Alternative measures of macroeconomic imbalances in the EU – design and verification

36 2019, XXII, 1 Economics DOI: 10.15240/ ul/001/2019-1-003 In oduc ion Yea s 2008 and 2009 we e pa icula ly a ec ed by he ou b eak o he global economic ecession which in addi ion o economic ins abili y was also a ec ed by poli ical ins abili y. Economies o mos coun ies in he wo ld el he impac o he i nancial c isis, no excluding he EU coun ies. Since he s a o he c isis, he e has been a subs an ial educ ion in he EU’s g ow h po en ial. In he EU, his was e l ec ed a he mac o-le el by 4.74% GDP educ ion in 2009 (Campos-So ia, Inchaus i-Sin es & Eugenio- Ma in, 2014). The high le els o ex e nal liabili ies and p i a e and public deb in many coun ies in he EU s ill cons i u e subs an ial ulne abili ies o g ow h, jobs and i nancial s abili y. The de elopmen o a ious indica o s a mic o and mac o le els in imes o las c isis in a ious coun ies was in es iga ed by se e al au ho s (Campos-So ia, Inchaus i-Sin es, & Eugenio-Ma in, 2015; Gugle , Weichselbaume , & Zulehne , 2015; Zhao, Jiang, & Li, 2014; Ta ulescu & Pa u i, 2014; Mazu ek & Mielco á, 2017; Klepáč & Hampel, 2018). In ecen yea s, pa ial examina ion o indica o s o mac oeconomic imbalances in he EU o OECD coun ies has ocused, o example, on he (im) balance o he cu en accoun balance (Gosse & Se ani a, 2014; Angelini & Fa ina, 2012; Zoz i, Chudík, & Dieppe, 2012), employmen (Que imi & Se gi, 2017; Ma ko i z, Boe , & Van Dick, 2014), ex e nal imbalances (Fogli & Pe i, 2015; Mau o & Pappada, 2014), labo p oduc i i y (Auzina-Emsina, 2014), he ade imbalance (Begle & Ni sch, 2014), se ing-up wages and p ices (Angelini, Dieppe, & Pie luigi, 2015), o he sa is ac ion o ci izens in ela ion o mac oeconomic indica o s (S acca, 2014). The Eu opean Commission moni o s he de elopmen o indica o s o mac oeconomic imbalances wi hin he “ale mechanism epo ” (AMR) in which indica o s o mac o-economic imbalances (MI) a e assessed ac oss EU Membe S a es (i.e. “Mac oeconomic Imbalance P ocedu e” MIP). MIP was es ablished in Decembe 2011 and implemen ed o he i s ime in 2012 as a consequence o economic and i nancial c isis. I s aim is de ec ion, p e en ion and co ec ion o mac oeconomic imbalances including i nancial s abili y o EU coun ies. I is based on wo egula ions No. 1176/2011 and No. 1174/2011 wi hin he so-called “Six-pack” aimed a imp o ing he economic go e nance o he EU. MIP should iden i y mac oeconomic ends which should be ins umen al in aking app op ia e policy esponses o mi iga e and manage mac oeconomic imbalances (AMR, 2014). Eu opean Union economic policy amewo ks ely mo e han e e on imely and high quali y socio-economic and i nancial s a is ics in membe s a es howe e i s quali y a ies. Mac oeconomic imbalances a e hus e l ec ed a na ional le el in a ious o ms and a ec mainly he capaci y o each economy o gene a e s ong and sus ainable g ow h and c ea e jobs. MIP could be conside ed o ale mechanism based on he e olu ion ends o moni o ed headline and auxilia y indica o s. Based on he de elopmen o indica o s moni o ed in he MIP he AMR gi es he Eu opean Commission ecommenda ions o Membe S a es o he adop ion o e ec i e solu ions which should be accompanied by a comp ehensi e and coo dina ed policy p ocedu es and decisions (AMR, 2015, p. 2). The lack o imely and decisi e mac oeconomic policy ac ion – o co ec domes ic and ex e nal imbalances – could led o he i nancial c isis and he G ea Recession (Ca e, Co a, Pagano, & Visco, 2011) so hey a e e y likely o ha e a - eaching consequences on economic and social li e. Selec ion o indica o s and hei moni o ing, e alua ion and eedback a e impo an o ALTERNATIVE MEASURES OF MACROECONOMIC IMBALANCES IN THE EU – DESIGN AND VERIFICATION S anisla Kološ a, Pa ol K áľ, Filip Flaška EM_1_2019.indd 36EM_1_2019.indd 36 8.3.2019 9:15:528.3.2019 9:15:52 37 1, XXII, 2019 Economics he economic, i nancial and social s abili y o he EU and indi idual EU membe s a es (see o example s udies o Babecký e al. (2012), Cso os and Szalaj (2014), Domonkos e al. (2017) and Li (2018)). These indica o s o mac oeconomic imbalances a e also subjec o c i icism (Cooling on, 2012; 