (9) C * = (AE + NAE) Y Y = # AE & + NAE + " AE j ( Y Y and ts net balance wll be j! (10) NB * = AE! C * == AE! & AE + NAE + # AE j ) Y ( Y = AE 1! Y & Y ( )! & NAE + # AE j ) Y ( Y j" Hence, all member states contrbute to the fnancng of all expendtures n proporton to ther shares n aggregate EU ncome. Countres have an ncentve to try to maxmze ther own allocated expendture because ts margnal cost s very low, partcularly n small and relatvely poor countres. Ths s true even f a dollar of such expendture s valued less than a dollar of contrbutons or of drect expendture by the natonal government. j" Member states ncentves under the proposed compensaton scheme Under the scheme proposed n ths paper, countres would pay or receve compensatory transfers that would have to be added to ther regular budget contrbutons. The amount of the compensatory transfer payable to country, T, wll be gven by (11) T = N t = N [!"(y! y)! rnb ] = N ["(y! y )! rnb ] where ρ s the slope coeffcent of the redstrbuton regresson (see Fgure 1 or 2 n the text). Notce that countres receve postve transfers f ther ncome per capta s below the Unon average (y < y) and have to make addtonal contrbutons (receve negatve transfers) otherwse. Operatng wth ths expresson and usng equaton (5) (12) T = N [!(y " y ) " rnb ] =!(y " y )N " RNB =!(y " y )N " ( AE " C ) " N (NAE + D) N Now, country s total adjusted contrbutons wll be gven by (13) C ** = C! T = C! "(y! y )N + ( AE! C ) + N N (NAE + D) = AE + N N (NAE + D) + "(y! y)n and ther net balances by (17) NB ** = AE! C ** = "(y! y )N! N (NAE + D) N Hence, under the proposed compensatng scheme, each country wll pay for all expendture allocated to t and for the fracton of "overhead expenses" (non-allocated expendture and the budget defct) that corresponds to t by populaton. In addton, there s a redstrbutve component of total contrbutons that s proportonal to the countrys ncome gap wth the EU average. Net balances are ndependent of own allocated expendture and depend only on the level of ncome and the amount of overhead expendture. 16

18 2. Adjustng for prce dfferences Dfferences n prce levels across EU member states are very mportant. Settng the average prce level of the entre EU to 100, relatve prces n 2005 ranged from Denmark s 134 to Lthuana s 50. As a result, comparsons of ncome levels wth data n nomnal euros can be very msleadng. Smlarly, expressng measures of natonal contrbutons and net balances at current natonal prces wll not adequately capture the sacrfces or benefts they ental. To avod these problems we have measured the ncome of EU member states and the budget flows across them n real terms, that s, n normalzed PPS unts or euros of equal purchasng power. Ths secton shows how prce correctons affect the calculatons and results descrbed n the prevous secton of ths appendx. We wll use prmes to denote varables measured n PPS unts normalzed so that the average prce level of the EU s equal to 1 n each year. Hence, Y wll be country s real natonal ncome,.e. GNI n normalzed PPS unts, and Y wll denote the same magntude measured n nomnal terms (current euros). By constructon, aggregate EU ncome wll be the same n real and nomnal terms,.e. Y = Y. The ndex of relatve country prces wll be gven by the rato of nomnal to real GNI (.e. of GNI n current euros to GNI n normalzed PPS unts) (15) P = Y Y Natonal prce ndces wll be normalzed so ther weghted average s equal to one: " (16) P = Y Y Y! 1 = Y = " P Y Y Y = " P. Y For most of our calculatons t wll be convenent to work wth the nverse of the prce ndex,.e. wth the relatve deflator, whch we wll denote by Q : (17) Q = 1 P = Y Y! Y = Q Y Real balances Measured n real terms, country s net balance wll be gven by (18) NB = NB Q = (AE! C )Q = AE! C = (ae! c )N = nb N Proceedng as before, but workng wth real magntudes, we can defne per capta and total relatve balances. We have, n partcular, (19) RNB = rnb N where relatve real balances per capta are calculated as before but workng wth allocated expendtures and contrbutons measured n normalzed PPS unts 17

