PANORAMA OCTOBER 2016 RESEARCH PAPER COFACE ECONOMIC PUBLICATIONS. Economic crisis rhymes with rising political risk

Transcription

1 2 RESEARCH PAPER a ecoomies: Will political risk spoil the party i 217? 2 Political tempo acceleratig i 3 Ecoomic crisis rhymes with risig political risk 6 What impact does political ucertaity have o growth i? PANORAMA OCTOBER 216 a ecoomies: Will political risk spoil the party i 217? COFACE ECONOMIC PUBLICATIONS By Coface Group Ecoomists T The UK referedum i Jue, parliametary electios i Spai i December 215 ad Jue The past 12 moths have see a multiplicatio i the umber of political deadlies with uexpected outcomes ad ucertai cosequeces i. However, the political tempo is set to quicke agai i the comig year, with the referedum i ad the presidetial electio i Austria o 4 December, possible parliametary electios i Spai o 25 December (the third i just over a year!), electios i the Netherlads o 15 March, presidetial ad parliametary electios i i April, May ad Jue, without forgettig the vote i Germay i autum 217. The questio iduced by this faster political tempo is simple: will ecoomic growth withstad the pace? Ideed, growth could be affected i various ways by icreased political ucertaity: delayed ivestmet ad corporate spedig decisios, a drop i household cofidece, a dowtur o equities markets ad the icrease i bod rates takig a toll o fiacig terms for ecoomic agets, as well as a lack of reforms ad a freeze o public spedig i the evet of a govermet vacuum. While it is easy to eumerate these trasmissio chaels, it is far more difficult to quatify the impact of this political ucertaity o growth. This is what we attempt to do here. After measurig political risk i 14 coutries i wester by takig ito accout risk idicators specific to the regio (risig euroscepticism, ati-immigratio setimet ad a fragmetatio of political scees), we have built a ecoometric model aimed at measurig the impact of a icrease i political ucertaity o GDP growth i five coutries: Germay, the UK,, ad Spai. This idicates that the icrease i political risk oted sice the crisis has deted growth by.2 poits o average, the impact o ivestmet (.5 poits) beig higher tha that o household cosumptio (.1 poits). Nevertheless, this average masks clear differeces, growth i the UK (.3 poits), (.4) ad Spai (.3) beig more affected by political risk tha i ad Germay. I the evet that these last four coutries suffer a rise i political ucertaity like that see i the UK at the time of the referedum i Jue, their growth would be deted by aroud.5 poits o average. Evetually, it is difficult to speak of political ucertaity without metioig the US i these last moths of 216. Our model shows a sigificat impact o the US ecoomy by a shock i political ucertaity to a similar extet of that see with the UK referedum (1.5 poits). The impact o a ecoomies would be eve worse, thereby cofirmig the systemic role of the US ecoomy ALL OTHER PANORAMAS ARE AVAILABLE ON

2 2 PANORAMA RESEARCH PAPER October 216 EUROPEAN ECONOMIES: WILL POLITICAL RISK SPOIL THE PARTY IN 217? Julie MARCILLY Chief Ecoomist Hawa LY Juior Ecoomist Vicet TERNISIEN Juior Ecoomist 1 POLITICAL TEMPO ACCELERATING IN EUROPE Chart 1 Mothly arrivals of migrats i Greece ad 25, Political risk has bee associated with emergig markets for several decades, but ow it also affects wester coutries ad especially Wester (as well as the US i the late stages of 216). The first sigs of this risig political risk i appeared i 211, via istitutioal blockages i the a Uio ad umerous political ucertaities withi the coutries sufferig from the sovereig debt crisis (Greece ad to a lesser extet Portugal, Irelad ad Spai). These risks have agai otched up a gear over the past year with two parliametary electios i Spai ad the UK referedum i Jue. The political tempo is set to quicke agai over the ext 12 moths, whe votes with ucertai outcomes are scheduled i four of the mai eurozoe coutries (ad eve the top five, if we iclude the parliametary electio i the Netherlads o 15 March 217). Geece O 4 December, is due to hold a referedum o its costitutioal reform aimig to reduce the powers of the Seate ad thereby improve capacity to reform the coutry. After aoucig that he would resig if the project is rejected, prime miister Matteo Rezi cotributed to trasformig this vote ito a referedum for or agaist him. Issues related to migrats, for whom the umber of arrivals has ot falle sice the agreemet siged with Turkey cotrary to Greece (see chart 1), are playig a sigificat role i the campaig, as is the situatio i the bakig sector. Although Mr Rezi has sice suggested that he could remai i place eve after a defeat, votig itetios remai very mixed ad the outcome ucertai, with the o vote uderpied by the 5star ati-system movemet ad the Norther League far-right party. I the evet of defeat ad the departure of the curret prime miister, the short-term political future of the coutry would be ucertai, with possibly the formatio of a techical govermet of atioal uio or the orgaisatio of fresh electios. 