2013); on he o he hand in he EU he e exis s poli ical consensus o i s e alua ion and p ac ical e idence. The e o e, i is meaning ul o make con ibu ion wi hin MI assessmen . The e alua ion o indica o s o he mac oeconomic imbalances should gi e an o e all pic u e abou de elopmen o a leas headline mac oeconomic indica o s as well as abou he de elopmen o indi idual indica o s. Each indica o in AMR epo s is epo ed sepa a ely o each EU coun y and ice e sa. F om AMR da a, we can ge pic u e abou he e olu ion o di e en MI indica o s in he EU, bu comple e pic u e abou he o e all si ua ion o indi idual coun ies in he i eld o mac oeconomic imbalances is missing. The e o e, we ocused on design and e i i ca ion o sui able al e na i e e alua ion ool which AMR lacks, and which could be usable o decision making p ocesses wi hin Eu opean Commission. The e o e we decided o build on he Knedlik and Schwaini z (2011) who p oposed o combine as many meaning ul single indica o s as possible in one composi e indica o as a ool o moni o ing o economic and i nancial s abili y. The main aim o his a icle is o p opose agg ega ed and pa ial indices o mac oeconomic imbalances (MI) and o p o ide an al e na i e complex e alua ion ool o each EU coun y on i s global posi ion in a leas he headline o sco eboa d indica o s. F om he p oposed indices we expec as and simple look on e olu ion o MI in EU coun ies, o make hem mo e comp ehensible o wide use, and o p opose simple eedback ool measu ing p og ess o EU membe s a es owa ds imp o ing he mac oeconomic balance including s abili y in i nancial sec o s. Such comp ehensi e e alua ion and moni o ing can be help ul in e alua ing he e i ciency and e ec i eness o he measu es aken o elimina e he nega i e de elopmen s in mac oeconomic balance ha jeopa dize he p ope unc ioning o economic and mone a y union, as s a ed in AMR (2015, p. 2). The es o he pape is a anged as ollows. In sec ion wo, we desc ibe da a used o de i e mac oeconomic imbalance indices. In sec ion h ee, we p esen me hodology we used o de elop de i e he indices. In sec ion ou , we discuss he esul s. Finally, sec ion i e d aws some conclusions. 1. Da a Desc ip ion The AMR uses a sco eboa d o ou een headline indica o s including new employmen indica o s plus a wide se o auxilia y indica o s ( e iew o all indica o s con ains s a is ical annexes o AMR). New employmen indica o s we e no included in ou esea ch due o ou ocus on gene al ex e nal and in e nal imbalance. These new indica o s we e p e iously auxilia y and ha e been added in 2016 o he headline MIP sco eboa d o complemen he in o ma ion p o ided by he unemploymen indica o al eady included wi hin in e nal imbalance headline indica o s (AMR, 2016). Indica o o in e nal imbalance “% change in de l a ed house p ices” was excluded due o da a incomple eness in many coun ies. In ou case, indica o s 1-5 ep esen ex e nal imbalances (In1-5) while indica o s 6-10 ep esen in e nal imbalances (In6-10). These indica o s ocus on he mos ele an dimensions o mac oeconomic imbalances and compe i i eness losses. Wi h e e ence o AMR we logically assume he desi able di ec ion o chosen headline indica o s as p esen ed in Tab. 1 (+/- means ha g owing/dec easing alues o indica o ha e posi i e impac on agg ega ed MI index, i.e. alues o agg ega ed MI index inc eases). Fo design and e i i ca ion o p oposed me hodology an annual da a o a iables we e used om Chap e 3 o he s a is ical annex o AMR 2015 (we use one yea da a which we e no a e aged). We wo ked wi h o i cial da a om Eu os a and o design and e i i ca ion o ou indices 2004-2013 ime pe iods we e used. Indices we e compu ed om alues o headline MI indica o s o yea s 2004-2012; yea 2008 means cu poin o p e-c isis and c isis pe iod; alues o MI indica o s om 2013 we e used o e i i ca ion o p oposed mean MI indices om i s in e p e a i e capabili y poin o iew. We coun all es ima ed alues o eal da a. Mal a and Poland we e elimina ed om ou analysis o he yea 2004 and Luxembou g o he pe iod 2004-2007 due o lack o da a in a ious headline indica o s. The in e connec ions among EU economies lead us o conse a i e in e p e a ions o ou esul s. EM_1_2019.indd 37EM_1_2019.indd 37 