Debt sustainability in Guatemala: Institutional arrangement and quantitative analysis

Ca alán-He e a, Juan Ca los Wo king Pape Deb sus ainabili y in Gua emala: Ins i u ional a angemen and quan i a i e analysis IDB Wo king Pape Se ies, No. IDB-WP-1140 P o ided in Coope a ion wi h: In e -Ame ican De elopmen Bank (IDB), Washing on, DC Sugges ed Ci a ion: Ca alán-He e a, Juan Ca los (2020) : Deb sus ainabili y in Gua emala: Ins i u ional a angemen and quan i a i e analysis, IDB Wo king Pape Se ies, No. IDB-WP-1140, In e -Ame ican De elopmen Bank (IDB), Washing on, DC, h ps://doi.o g/10.18235/0002648 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/234710 S anda d-Nu zungsbedingungen: Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen Zwecken und zum P i a geb auch gespeiche und kopie we den. Sie dü en die Dokumen e nich ü ö en liche ode komme zielle Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich machen, e eiben ode ande wei ig nu zen. So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen (insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en, gel en abweichend on diesen Nu zungsbedingungen die in de do genann en Lizenz gewäh en Nu zungs ech e. Te ms o use: Documen s in EconS o may be sa ed and copied o you pe sonal and schola ly pu poses. You a e no o copy documen s o public o comme cial pu poses, o exhibi he documen s publicly, o make hem publicly a ailable on he in e ne , o o dis ibu e o o he wise use he documen s in public. I he documen s ha e been made a ailable unde an Open Con en Licence (especially C ea i e Commons Licences), you may exe cise u he usage igh s as speci ied in he indica ed licence. h ps://c ea i ecommons.o g/licenses/by-nc-nd/3.0/igo/legalcode Abs ac 1 This pape assesses deb sus ainabili y in Gua emala. Deb s abili y has been achie ed a e y low expendi u e le els, a he expense o adequa e p o isioning o public goods and se ices and a widening gap in social de elopmen and in as uc u e. Since iscal ou comes a e no independen om iscal policy a angemen s and p ocedu es, he pape also se s o h a hypo hesis o possible ins i u ional a angemen s ha ha e allowed o he con ainmen o iscal de ici s o o e 20 yea s. The pape a gues ha embedded in he legal amewo k and ins i u ional a angemen , he e is an “implici ” iscal ule ha a o s s abili y. The pape explo es cha ac e is ics o how iscal policy is conduc ed, showing ha go e nmen expendi u es a e p o-cyclical, p o iding oom o imp o emen in cycle managemen . Fiscal policy has been mainly conce ned wi h s abili y a he han o he possible goals like boos ing long- un g ow h, a enua ing business cycles, imp o ing human de elopmen indica o s, dealing wi h edis ibu ion issues and o he goals ha iscal policy could pu sue. JEL classi ica ions: H62, H63, E62, E02 Keywo ds: Fiscal sus ainabili y, Gua emala, Deb , Fiscal policy 1 This s udy was unde aken in he Economic Resea ch Depa men o Banco de Gua emala wi hin he amewo k o he Cen e o La in Ame ican Mone a y S udies (CEMLA) Join Resea ch P og am 2019 in collabo a ion wi h he Financial S abili y and De elopmen (FSD) Ne wo k o he In e -Ame ican De elopmen Bank (IDB). The au ho s g a e ully acknowledge commen s p o ided by Alejand o Izquie do, P incipal Economis o he Resea ch Depa men o he IDB, and commen s ecei ed du ing he in e nal pee e iew p ocess. The au ho addi ionally app ecia es aluable commen s by Juan Ca los Cas añeda Fuen es and Héc o Wil edo Ma ínez Méndez. The opinions exp essed in his publica ion a e hose o he au ho and do no e lec he iews o CEMLA, he FSD Ne wo k, he IDB, o he Cen al Bank o Gua emala. Email: jcch@bangua .gob.g ; phone: +(502) 2429-6000x3653; ax: +(502) 2429-6086. Economic Resea ch Depa men , Banco de Gua emala. 