2, 15, 1, 5, Jauary 214 April 214 July 214 October 214 Jauary 215 April 215 July 215 October 215 Jauary 216 April 216 July 216 Source: UNHCR

3 PANORAMA 3 I Spai, the risk of a third parliametary electio i just over a year, beig held o 25 December, has ot disappeared. Ideed, egotiatios betwee the four mai parties, which have started after the 26 Jue electio, have so far ot bee more efficiet tha those held after the December 215 vote. Parliamet has util 31 October to elect a govermet, uder threat of beig dissolved agai. The outcome of these egotiatios will otably deped o the positio of the PSOE leftwig party, udermied by dissesios regardig which path to follow: either to cotiue opposig the vote to ivest a coalitio govermet betwee the PP right-wig party of curret prime miister Mariao Rajoy ad the ati-corruptio cetreright party Ciudadaos, or to abstai. However, the forced resigatio of the PSOE s secretary geeral, Pedro Sachez o 1 October 216 could ope the way to a improvemet i egotiatios with the PP. Therefore, fresh electios could weigh o corporate ad household cofidece (eve if this has ot bee the case so far), especially sice the o-approval of the 217 budget before the ed of the year would be syoymous with freezig a large amout of public spedig at the curret level. Meawhile, the Frech presidetial electio o 23 April ad 7 May will see a showdow betwee the cadidates of the three mai parties: the rightwig Les Républicais, the left-wig Parti Socialiste ad the far-right Frot Natioal party. So far, oly the Frot Natioal cadidate is kow: Marie Le Pe is virtually certai to get to the secod roud of votig accordig to opiio polls. Les Républicais cadidate, probably former presidet Nicolas Sarkozy or former prime miister Alai Juppé, is to be chose after a first vote orgaized for 2 ad 27 November. The wier will be the favourite to accompay Marie Le Pe ad the wi the electio. Fially, curret presidet Fraçois Hollade is set to aouce before the ed of the year whether he will stad for electio ad take part i the first vote orgaized by the Parti Socialiste o 22 ad 29 Jauary. Two of his former miisters (Araud Motebourg ad Beoît Hamo) have already aouced their participatio i this first vote, whereas two others could also preset themselves for the presidetial electio: Emmauel Macro ad Cécile Duflot (Les Verts party). Oe shall ot forget likely cadidates, who are Jea-Luc Mélecho (far left) ad Fraçois Bayrou (cetre-right). I this cotext of multiple cadidates, sources of ucertaity are umerous: who will face Marie Le Pe i the secod roud? If abstetio is high i the secod roud, could she wi the electio? Will the party that wis the electio, which ever oe it is, maage to obtai a majority of seats i the Natioal Assembly followig the legislative electios i Jue? I Germay, the ext parliametary electios are to be held betwee 23 August ad 22 October 217. The usual parties (right CDU/CSU, cetreleft SPD, liberal FDP, the gree party ad the far left Like party) are set to face the ew far-right AfD party (Alterative für Deutschlad), whose score (curretly more tha 1% of votig itetios) will deped o the make-up of the traditioal-party coalitio. The most likely sceario remais a major coalitio betwee the CDU/CSU ad the SPD, withi which Agela Merkel would remai chacellor, eve though votig itetios i favour of her party have declied over the past year. 2 ECONOMIC CRISIS RHYMES WITH RISING POLITICAL RISK Coface assessmet of political risk i emergig markets... Followig the Arab sprig, which showed that uderlyig tesio i emergig coutries could express itself ad result i political chages ad revolutios, Coface adapted its methodology of measurig political risk (1) i order to take ito accout this ew ladscape. The series of populist risigs i North Africa coutries, the Middle East ad beyod appeared to be the result of socio-cultural mometum testifyig to the populatios aspiratio for greater justice ad icreased respect for idividuals. Therefore, i