8.3.2019 9:15:528.3.2019 9:15:52 38 2019, XXII, 1 Economics A e ou b eak o global c isis indica o s P i a e sec o deb , Gene al go e nance sec o deb , Unemploymen , Ne in e na ional in es men s in a e age exceeded i s h esholds and hese indica o s can be also included in o g oup o indica o s wi h he highes in l uence on agg ega ed MI index a e 2008 (Appendix – Tab. 11). In yea s 2004-2012 Real e ec i e exchange a e, Nominal uni labo cos s and Expo ma ke sha es (wi h he excep ion in 2010) we e wi hin h esholds. A e 2008 a ouse disc epancies among EU coun ies in indica o s o Real e ec i e exchange a e, Expo ma ke sha e, P i a e c edi l ows and To al i nancial sec o liabili ies. Wes e n EU coun ies did no show sha p l uc ua ion o abo e men ioned indica o s di ec ly in 2008 bu wi h 2 yea ime-lag. I we compa e c isis yea 2012 wi h p e-c isis yea 2004 we can see de e io a ion in mo e han 50% o EU coun ies in headline indica o s like Unemploymen (20% inc ease), P i a e sec o deb (26.75% inc ease), Gene al go e nance sec o deb (40% inc ease). Global c isis a ec ed he mos Expo ma ke sha e (dec ease om -0.1% o -4.9%) and Ne in e na ional in es men s (in mo e han 50% o EU coun ies his indica o was wo sen a 200%). A e 2008 mo e han hal o EU coun ies exceeded h esholds in Ne in e na ional in es men s, P i a e sec o deb and Gene al go e nance sec o deb . 2. Me hodology The p incipal componen analysis (PCA) can be seen in some sense as a s anda d me hod o compu ing o agg ega ed indices as i was used o c ea ing many indices o di e en pu pose (e.g. Bolcá o á and Kološ a (2015) agg ega ed SD index in EU; D ehe e al. (2006) KOF index o globaliza ion; Flo ida e al. (2011) global c ea i i y index; Sasaina and Sa elli (2012) CPI o T anspa ency In e na ional; Kološ a, Sabelo á and K áľ (2018) agg ega ed and pa ial LR indices, ec .). This me hod enables o educe he dimensionali y o a da a se which includes a la ge numbe o in e - ela ed a iables while e aining as much as possible he a ia ion p esen in he da a se (Jolli e, 2002, p. 10). In o de o o e come de i ciencies wi hin compu a ion and in e p e a ion o indices we decided o use a cons ained PCA implemen ed in he R package nsp comp (Sigg & Buhmann, 2008). Acco ding Sigg and Buhmann (2008), in he case o cons ained PCA we s a wi h he s anda d p oblem o ge he i s p incipal componen , i.e. (1) whe e C ϵ R D×D is he posi i e semi-de i ni e co a iance ma ix o he da a, bu we assume wo addi ional cons ain s imposed on w, namely spa si y IIwII0 ≤ K1 and non-nega i i y w ≥ 0. P emise Headline indica o Uni o measu e Th eshold + Cu en accoun balance – In1 % o GDP -4% /+6% o GDP (eu o/non-eu o a ea) + Ne in e na ional in es men s – In2 % o GDP -35% o GDP - Real e ec i e exchange a e – In3 % change +-5% / +-11% (eu o/non-eu o a ea) + Expo ma ke sha es – In4 % change -6% - Nominal uni labo cos s - In5 % change +9% /+12% (eu o/non-eu o a ea) - P i a e c edi sec o l ow – In6 % o GDP 14% o GDP - P i a e sec o deb – In7 % o GDP 133% o GDP - Gene al go e nance sec o deb – In8 % o GDP 60% GDP - Unemploymen a e – In9 % 10% - To al i nancial sec o liabili ies – In10 % change +16.5% Sou ce: own, using AMR Tab. 1: O e iew o used headline MI indica o s in EU EM_1_2019.indd 38EM_1_2019.indd 38 8.3.2019 9:15:528.3.2019 9:15:52 39 1, XXII, 2019 Economics Sigg and Buhmann (2008) p oposed he expec a ion-maximiza ion algo i hm (Tab. 2) o i nd a solu ion o he abo e-men ioned p oblem. In he o iginal PCA, he i s p incipal componen maximizes a iance o he p ojec ed da a, he second p incipal componen maximizes a iance o he p ojec ed da a oo bu unde he cons ain ha i is o hogonal o he i s componen , analogously he hi d p incipal componen maximizes a iance and i is o hogonal o he i s wo componen s e c. In he case o cons ained PCA, he addi ional cons ain s allow us o educe signi i can ly he numbe o componen s, and hence o simpli y in e p e a ion, ading i o a small loss in explained a iance. We assume ha he esul ing index should be a linea combina ion o he o iginal indica o s (2) p ese ing as much a iabili y o o iginal da a as possible and wi h coe i cien s c1, c2, ..., c10 ollowing p emises acco ding o signs lis ed in Tab. 