7a A e. 22-01, zona 1. Ciudad de Gua emala, C.A. Email: jcch@bangua .gob.g ; phone: +(502) 2429-6000x3653; ax: +(502) 2429-6086. 2 1. In oduc ion Al hough iscal adjus men seems inc easingly necessa y in se e al coun ies in La in Ame ica, in o de o p e en a si ua ion in which deb sus ainabili y is comp omised, he e a e also coun ies like Gua emala whe e s ic expendi u e con ols ha e led o mode a e budge de ici s and low deb a ios, e lec ing a long-s anding commi men o p uden iscal and mone a y managemen . Clea ly, iscal sus ainabili y is desi able, and by any means his should be jeopa dized, bu in Gua emala deb sus ainabili y has been achie ed a e y low expendi u e le els, a he expense o adequa e p o isioning o public goods and se ices and a widening gap in social de elopmen and in as uc u e. In a coun y wi h hese social de elopmen lags, i is necessa y o unde s and and e alua e how iscal policy is ope a ing and how hese iscal esul s a e a ained. In o de o documen deb dynamics and sus ainabili y, I implemen ou di e en me hodologies: i) he s anda d o long- un app oach, ii) he sho - un app oach, iii) he Fan Cha app oach and i ) he p obabilis ic model de eloped by Mendoza and O iedo (2004). The gene al conclusion ac oss me hodologies is ha public deb lies wi hin easonably s able le els. Fiscal de ici s ha e been con ained o o e 20 yea s, which ansla es in o he obse ed low le el o deb . Behind his esul , he e a e cul u ally-shaped ins i u ions ha go agains he p oblem o excessi e public indeb edness; in Gua emala he e is a s ong bias owa ds cu ing spending as opposed o deb inc eases whene e he e is a decline in go e nmen e enues. This is an idiosync a ic cha ac e is ic o he coun y. Fu he mo e, I a gue ha legal and ins i u ional a angemen s cons i u e an “implici ” iscal ule a o ing s abili y o public inances. Fiscal policy in Gua emala has been mainly conce ned wi h mac oeconomic s abili y a he han o he possible goals like boos ing long- un g ow h, a enua ing business cycles, imp o ing human de elopmen indica o s, dealing wi h edis ibu ion issues and o he goals ha iscal policy could pu sue. Despi e Gua emala’s success in e ms o iscal sus ainabili y, go e nmen expendi u e has been p o-cyclical. This endency his migh be exace ba ed by p e-commi men s o e enue o speci ic spending lines, since hey ie down expendi u es o e enues, making expendi u es almos as p o-cyclical as e enues. On he one hand, hese a angemen s help o s abilize deb , bu on he o he , hey a e de imen al in e ms o adequa e cycle managemen . I use dynamic co ela ions o show he posi i e associa ion be ween ou pu and leads and lags o ax income and expendi u es. I a gue ha adequa e cycle managemen equi es expendi u es o be coun e -cyclical. Thus he e is s ill oom o imp o emen in e ms o cycle managemen , pa icula ly when i is appa en ha 3 capi al spending is highly p o-cyclical; his is p ecisely he ype o expendi u e ha could be used o implemen sound coun e cyclical policy. The es o he pape is o ganized as ollows. Some s ylized ac s and desc ip i e s a is ics a e gi en in Sec ion 2, along wi h a discussion o how iscal esul s a e in luenced by ins i u ional a angemen and cul u e. This sec ion addi ionally documen s p o-cyclicali y o expendi u es. Sec ion 3 p esen s he s anda d deb sus ainabili y analysis, and Sec ion 4 concludes wi h some inal ema ks. 2. S ylized Fac s 2.1 O e iew Fo mo e han 20 yea s, deb has emained low and s able, luc ua ing be ween 