order to measure political risk i emergig markets (2), Coface decided to use idicators stemmig from two sources of ratioales. The model aims to idetify, i emergig markets, the aspiratio ad ability of populatios to rise agaist a system that iduces political ad ecoomic frustratios. As such, idicators of pressure for chage are take ito accout: GDP per capita: the higher the level of developmet, the more access to essetial goods (food, housig, health) is guarateed to a sizeable share of the populatio; similarly, the higher the populatio s expectatios are i terms of satisfactio of additioal eeds (3) ; (1) Political risk ca be defied as all political or admiistrative evets ad decisios, whether atioal or iteratioal, that ca cause ecoomic, commercial or fiacial losses for compaies that import, export or ivest abroad. (2) Coface Coutry Risk Paorama Chages i emergig coutry risk March 213 (3) Maslow s psychological theory of eeds. Abraham Maslow s hierarchy of idividual eeds follows the idea that these chage depedig o the level of satisfactio. Whe basic survival eeds are satisfied, other eeds must be satisfied too (security, belogig, esteem, self-actualisatio i the order defied by Maslow).

4 4 PANORAMA Uemploymet: a high uemploymet rate excludes people from professioal isertio ad social itegratio, thereby exacerbatig frustratio. Iflatio: a high iflatio rate ca make access to goods difficult, especially foodstuffs. The GINI coefficiet: a uequal distributio of icome (measured by the GINI idex) makes the satisfactio of certai social eeds difficult, such as access to housig. These traditioal ecoomic ad social idicators are key variables. However, they do ot fully exhaust the logic of this exasperatio. Frustratio is ot just ecoomic, but ca be political: Expressio ad resposibility : as uderscored by Armatya Se, freedom costitutes the aim of the developmet process. Freedom of speech, freedom of associatio ad the existece of free media is therefore a major issue. As such, their absece is a major source of discotet. Corruptio breeds a feelig of ijustice ad frustratio favourig protestatio. These variables reflect ecoomic, social, ad political frustratios i a coutry. I a secod stage, Coface measures the ability of coutries to trasform this pressure ito chage. Istrumets of chage are measured by the rate of higher educatio, the literacy rate i adults, access to iteret, the proportio of youg people, fertility rates, urbaisatio rate ad the female participatio rate. Political risk therefore combies these two modules (pressure for chage ad istrumets for chage), made up of six aggregates each. Table 1 Compoets i the Coface political risk idicator for emergig coutries PRESSURE INSTRUMENTS VARIABLE GDP per capita Gii Iflatio Uemploymet Political freedom Corruptio Youg people aged 2-29 Level of educatio Female participatio rate Fertility rate Iteret access Urbaisatio rate SOURCE IMF IMF Oxford Ecoomics UN... adapted to wester a coutries Uder the framework of developed coutries such as those that we study here (4), takig ito accout istrumets does ot seem relevat. As such, we estimate that if aspiratios for chage were preset, populatios would ot ecouter barriers either for expressig their exasperatio, or trasformig them ito political chage, ulike i emergig markets where the meas for provokig chage ca be blocked by several factors. For all EU coutries, the literacy rate is ideed above 95% (5). Furthermore, a majority of households i the EU have access to the iteret (81% i 214 accordig to Eurostat). Except for, with 2.1 childre per woma, fertility rates i are below 2 childre per woma. Fially, a Uio coutries are all highly urbaised. I terms of pressure to chage, we have take ito accout the uemploymet rate (6) ad corruptio, as a source of social discotet. We have also added GDP growth as well as the aual chage i the structural primary budget balace (% of GDP), which measures efforts i budgetary austerity that ca be upopular, especially sice icreased tax pressure is ofte actually cocetrated o the middle class, which is the largest class ad accouts for the lio s share of the electoral body. Other variables help idetify pressures existig i certai a coutries. These idices are measured by surveys, such as public opiio o immigratio or euroscepticism. Two surveys, udertake twice a year by Eurobarometer, deal with chages i euroscepticism ad fears regardig migratio. As such, we take ito accout the followig idicators: The percetage of idividuals