1. In o he wo ds, we a e looking o he i s p incipal componen wi h loadings o indica o s sa is ying some cons ain s imposed by hei expec ed in l uence on he index. In o de o ge c1, c2, ..., c10, da a we e s anda dized, i.e. a co ela ion ma ix was used, because a iances o he selec ed indica o s a y in o de s o magni ude (Appendix – Tab. 9). Then cons ain s we e en o ced on loadings o he o iginal indica o s and he i s p incipal componen , i.e. index o mac oeconomic imbalance (agg ega ed o o e all MI index) was ex ac ed, sepa a ely o each yea . The ac , ha he p oposed MI index possesses p ope ies we equi e om a good MI measu e, can be e i i ed in wo ways. Fi s , by i s quali a i e analysis whe e we ho oughly in e p e he alues o he index o see how close hese in e p e a ions a e o hose we can ge om indi idual headline indica o s. Second, we can use a mul iple ime se ies model o see whe he o ecas ing abili y o he index co esponds o ou expec a ion o p e-c isis, c isis and pos -c isis pe iods. In he pape , o simplici y, we u ilize a simple andom o es mul iple eg ession model (B eiman, 2001) o c ea e a one-s ep- ahead p edic ions whe e a alue o he index o a coun y a ime is p edic ed using alues o he index o a coun y a imes – 1, – 2 and – 3. We also use a coun y g oup a iable (1 – EU 15, 2 – V4 coun ies, 3 – o he EU coun ies) as a p edic o . Ou assump ion is ha he p edic ion abili y o olde alues o he index is qui e good in p e-c isis pe iod, i de e io a es du ing a c isis and hen i imp o es once he c isis is e ea ing. The basic building block o andom eg ession o es s is a eg ession ee. In he Sou ce: Sigg and Buhmann (2008) Tab. 2: Algo i hm: EM o cons ained PCA EM_1_2019.indd 39EM_1_2019.indd 39 8.3.2019 9:15:528.3.2019 9:15:52 40 2019, XXII, 1 Economics pape , we we e wo king wi h andom o es eg ession whe e indi idual eg ession ees a e cons uc ed ia CART algo i hm, i.e., we assume bina y ees whe e each pa en node is spli in o exac ly wo child nodes using esidual sum o squa es (RSS) as a spli ing c i e ion. RSS is de i ned as ollows (B eiman, 2001): (3) whe e ¯yl (¯y ) is he mean alue o a esponse a iable Y in he le ( igh ) child node, espec i ely. I we assume ha a ee is cons uc ed e ically om op o bo om, he nodes a he bo om a e called e minal nodes. The p edic ed alue o each e minal node is hen he a e age alue o a esponse a iable o all obse a ions in he node. When cons uc ing a andom o es , we i mul iple eg ession ees o boo s ap samples o he o iginal da a sample using he ollowing algo i hm in Tab. 3 (B eiman, 2001). Analogously, indices desc ibing ex e nal and in e nal imbalances we e de i ed using indica o s In1-In5 (ex e nal MI index) and In6-In10 (in e nal MI index), espec i ely. We applied he ob ained MI indices o anking he EU coun ies. Un o una ely, compu a ion o p oposed MI indices is qui e imp ac ical as o each yea we ha e o un cons ained PCA which s ongly depends on a ailabili y and quali y o da a. We can elimina e his p oblem i we eplace indices based on coe i cien s ex ac ed yea ly by indices based on a e age coe i cien s in a chosen pe iod. We decided o a e age coe i cien s om he pe iod 2004-2012 consis ing om bo h p e-c isis and c isis yea s. Finally, we compa ed p ope ies o hese so called mean based MI indices o hose based on yea ly compu ed MI indices. All compu a ions we e done in s a is ical so wa e R (R Co e Team, 2016). The da a and co esponding R code can be eques ed om au ho s. Weigh s o headline indica o s enable us o moni o changes o i s signi i cance impac on agg ega ed MI index o e ime. Weigh s o indices ha e e ol ed in dependence on in ensi y o i s yea o yea changes (highe in ensi y o changes means highe weigh and ice e sa). Indica o s can be di ided acco ding o i s signi i cance (Tab. 4) in o ou g oups: 1. S able – Cu en accoun balance, Ne in e na ional in es men s; 2. G owing in l uence in c isis – Gene al go e nance sec o deb , Unemploymen ; 3. Dec easing in l uence in c isis – Real e ec i e exchange a e, pa ially also Nominal uni labo cos s, P i a e c edi sec o l ow, To al i nancial sec o liabili ies; 4. Negligible impac – Expo ma ke sha es, P i a e sec o deb . A e ou b eak o global c