20 and 25 pe cen o GDP. In 2018, o ins ance, cen al go e nmen deb amoun ed o 24.5 pe cen o GDP, one o he lowes le els in La in Ame ica (see Figu e 1). The la ges sou ce o inancing o cen al go e nmen is he domes ic ma ke , which accoun s o abou 57 pe cen o he o al deb s ock. Ex e nal inancing comes mos ly om mul ila e al loans (28 pe cen ) and ex e nal bonds (16 pe cen ). Due o weak e enue collec ion, he s ock o deb as a sha e o o al go e nmen e enue has been ising in ecen yea s, eaching 231 pe cen in 2018 (see Figu e 2). Figu e 1. Go e nmen Deb Ra ios 4 Figu e 2. Composi ion and Deb Bu den Since cen al go e nmen deb is a s ock a iable e lec ing accumula ed de ici s, his obse ed s abili y e lec s a ack eco d o low and s able iscal de ici s obse ed o e nea ly wo decades (see Figu e 3). Despi e ad e se shocks, cen al go e nmen inances ha e emained s able. The de ici widened ma kedly in 2009-2010 in he mids o he global inancial c isis bu na owed o e he ollowing yea s as economic condi ions imp o ed, eaching a low o 1.1 pe cen in 2016 be o e ising aised again o 1.7 pe cen in 2018 as he esul o a mo e expansiona y iscal policy in he con ex o a weake economy. Figu e 3. O e all and P ima y Balances, % o GDP In spi e o being able o main ain low iscal de ici s, cen al go e nmen e enue in ake emains a key challenge. Gua emala has consis en ly anked in he lowes posi ions in e ms o go e nmen e enue as pe cen age o GDP, 11.4 pe cen on a e age om 1999 o 2018 (10.6 5 pe cen in 2018). Tax e enue is ai ly s able, as 94 pe cen o o al e enue comes om ax collec ion, bu i has been insu icien o sa is y social de elopmen needs. Low go e nmen e enue has cons ained expendi u es, also anking among he lowes in La in Ame ica a jus 12.1 pe cen o GDP in 2016 (12.3 pe cen in 2018) and limi ed he esou ces a ailable o social de elopmen , as expendi u es on heal h and educa ion also ank a he bo om o La in Ame ican coun ies (see Figu e 4). Al hough Gua emala was ecen ly upg aded and classi ied as an uppe middle-income coun y (Wo ld De elopmen Indica o s Da a Bank), a es o inequali y and po e y a e highe han La in Ame ican a e ages. The po e y headcoun a io a na ional po e y lines was 0.6 pe cen o popula ion, which was highe han he a e age o La in Ame ica (0.3 pe cen ). Mo eo e , he Human De elopmen Index was 0.65 in 2017 (Human De elopmen Da a 2017, UNDP), lowe han La in-Ame ican a e age (0.73). Figu e 4. Low Re enue- o-GDP Ra io and Low Expendi u e Le els While he ac o ha ing one o he lowes e enue- o-GDP a ios has limi ed he esou ces a ailable o sa is y social de elopmen needs, i has no jeopa dized he s abili y o public inances (as shown in Sec ion 3), because i seems ha in Gua emala i is easie o cu expendi u es han o inc ease deb . This is unusual, since mos democ acies ha e a endency o pu sue sub-op imal iscal policies, leading o he accumula ion o excessi e deb . I a gue ha , in he pa icula case o 12 in expo s and wo ke s’ emi ances), i seems ha he go e nmen employed a coun e -cyclical measu e, inc easing expendi u es (as a sha e o GDP) as ou pu and ax e enue dec eased. A e 2010, expendi u es began o adjus , con e ging o he le el o e enues. To e u n o Figu e 8, in bo h panels i is e iden ha he adjus men o expendi u es a e he c isis came almos en i ely h ough a educ ion o “ lexible” expendi u e, despi e he ac ha he s imulus o 2009 was enginee ed mainly h ough “ igid” expendi u e. I we obse e he yea -o e -yea change in expendi u e- o-GDP a ios (see Figu e 9), hen i is ob ious ha he s imulus o 2009 was mainly p o ided by an inc ease in “ igid” expendi u e. I is also