withi the coutry that aswered immigratio to the questio What are the two mai sources of cocer i your coutry. The percetage of people expressig a egative opiio whe the questio What image of the EU do you have? is asked. Fially, the fragmetatio of political scees is a sig of political istability. We therefore take accout of the chage i the umber of seats i the parliametary majority (7) after each electio. A low majority ca be the sig of a large umber of parties represeted i the electios. The more parties there are i parliamet, the more difficult it is to gai a majority ad therefore implemet reforms. To measure political risk, i the a case, we therefore combie eight variables. GDP growth, the uemploymet rate, corruptio ad euroscepticism have the highest relative weights, 15%. (4) 14 coutries: Germay, Austria, Belgium, Demark, Spai, Filad,, Greece, Irelad,, the Netherlads, Portugal, Swede ad the UK. (5) 214 huma developmet report. (6) A alterative idicator is the uemploymet rate for youg people whose parets are fully aware. Note that takig this ito accout would ot have chaged the rakig preseted below. (7) Lower chamber i the case of a two-chamber parliamet.

5 PANORAMA 5 Table 2 Compoets i the Coface political risk idicator for wester a coutries VARIABLE GDP growth Gii Uemploymet Budget balace / GDP Corruptio Euroscepticism Ati-immigratio setimet Political fragmetatio Chart 2: Coface political risk idex for wester a coutries 7% 6% 5% 4% 3% 2% 1% % Chart 3 Coface political risk idex for wester a coutries Gap betwee 27 ad 216 3% 25% 2% 15% 1% 5% EU average Greece Germay SOURCE IMF, Coface estimates Oxford Ecoomics IMF Fiscal Moitor Eurobarometer Eurobarometer Natioal Filad Swede Austria Netherlad % % 1% 2% 3% 4% 5% 6% 7% Political score i 216 Spai Portugal Irelad Germay Demark Uited Kigdom Belgium Greece Expressed o a scale of to 1, this political risk idex rose by 13 poits o average i the coutries studied betwee 27 ad 216 (see chart 2). Usurprisigly, political risk has therefore icreased sigificatly i. More precisely, Greece ad, which already had the highest score i 27, are also the coutries for which the score has icreased the most. All of the idicators i these coutries deteriorated over the period. Note for example, corruptio, which was evertheless already at the highest level of the 14 coutries studied i 27, as well as icome iequality. I Greece, risk reached a peak of 67% i 212, whereas the majority lost 2% of its seats i parliamet. The lack of recovery, high uemploymet, budgetary strictess imposed by iteratioal istitutios, the migrat crisis ad high euroscepticism (51% of Greek people have a egative opiio of the EU i 216), are all factors explaiig why the score is still above 6%. I, pressure is similar; Italias stated that they were more worried about immigratio tha the Greeks i 215. Portugal (see chart 3), like Greece i 27, shows the highest iequalities i the sample: these have ot arrowed sice 27. The overall score remais fairly high give the still very high uemploymet rate, which is still icreasig. The electio i 215 of a left-wig coalitio, partly ati-eu, testified to risig euroscepticism i the coutry. Whereas the ecoomic recovery has bee cofirmed i Spai, iequalities, high uemploymet (22% i 215) ad parliametary istability (there is still o majority after two parliametary electios) are slowig the declie i pressures. Political risk has also icreased, i a lesser way, i ad i Germay accordig to this idicator, reflectig the deterioratio i the ecoomic ad social eviromet. The populatio s huge cocer over immigratio as well as a icreasigly egative opiio o the EU partly explais this tred. Usurprisigly, these two variables were also the origi of the icreasig level of risk see i Filad (+2 poits, the highest growth after Greece at 29 poits), the Netherlads (+19 poits) ad i Austria (+18 poits). Fially, the case of the UK highlights the limits of our model, it associates the coutry with a fairly low level of risk, which has icreased very slightly (see chart 3). Ideed, the model does ot capture very well the low-skilled part of the maual labour workforce i the case of coutries where uemploymet remais low (4.9% i July 216). I additio, whereas the coutry voted to leave the EU i Jue 216, oly 26% of the populatio has a egative opiio of the EU (vs. 32% i 215) accordig to the Eurobarometer survey. I cotrast, the survey shows a rise i migratory fears i 216 (8). (8) For 38% of the populatio i 216, this is a major source of cocer.