isis, indica o s Unemploymen and Gene al go e nance sec o deb had he highes weigh s among in e nal imbalance indica o s, i.e. ill 2008 he EU coun ies a oided majo l uc ua ions o hese indica o s (Tab. 4). Each yea agg ega ed MI index consis s o a ious composi ion o componen s (especially a e 2008), which had di e en yea ly based weigh s. The e o e, we c ea ed an index composed o he a e age weigh s o each indica o which ca ch bo h pe iods o expansion as well as economic ecession. In his way we C ea e N boo s ap samples om he da a by selec ing n cases a andom wi h eplacemen . A each o he boo s ap samples i a eg ession ee as ollows: A each node selec m a iables a andom ou o all M possible a iables (independen ly o each node). Find he bes spli on he selec ed m a iables. A e age i ed ees o ge p edic ions. Sou ce: B eiman, 2001 Tab. 3: Algo i hm: Random o es eg ession EM_1_2019.indd 40EM_1_2019.indd 40 8.3.2019 9:15:538.3.2019 9:15:53 41 1, XXII, 2019 Economics ob ained an agg ega ed mean MI index wi h weigh s o all ep esen ed headline indica o s ha can easily calcula e he alue o he o e all MI index subs i u ing alues o each coun y. Indica o s ha ha e a high weigh in he yea ly based agg ega ed MI index emained s ong a e a e aging, and opposi e applied o indica o s ha ha e li le o no weigh . Co ela ions among agg ega ed MI index cons uc ed on an annual basis (Me hod 1) wi h a mean MI index using he a e age weigh s (Me hod 2) a e in he Tab. 5. We ound e y s ong di ec co ela ion be ween he agg ega ed MI indices and ex e nal MI indices (cons uc ed using bo h me hods) o all EU coun ies and howe e in c isis pe iod co ela ions educed (mo e isible in non-EU coun ies) i s ill emained e y s ong. We deduce ha in economically s able pe iod bo h e sions o cons uc ion o he MI index can 2004 2005 2006 2007 2008 2009 2010 2011 2012 AVG STDEV In1 0.522 0.500 0.517 0.493 0.562 0.572 0.432 0.552 0.442 0.508 0.049 In2 0.443 0.340 0.352 0.385 0.562 0.203 0.052 0.552 0.573 0.381 0.175 In3 -0.258 -0.336 -0.272 -0.338 -0.282 0 0 0 0 -0.164 0.159 In4 0000000.517 0 0.162 0.075 0.174 In5 -0.491 -0.500 -0.449 -0.422 -0.259 -0.225 -0.329 0 0 -0.297 0.194 In6 -0.256 -0.196 -0.392 -0.360 -0.377 -0.548 -0.374 0 0 -0.278 0.185 In7 000000-0.180 0 0 -0.021 0.060 In8 00000-0.295 -0.465 -0.323 -0.387 -0.168 0.199 In9 -0.240 -0.098 -0.016 -0.024 -0.281 0 0 -0.535 -0.548 -0.189 0.223 In10 -0.318 -0.472 -0.423 -0.432 0 -0.441 -0.218 0 0 -0.259 0.207 Sou ce: own EU 2004 2005 2006 2007 2008 2009 2010 2011 2012 In1-In10 0.982 0.934 0.962 0.976 0.939 0.947 0.697 0.681 0.549 In1-In5 0.989 0.992 0.990 0.989 0.989 0.918 0.986 0.888 0.787 In6-In10 0.646 0.691 0.332 0.526 0.737 0.643 0.727 0.494 0.542 EU 15 In1-In10 0.960 0.947 0.943 0.912 0.871 0.964 0.857 0.854 0.704 In1-In5 0.982 0.996 0.989 0.982 0.982 0.957 1.000 0.936 0.857 In6-In10 0.635 0.881 0.938 0.899 0.804 0.611 0.768 0.646 0.254 non EU 15 In1-In10 0.964 0.918 0.962 0.967 0.846 0.890 0.835 0.456 0.412 In1-In5 0.993 0.967 0.989 0.989 0.978 0.758 0.967 0.495 0.703 In6-In10 0.720 0.225 -0.148 0.357 0.698 0.659 0.681 0.555 0.835 Sou ce: own Tab. 4: Weigh s o indi idual indica o s in he MI index Tab. 5: Co ela ion be ween MI indices and mean based MI indices EM_1_2019.indd 41EM_1_2019.indd 41 8.3.2019 9:15:538.3.2019 9:15:53 42 2019, XXII, 1 Economics be used; a he ime o economic ecession o mo e app op ia e al e na i e compu a ion o agg ega ed and ex e nal MI indices we conside i s me hod. Lowe and l uc ua ing alues o co ela ion coe i cien s we e be ween In e nal MI indices; in his pa ial index we expec o ake o mo e con enien i s me hod conside ing he ola ili y o he balance o componen s weigh s. These me hodological i ndings a e impo an o ou u he in es iga ions. The a e age agg ega ed MI index may be used o es ima ed assessmen o MI by subs i u ion o es ima ed o expec ed da a using a e age weigh s. Ano he possibili y how o p edic e olu ion o MI could by calcula ion o weigh s om es ima ed da a using ou me hodology and, on ha basis, de e mine he p ognosis o de elopmen o MI in EU coun ies. The same me hodology was chosen o he cons uc ion o pa ial indices o ex e nal and in e nal MI. This allowed us o in es iga e