19 (20) rnb = nb! nb and nb s the average per capta real balance of the entre Unon (21) nb = NB N = 1 "! NB N # & = 1 N ( AE ( C ) The only complcaton ths ntroduces s that total contrbutons and total expendtures wll generally be dfferent from zero when measured n PPS unts even f the EU budget s balanced n nomnal terms. Hence, the budget defct calculated wth deflated fgures wll nclude a (postve or negatve) prce dfferences term that wll depend on the way n whch net balances are dstrbuted across countres wth hgh and low prce leveles. Usng equaton (3) we now have: (22) NB =! NB =! NB Q =! NB (Q " 1+ 1) =! NB (Q " 1) +! NB =! NB (Q " 1) + NB =! NB (Q " 1) " (NAE + D) # X " (NAE + D) where X s the sum of natonal balances weghted by the correspondng relatve prce levels or, for short, the prce correctons term. Notce that f Q = 1 for all then X = 0. Usng ths expresson, country s relatve real balance wll be gven by " NB (23) RNB = N # N! NB N & = N " # and ts value n curent euros wll be AE! C N! X! (NAE + D) N & = (AE! C ) + N (NAE + D! X) N (24) RNB P = (AE! C )P + N N (NAE + D! X)P = (AE! C ) + N N (NAE + D! X)P That s, workng wth real relatve balances amounts to allocatng the sum of overhead expendtures (non-allocated expendtures and the budget defct adjusted by the prce correctons term) across countres on an equal per capta bass n real terms. Hence, when we convert relatve balances to nomnal terms, countres wth hgher prce levels wll be allocated a greater share of overhead expendtures. Compensatory transfers and total contrbutons under the proposed scheme Under our proposal, the real compensatory transfer payable to country, T, wll be gven n a frst approxmaton by (25) T = N t = N [!"(y! y )! rnb ] = "(y! y )N! RNB Member states pay postve transfers f ther real ncome per capta exceeds the Unon average. Notce that compensatory transfers add up to zero n real terms because they are defned as the resduals of the redstrbuton regresson. When transformed nto nomnal terms, however, they may not add up to zero and wll n general have to be adjusted. 18

20 Let PT and NT be the total amounts n nomnal terms of postve and negatve transfers, PT =! P and NT =! " T P T :T >0 :T <0 To get thngs to balance, we wll reduce all postve transfers and ncrease all negatve ones (.e. ncrease net contrbutons) n the same proporton, a. We seek the value of the adjustment coeffcent a such that (1! a)pt = (1 + a)nt Solvng ths equaton we have (26) a = PT! NT NT + PT Hence, fnal compensatory payments n nomnal terms to country wll be gven by # T * = (1! a)t P = (1! a) ["(y! y )N! RNB ]P f T > 0 & (1+ a)t P = (1+ a) ["(y! y )N! RNB ]P f T < 0 or, usng equaton (24), ) # (1! a) "(y! y )N P! (AE! C )! N N (NAE + D! X)P & ( f T > 0 (27) T * + = * + # (1+ a) "(y! y )N P! (AE! C )! N N (NAE + D! X)P & ( f T < 0, + Now, country s total adjusted contrbutons wll be gven n nomnal terms by ) # C! (1! a)t P = ac! (1! a) "(y! y )N P! AE! N N (NAE + D! X)P & C * + ( = * + # C! (1+ a)t P =!ac! (1+ a) "(y! y )N P! AE! N N (NAE + D! X)P & (, + f T > 0 f T < 0 or ) # ac + (1! a)!"(y! y )N P + AE + N N (NAE + D! X)P & (28) C * + ( = * + #!ac + (1+ a)!"(y! y )N P + AE + N N (NAE + D! X)P & (, + f T > 0 f T < 0 Snce a wll generally be close to zero, ths expresson says that, to a frst approxmaton, each country s contrbuton to the Unon has three components: - full repayment for all EU expendtures allocated to t (AE ), - an equal per capta share n real terms of overhead expendtures (nomnal contrbutons to the fnancng of overhead expendtures ncrease proportonately wth the country s prce level), N " N (NAE + D! X)P # &, and 19

21 - a redstrbutve payment whch s postve f the country s real ncome per capta s above the Unon average and negatve otherwse,!(y " y ) References Begg, I. (2005). Fundng the European Unon. A Federal Trust Report on the European Unon s Budget. Federal Trust, London. Boldrn, M. and F. Canova (2001). Income dspartes and regonal polces. Economc Polcy 32, pp Commsson of the European Communtes (CEC, 1998). Fnancng the European Unon. Commsson report on the operaton of the own resources system. Brussels. Commsson of the European Communtes (CEC, 2004). Fnancng the European Unon. Commsson Report on the operaton of the own resources system. Brussels. Commsson of the European Communtes (CEC 2007). EU budget Fnancal report. Offce for offcal publcatons of the European Communtes. Luxembourg. Councl of the European Unon (CEU 2005). Fnancal perspectve Brussels. de la Fuente, A. and R. Doménech (2001). "The redstrbutve effects of the EU budget: an analyss and a proposal for reform." Journal of Common Market Studes 39(2), pp de la Fuente, A., V. Rant and R. Doménech (2008). The redstrbutve effects of the EU budget: an update and projectons for Mmeo. Department for Envronment, Food and Rural Affars (DEFRA, 2005). A vson for the Common Agrcultural Polcy. HM Treasury. London. Economc Polcy Commttee and European Commsson (ECP, 2006). The mpact of agng on publc expendture: projectons for the EU25 Member States on pensons, health care, long-term care, educaton and unemployment transfer. Annex. European Economy.Specal report no. 1/2006. Sapr, A., Ph. Aghon, G. Bertola, M. Hellwg, J. Psan-Ferry, D. Rosat, J. Vñals and H. Wallace (2003). An agenda for a growng Europe. Makng the EU economc system delver. Report of an ndependent hgh-level study group establshed on the ntatve of the Presdent of the European Commsson. Brussels. 20