6 6 PANORAMA 3 WHAT IMPACT DOES POLITICAL UNCERTAINTY HAVE ON GROWTH IN EUROPE? Ecoomic growth ad political risk are itercoected. As set out i the previous part, deterioratio (improvemet) i ecoomic coditios is likely to cause a icrease (decrease) of political risk: a rise i uemploymet, iflatio or icome iequalities ca provoke social discotet. Meawhile, icreased political risk ca affect the ecoomic eviromet. The impact of this political ucertaity o growth primarily ivolves two trasmissio chaels that are likely to support each other mutually: 1) a declie o equity markets ad a icrease i bod rates takig a toll o fiacig terms for ecoomic agets (state, compaies ad households) ad hece their ivestmet ad spedig outlook, 2) a lower level of corporate ad household cofidece promptig delays or cacellatios i ivestmet or spedig decisios. If these last for some time, a third chael ca also occur: that of budget policies. I the evet of a log-lastig govermet vacuum, the freeze o public spedig causes a egative impact o activity. Various studies cofirm these itercoectios. I 211, Aise ad Vaiga (9) cocluded that a higher degree of political ucertaity (measured by the umber of chages of govermet) is associated with lower growth i GDP per capita, after studyig the statistical relatioship betwee these two variables i 169 coutries betwee 196 ad 24. These results cofirm those of Alesia (1996) (1) accordig to which GDP growth i 113 coutries betwee 195 ad 1982 was sigificatly lower whe the likelihood of the collapse of a govermet was high. However, recet evets i oly seem to partially cofirm the theory. For example, the Spaish ecoomy does ot seem to be pealized by the lack of govermet sice ed-215 if we look at growth figures for the first semester of 216. The same was also true of Belgium i I the UK, the ecoomic cosequeces of ucertaity caused by the 23 Jue vote are visible, but ot as bad as expected (see Coface Q3 216 Coutry Risk Barometer Paorama). To get a better picture, we set out to quatify the cosequeces of risig political risk. However, the idicator preseted i the previous sectio caot be used uder the framework of a ecoometric model sice its frequecy is aual. I order to measure political istability, we therefore used the EPU idex (Ecoomic Policy Ucertaity) developed by Baker, Bloom ad Davis, ad built o the basis of the umber of key-words, defiig ucertaity, foud i the press of the coutry studied. It relates three differet fields: the ecoomy, politics ad ucertaity that are themselves subdivided ito differet categories such as public spedig, fiacial regulatios ad iterior security (11). Box 1 Statistical model measurig the impact of political ucertaity o ecoomic activity. Our model is made up of two variables: the measure of political ucertaity (EPU, Ecoomic Policy Ucertaity, developed by Baker, Bloom ad Davis (12), associated respectively with GDP growth, growth i household cosumptio ad that i ivestmet measured by the gross fixed capital formatio (13). All the variables are seasoally adjusted, quarterly ad available betwee Q2 21 ad Q2 216 (61 observatios). We focus o five coutries i wester (, Germay,, Spai ad the UK) ad the US (see box 2 page 9). Macroecoomic variables are expressed i quarterly variatio with the ull hypothesis of ostatioarity rejected for all series except for Spaish GDP. However, this series becomes statioary if we take ito accout the period from Q1 28 to Q We therefore focus o this period for Spai. The EPU variable is a idex that has bee log-liearized ad we have (9) Aise A. ad Vaiga F. (211), "How Does Political Istability Affect Ecoomic Growth?", Workig Paper o. WP/11/12 of the Iteratioal Moetary Fud. (1) Alesia, A. ad Perotti, R. (1996), "Icome distributio, political istability, ad ivestmet", a Ecoomic Review 4, (11) The VIX idex measurig implied volatility o the S&P5 equity markets idex could also have bee used. Its results are similar to those of the EPU (see Ferrara ad Guéri, 216). (12) Measurig Ecoomic Policy Ucertaity, Scott R. Baker, Nicholas Bloom, ad Steve J. Davis, NBER Workig Paper No , October 215. (13) Source: Eurostat.