no only o e all MI among EU coun ies bu also ex e nal MI and in e nal MI. The weigh s o componen s o ex e nal MI index (Tab. 6) de eloped in e y iden ical way like weigh s o agg ega ed MI index; Cu en accoun balance and Ne in e na ional in es men s we e he mos s able again and hese indica o s ha e a s ong impac on ex e nal MI index. A e ou b eak o economic and i nancial c isis he es o indica o s o ex e nal MI educed i s impac on ex e nal MI index wi h he excep ion o Expo ma ke sha e. Di e ences we e be ween weigh s o indica o s o agg ega ed MI index and weigh s o indica o s o in e nal MI index (Tab. 7) – a e ou b eak o c isis he highes di e ences we e in indica o s o Unemploymen , To al i nancial sec o liabili ies (in bo h weigh s inc eased) and Gene al go e nance sec o deb (weigh dec eased); none o he in e nal MI indica o s was s eadily e ol ing. Those ac o s had an impac on he o e all de elopmen o he agg ega ed MI index in indi idual EU coun ies as well as on pa ial MI indices cons uc ed by bo h me hods. 3. Resul s and Discussion We s a ou p esen a ion o esul s by compa ison o he one-head p edic ion abili y o he index based on andom o es eg ession models wi h ou assump ion s a ed in he 2004 2005 2006 2007 2008 2009 2010 2011 2012 AVG STDEV In1 0.608 0.578 0.629 0.547 0.606 0.717 0.717 0.709 0.672 0.642 0.064 In2 0.512 0.487 0.477 0.500 0.569 0.697 0.581 0.706 0.690 0.580 0.095 In3 -0.282 -0.394 -0.356 -0.459 -0.376 0.000 -0.216 0.000 0.000 -0.231 0.186 In4 0.000 0.000 0.000 0.000 0.000 0.000 0.227 0.000 0.269 0.055 0.110 In5 -0.538 -0.523 -0.500 -0.491 -0.410 0.000 -0.224 0.000 0.000 -0.298 0.242 Sou ce: own 2004 2005 2006 2007 2008 2009 2010 2011 2012 AVG STDEV In6 0.000 0.710 0.727 0.742 0.000 0.000 0.000 0.724 0.000 0.323 0.383 In7 0.718 0.704 0.190 0.424 0.672 0.000 0.000 0.689 0.076 0.386 0.320 In8 0.000 0.000 -0.659 -0.520 -0.079 -0.305 -0.407 0.000 0.000 -0.219 0.259 In9 0.000 0.000 0.000 0.000 0.000 0.647 0.614 0.000 0.718 0.220 0.331 In10 -0.696 0.000 0.000 0.000 -0.736 -0.699 -0.676 0.000 -0.691 -0.389 0.369 Sou ce: own Tab. 6: Weigh s o indi idual indica o s in he ex e nal MI index (In1-5) Tab. 7: Weigh s o indi idual indica o s in he in e nal MI index (In6-10) EM_1_2019.indd 42EM_1_2019.indd 42 8.3.2019 9:15:538.3.2019 9:15:53 43 1, XXII, 2019 Economics me hodology pa o he pape . We cons uc ed six andom o es models (Tab. 8). Based on Tab. 8 we can conclude ha p edic ion abili y o ou index is cohe en wi h ou assump ion, i.e. i is sa is ac o y be o e he c isis, hen de e io a es a hen slowly eco e s a e he peak o he c isis. In he p e-c isis pe iod, mos EU coun ies ha e posi i e alues o agg ega ed MI index calcula ed by bo h me hods and bo h indices eac ed o he c isis by educing o alues o all coun ies (Fig. 1). This also applied o pa ial index o he ex e nal MI. Non-EU15 coun ies ha e expe ienced a adical educ ion o he alues immedia ely a e he ou b eak o he c isis, while he EU-15 wi h one yea lapse. In he a e ma h o he i s e ec s o he c isis mo e han 50% non-EU coun ies had highe alues o agg ega ed MI index han 75% o he EU15 coun ies. In 2011 and 2012 we can see a mo e balanced alue o agg ega ed MI index, pa icula ly mean based, bu up o ha ime, mos EU coun ies ha e no eached he p e-c isis le els. In he p e-c isis pe iod we e he lowes alues o agg ega ed MI indices in La ia, Es onia, Bulga ia, Romania; a e 2008 had he lowes alues Po ugal, G eece, Spain, Cyp us; he coun ies wi h he highes alues du ing all in es iga ed pe iod o ime we e Denma k, Luxembou g, Ne he land, Sweden and Ge many. Pa adoxically, he bes alues o he pa ial in e nal MI index we e in coun ies wi h low le els o o e all and ex e nal MI indices excluding Po ugal and Cyp us whe e alues l uc ua ed. A e he ou b eak o he c isis, he g ea e decline was in all indices in non-EU15 coun ies which esponded o he c isis mo e sensi i e and signalized simila end in EU15 coun ies whe e he same scena io ook place wi h one yea lag. Tab. 9 p o ides colo ed classi i ca ion o coun ies acco ding o eached alues o index wi h compa ison o indices cons uc ed by he i s and he second me hod. I we compa e coun ies like Belgium, Denma k, Ne he land, Luxembou g, Aus ia, Ge many, Sweden, Finland and Mal a o coun ies such as Cyp us, G eece, Spain, Po ugal, and a e 2008 also I aly we will i nd ha he i s g oup o coun ies shows high alues o he agg ega ed MI index in compa ison o he o he g oup. Also a cha ac e is ic ea u e o he i s g oup o coun ies is imp o emen in alues o agg ega ed MI index a e he i s impac o he c isis in 2011 and 2012 – i seems ha in hese coun ies we e applied measu es which p ope ly sol ed issues o c isis in conc e e socio-economic en i onmen s, in e nal s uc u e o ma ke s and posi ion o hese coun ies a global scale can be conside ed o mo e s able in compa ison wi h second g oup o coun ies mos ly om sou h o he EU. F ench and I aly had high MI index alues in he p e- c isis pe iod bu a e c isis bo h coun ies we e no able o each p e-c ises le el. Bal ic S a es and Viseg ad g oup showed in gene al mos ly nega i e agg ega ed MI index alues and i s de elopmen a ied – s ong weigh o long- e m deep nega i e Ne in e na ional in es men posi ion and Cu en accoun balance could be one o he ac o s which keep low o nega i e alues o MI indices om long e m pe spec i e. UK has eached in 2011-2012 simila alues like Bal ic and Viseg ad coun ies, while du ing c isis s ong weigh o changes in Real e ec i e exchange a e could posi i ely impac s on o e all alues o bo h MI indices. Yea 2008 showed he onse o he c isis and wo sening o MI index in all EU coun ies excep Ne he land, Es onia and UK which on he con a y be e P edic ed alue P edic o s Va iance explained Model 1 y2007 y2004, y2005, y2006, g oup 88% Model 2 y2008 y2005, y2006, y2007, g oup 77% Model 3 y2009 y2006, y2007, y2008, g oup NA Model 4 y2010 y2007, y2008, y2009, g oup 56% Model 5 y2011 y2008, y2009, y2010, g oup 52% Model 6 y2012 y2009, y2010 y2011, g oup 64% Sou ce: own Tab. 8: Random o es eg ession models EM_1_2019.indd 43EM_1_2019.indd 43 8.3.2019 9:15:538.3.2019 9:15:53 44 2019, XXII, 1 Economics Fig. 1: Boxplo s MI indices and mean MI indices – agg ega ed (In1-10), ex e nal (In1-5) and in e nal (In6-10) MI Sou ce: own EM_1_2019.indd 44EM_1_2019.indd 44 8.3.2019 9:15:538.3.2019 9:15:53 51 1, XXII, 2019 Economics R Co e Team. (2016). R: A language and en i onmen o s a is ical compu ing. Vienna: R Founda ion o S a is ical Compu ing. Re ie ed om h ps://www.R-p ojec .o g/. Sasiana, M., & Sa elli, A. (2012). Co up ion Pe cep ions Index 2012 S a is ical Assessmen . Luxembou g: Publica ions O i ce o he Eu opean Union. Sigg, C. D., & Buhmann, J. M. (2008). Expec a ion- Maximiza ion o Spa se and Non-Nega i e PCA. In P oceedings o he 25 h In e na ional Con e ence on Machine Lea ning (pp. 960-967). h ps://doi.o g/10.1145/1390156.1390277. S a is ical annex AMR. (2015). S a is ical annex o Ale Mechanism Repo 2015. Eu opean Commision. A ailable om h p://ec.eu opa.eu/eu os a /web/mac oeconomic- imbalances-p ocedu e/publica ions. S acca, L. (2014). Financial imbalances and household wel a e: Empi ical e idence om he EU. Jou nal o Financial S abili y, 11, 82-91. h ps://doi.o g/10.1016/j.j s.2013.12.001. Ta ulescu, A., & Pa u i, A. (2014). S uc u al Funds And Economic C ises: Romania’s Abso p ion Pa adox. P ocedia Economics and Finance, 16, 64-72. h ps://doi.o g/10.1016/ S2212-5671(14)00775-8. Zhao, X., Jiang, X., & Li, Z. (2014). The impac o he economic c isis on he i nancial pe o mance o mul ina ional co po a ions. In e na ional Re iew o Economics and Finance, 37, 55-68. h ps://doi.o g/10.1016/j. i e .2014.11.013. Zoz i, M., Chudik, A., & Dieppe, A. (2012). Thousands o models, one s o y: Cu en accoun imbalances in he global economy. Jou nal o In e na ional Money and Finance, 31, 1319-1338. h ps://doi.o g/10.1016/j. jimon i n.2012.02.003. Ing. S anisla Kološ a, PhD. Ma ej Bel Uni e si y Facul y o Economics Depa men o Public Economics and Regional De elopmen Slo akia s anisla [email p o ec ed] RND . Pa ol K áľ, PhD. Ma ej Bel Uni e si y Facul y o Economics Depa men o Quan i a i e Me hods and In o ma ion Sys ems Slo akia [email p o ec ed] Ing. Filip Flaška, PhD. Ma ej Bel Uni e si y Facul y o Economics Depa men o Public Economics and Regional De elopmen Slo akia i lip. l [email p o ec ed] EM_1_2019.indd 51EM_1_2019.indd 51 8.3.2019 