7 PANORAMA 7 derived the first differece of this log-liearizatio. I order to study the impact of ucertaity o GDP growth, we use autoregressive vectoral models (VAR(p)), with p beig the umber of lags chose. I order to specify the optimal umber of lags, we have based ourselves o stadard iformatio criteria (AIC, BIC, Schwarz) ad we select the optimal specificatio for each variable by miimizig these criteria. Depedig o the coutry cosidered, the umber of lags varies depedig o the robustess of the model s specificatio. Similarly, the presece of a costat depeds o its sigificace i the model. I order to avoid multi-colliearity problems, we estimate ivestmet, cosumptio ad GDP separately. We therefore aim to estimate the followig model for i = 1,, 6 coutry, with EPU - political ucertaity, GDP - growth i GDP, coso, which is that i household spedig ad ivest, which is that i ivestmet: After estimatig our equatios, we simulate a shock o political ucertaity ad observe its impact o GDP growth, household spedig ad the ivestmet. We use a Cholesky decompositio, which helps us i modelizig a orthogoal impulse respose fuctio represeted i cumulative form. Sice the 28 crisis, political ucertaity has icreased sigificatly (see chart 4). The EPU idex gaied 117 poits betwee ad ad the UK are the coutries where it rose the most (respectively 177 ad 19 poits). I ad Spai, the icrease oly reaches 63 ad 69 poits. Fially, Germay is i a itermediary positio (+98 poits). I this cotext, we have measured the impact of this icrease i political ucertaity sice 28 o ecoomic activity i by usig the autoregressive vector model (see box 1 page 6 ad 7). We have firstly measured the impact of a 1-poit icrease i the EPU idex o GDP, cosumer spedig ad ivestmet, that is to say a idetical shock for all coutries. Chart 4 EPU political ucertaity idex Uited Kigdom Germay Spai Source: Ecoomic Policy Ucertaity

8 8 PANORAMA GDP would decrease by almost.2% o average i, with all of the impact beig felt i the year followig the shock (see chart 5). I detail, ivestmet spedig (-.5%, see chart 7) would fall further tha household cosumptio (-.1%, see chart 6), i a sig that corporate cofidece is more sesitive to political ucertaity tha household cofidece. The esuig delays i compay ivestmet decisios therefore has a higher impact o growth tha that caused by a declie Chart 5 Cumulative impact o GDP of a shock o the EPU idex (+1 poits) Germay Spai Uited Kigdom i household cofidece (eve takig ito accout the relative weight of household cosumptio ad ivestmet i GDP (14) ). Spai is the coutry where growth is the most affected by this uiform political shock. It is followed by, the UK ad Germay. This rakig is modified whe we apply ot the same shock to the five coutries, but that effectively observed i the post-crisis period. therefore stads out as the coutry that has suffered the most from political risk (-.4 poits), followed by the UK ad Spai (-.3 poits each). closes the market, with a virtually zero impact. A possible explaatio for this is a addictive effect: havig witessed umerous govermets topplig ad early electios over several decades, Italia ecoomic agets are used to uexpected political chage. This seems to be eve more true for Italia compaies tha for households, with ivestmet spedig beig less pealized tha household cosumptio (cotrary to other a coutries) Chart 6 Cumulative impact o household cosumptio of a shock o the EPU idex (+1 poits) Germay Spai Uited Kigdom Chart 8 Cumulative impact o GDP of a shock o the EPU idex see betwee ad Germay Spai Uited Kigdom Chart 7 Cumulative impact o ivestmet of a shock o the EPU idex (+1 poits) Chart 9 Cumulative impact o GDP of a Brexit shock o the EPU idex (+34 poits) Germay Spai Uited Kigdom Germay Spai Uited Kigdom (14) Ivestmet spedig i totals 19.8% of GDP whereas cosumptio accouts for 53.4% of GDP. Takig ito accout these weightigs, ivestmet spedig evertheless remais the largest cotributor to the overall shock o growth, accoutig for 51% of the shock, vs. 21% for cosumer spedig.