9:16:028.3.2019 9:16:02 52 2019, XXII, 1 Economics Appendix y In1 In2 In3 In4 In5 In6 In7 In8 In9 In10 y In1 In2 In3 In4 In5 In6 In7 In8 In9 In10 04 M -3 -27 2 1 2 11 108 48 9 18 09 M-1 -39 0 -1 4 1 162 54 9 3 SD6302 5 2 65125314 SD 6535 6 4 7 76294 8 VC NA NA 122 445 95 57 47 52 38 78 VC NA NA NA NA 95 744 47 54 39 308 Q1 -8 -48 1 -3 0 7 68 29 6 9 Q1 -4 -84 -1 -4 3 -2 116 35 7 -1 Q2 -4 -25 1 0 2 9 107 44 8 14 Q2 -1 -43 1 0 4 1 146 53 8 3 Q3 2 -8 2 5 3 13 159 63 11 20 Q3 3 -5 2 2 6 4 215 68 10 8 05 M -3 -28 0 -2 3 15 116 48 9 22 10 M-1 -37 -3 -7 -1 1 158 60 10 4 SD 6 35 5 6 5 10 54 26 3 11 SD 5543 5 3 9 70314 7 VC NA NA 1125 NA 144 67 46 54 38 51 VC NA NA NA NA NA 621 44 52 43 163 Q1 -8 -52 -2 -6 1 9 76 27 7 15 Q1 -4 -83 -4 -11 -2 0 116 38 7 -1 Q2 -4 -25 -1 -5 2 14 109 43 8 18 Q2 -1 -46 -3 -8 -1 3 140 57 9 3 Q3 1 -7 1 1 4 17 164 65 10 26 Q3 3 -3 -2 -4 0 5 213 80 12 8 06 M -5 -31 1 0 3 17 124 46 8 18 11 M-1 -34 0 0 0 3 157 64 10 4 SD 8 39 2 5 4 11 54 27 3 13 SD 462 1 6 2 6 71354 7 V NA NA 400 4800 141 65 44 58 34 73 VC NA NA 700 NA 550 191 45 55 43 186 Q1 -10 -62 -1 -3 1 10 89 27 6 11 Q1 -4 -76 -1 -4 0 -1 114 41 7 0 Q2 -4 -29 0 0 2 14 116 43 7 14 Q2 -1 -50 0 -2 1 3 143 59 8 4 Q3 1 -3 1 3 3 19 175 63 9 22 Q3 1 1 1 0 2 5 208 82 13 5 07 M -6 -37 3 4 5 18 133 44 7 19 12 M1 -35 -2 -4 2 0 156 68 11 2 SD9393 5 6 95427210 SD 4622 4 3 5 73355 5 VC NA NA 115 150 129 50 41 60 32 51 VC 780 NA NA NA 144 1500 47 51 48 212 Q1 -11 -73 1 1 2 12 101 26 5 11 Q1 -2 -79 -3 -6 1 -2 112 43 7 -1 Q2 -5 -28 1 3 3 18 120 38 6 17 Q2 -1 -51 -2 -5 3 1 136 62 10 3 Q3 0 -6 3 6 5 23 184 61 8 26 Q3 3 5 -2 -3 4 3 199 87 14 5 08 M -5 -35 3 0 7 14 154 46 6 6 SD 8 47 5 6 5 10 74 28 2 11 VC NA NA 182 2950 78 73 48 60 30 176 Q1 -10 -76 1 -4 3 9 112 26 5 2 Q2 -6 -38 2 -1 5 14 140 42 6 7 Q3 0 -7 4 3 7 18 206 63 8 9 Sou ce: own No e: M – Mean, SD – S anda d de ia ion, VC – Va ia ion coe i cien Tab. 11: Desc ip i e s a is ics o used annual headline indica o s o MI in EU EM_1_2019.indd 52EM_1_2019.indd 52 8.3.2019 9:16:028.3.2019 9:16:02 53 1, XXII, 2019 Economics Abs ac ALTERNATIVE MEASURES OF MACROECONOMIC IMBALANCES IN THE EU – DESIGN AND VERIFICATION S anisla Kološ a, Pa ol K áľ, Filip Flaška As a consequence o i nancial and economic c isis in he EU he mac oeconomic imbalances (MI) ha e been moni o ed since 2012. Annual Ala m Mechanism Repo (AMR) ocused on assessmen o MI con ains he in e linkages be ween he eal economy and he i nancial sec o . F om AMR da a, we can ge pic u e abou he e olu ion o di e en MI indica o s in he EU, bu comple e pic u e abou he o e all si ua ion o indi idual coun ies in he i eld o MI is missing. The e o e, we ocused on design and e i i ca ion o sui able al e na i e e alua ion ool which AMR lacks, and which could be usable o decision making p ocesses wi hin Eu opean Commission. The main aim o he a icle is o p opose agg ega ed and pa ial indices o MI using cons ained PCA which can: i) p o ide a complex e alua ion o each EU coun y on i s global posi ion in headline indica o s; ii) make MI indica o s mo e comp ehensible o wide use; iii) design simple al e na i e assessmen ool use ul o moni o ing whe he measu es aken by he EU and he membe s a es a e di ec ed owa ds imp o ing he mac oeconomic balance. The p oposed MI indices a e e i i ed om a quan i a i e as well as quali a i e poin o iew. Resul s o p oposed assessmen ools showed ha : Luxembou g, Ge many, Ne he land and Sweden can be conside ed o he mos s able EU coun ies om MI poin o iew; Pos -Communis coun ies deal wi h economic c isis be e han some coun ies om sou h o he EU; a e i s c isis yea s he bes imp o emen in MI index eached Bal ic and Viseg ad coun ies oge he wi h Luxembou g and Ne he land; o he EU coun ies in 2012 did no each MI index alues om p e-c isis pe iod. This s udy indica es ha posi i e and s able alues o cu en accoun balance and ne in e na ional in es men posi ion as % o GDP ha e high weigh on mac oeconomic s abili y o EU coun ies. Key Wo ds: P incipal componen analysis, mac oeconomic imbalances, agg ega ed index, indica o s. JEL Classi i ca ion: C8, E6. DOI: 10.15240/ ul/001/2019-1-003 EM_1_2019.indd 53EM_1_2019.indd 53 8.3.2019 9:16:028.3.2019 9:16:02