9 PANORAMA 9 After havig measured the impact o ecoomic activity of the political shock observed sice the Lehma Brothers crisis, we questio ourselves about a evetual major shock i comig moths. More precisely, the aim is to quatify the impact o activity i, Germay, ad Spai of a shock of a similar extet to that see i the UK sice the begiig of 216 due to the referedum o Brexit (see chart 9 page 8). This extesive Brexit sceario would otably be relevat if sources of potetial risk to come betwee ow ad the ed of 217 (Italia referedum, political blockage i Spai, electios i ad Germay, see first part) were to actually occur. I order to modelize our Brexit type sceario, the amplitude of the ucertaity shock is set as beig the icrease i the EPU idex i the UK durig this short period, amely 34 poits betwee Q1 ad Q3 216 (15). What do the results show? Firstly, UK growth will lose.5 poits i oe year due to the ucertaity caused by the referedum (see chart 9). The, the Spaish ecoomy would usurprisigly be the most affected by a political shock of a similar extet as the UK referedum, sice the model is liear. Its growth would be deted by aroud 1.2 poits after oe year. The effect would be aroud.4-.5 poits for Germay i the evet of a rise i risk ext autum. Italia growth would oly lose.2 poits accordig to the model, for example, if the outcome of the referedum o 4 December were to geerate reewed ucertaity. Fially, Frech growth would lose aroud.7 poits, which would imply a hike of aroud 1.2% i atioal compay isolvecies (16). Box 2 The possible electio of Doald Trump, a risk for a ecoomies? Usig the same methodology, we try to assess the impact of a icrease i political ucertaity i the US, similar to that observed i the UK, sice the begiig of the year o US growth, but also o a ecoomies. For istace, this is likely to occur i the evet of a victory by Doald Trump over Hillary Clito o 8 November. Accordigly to this sceario, US growth would lose 1.5 poits i oe year. The impact of this shock would be worse for a ecoomies (-2 poits of growth after a year),, where it would be slightly over oe poit ad Germay, beig two exceptios. Apparetly more surprisig, these results cofirm the systemic role of the US ecoomy already observed after the Lehma crisis. Although the crisis origiated i the US, GDP growth i the coutry oly dropped by 2.8% i 29, far lower tha the 4.5% declie see i the eurozoe (17). Chart 1 Respose fuctio of GDP growth followig a ucertaity shock caused by the electio of D. Trump Uited Kigdom Germay Spai Uited States (15) Oly data for July ad August were available for Q3 whe this study was completed. (16) Based o the Coface forecast model for Frech isolvecies (where DEF is default, ad Margi, the margi rate for o-fiacial compaies measured by the EBITDA/VA ratio): DEF t = c(1)*def t-1 + c(2)*gdp t + c(3)*margi t-2 + c(4) (17) Citigroup estimates that if Trump becomes the US presidet, global GDP would arrow by.7-.8 percetage poits via the fiacial markets that would be drive by ucertaity.

10 RESERVATION This documet is a summary reflectig the opiios ad views of participats as iterpreted ad oted by Coface o the date it was writte ad based o available iformatio. It may be modified at ay time. The iformatio, aalyses ad opiios cotaied i the documet have bee compiled o the basis of our uderstadig ad iterpretatio of the discussios. However Coface does ot, uder ay circumstaces, guaratee the accuracy, completeess or reality of the data cotaied i it. The iformatio, aalyses ad opiios are provided for iformatio purposes ad are oly a supplemet to iformatio the reader may fid elsewhere. Coface has o results-based obligatio, but a obligatio of meas ad assumes o resposibility for ay losses icurred by the reader arisig from use of the iformatio, aalyses ad opiios cotaied i the documet. This documet ad the aalyses ad opiios expressed i it are the sole property of Coface. The reader is permitted to view or reproduce them for iteral use oly, subject to clearly statig Coface's ame ad ot alterig or modifyig the data. Ay use, extractio, reproductio for public or commercial use is prohibited without Coface's prior agreemet.please refer to the legal otice o Coface's site. Photo : Foltolia - Layout : Les éditios stratégiques COFACE SA 1, place Costes et Bellote 9227 Bois-Colombes