e iden ha he subsequen adjus men included a la ge componen o capi al expendi u e educ ion. Figu e 9. Expendi u e- o-GDP, YoY Change (%) Displaying se ies o expendi u es and e enues ha coin eg a e in he long un is ce ainly a basic cha ac e is ic o op imal (o a leas app op ia e) deb managemen . In Gua emala his seems o be he case. Ne e heless, in e ms o cycle managemen , he e is s ill conside able oom o imp o emen . While expendi u es a e p o-cyclical, adequa e cycle managemen equi es hem o be coun e -cyclical. In o de o implemen a success ul coun e -cyclical expendi u e policy, while p ese ing he good esul s obse ed so a in e ms o deb sus ainabili y and con ainmen o de ici s, Capi al Expendi u e should be used as he main iscal policy ool. The e a e wo main easons o his policy ecommenda ion. Fi s , since Capi al Expendi u e is lexible, he s imulus can be wi hd awn wi hou majo poli ical opposi ion, and he e o e his can be unde aken in a 13 imely manne wi hou a ec ing iscal inances in he medium e m. Secondly, as men ioned abo e, public in es men may ha e complemen a i ies wi h p i a e in es men , posi i ely a ec ing posi i ely he g ow h a e o he economy, and in u n, imp o ing iscal sus ainabili y. 3. Deb Sus ainabili y Analysis The e is a la ge li e a u e on indica o s o public deb sus ainabili y and empi ical es s o iscal sol ency. Mos o he empi ical li e a u e a emp s o cons uc simple indica o s ha can be used o assess deb sus ainabili y (e.g., Blancha d, 1990; Blancha d e al., 1991; Bui e , 1985) and o de elop o mal econome ic es s ha can de e mine whe he he hypo hesis ha he In e empo al Go e nmen Budge Cons ain (IGBC) holds can be ejec ed (o no ) by he da a; see, o example, Bohn (2007), Mendoza and Os y (2008) and Ghosh e al. (2013). Since he deb - o-GDP a io in Gua emala has been ma kedly s able du ing he las 20 yea s, as no ed abo e, he e is no ob ious eason o o e complica e deb sus ainabili y analysis by applying s a e-o - he a echniques. I he e o e con ine he exe cise o a classic deb sus ainabili y analysis using ou di e en me hodologies: i) he s anda d o long- un app oach, ii) he sho - un app oach, iii) he Fan-Cha app oach and i ) he p obabilis ic model de eloped by Mendoza and O iedo (2004) o wha hey call he “Na u al Deb Limi .” All ou me hodologies a e implemen ed using he Templa e o Deb Sus ainabili y Analysis de eloped by he In e -Ame ican De elopmen Bank (see Bo ensz ein e al., 2010). 3.1 The Long-Run App oach The Long- un o S anda d App oach ollows he con ibu ions o Blancha d (1990) and Bui e (1985). I is based on he analysis o he p ima y su plus (PS) equi ed o main ain he a io o deb - o-GDP cons an in he long- un. The analysis ies o ind he equi ed PS o s abilize deb - o-GDP a a gi en le el—usually (and in his case) he cu en deb le el, gi en he condi ions consis en wi h he long- un o he s eady s a e. The analysis s a s om he cu en pe iod budge cons ain o he non- inancial cen al go e nmen , which equa es he lows o go e nmen e enues and expendi u es wi h changes in he s ock o public deb : G + (1 + R )D −1 = T + D (1) 14 whe e G a e expendi u es, R is he in e es a e paid o e deb , D is he s ock o (one pe iod, non- con ingen ) public deb and T is go e nmen e enues. A e conside ing ha a sha e (α) o deb is issued in US dolla s and o which an in e es a e ( ) is paid, we can ew i e equa ion (1) in eal e ms and as a ac ion o GD, o ind a di e ence equa ion desc ibing he way deb e ol es o e ime: (2) whe e lowe -case le e s ep esen he same a iables as in equa ion (1) bu in eal e ms and exp essed as sha es o GDP. Le γ be eal ou pu ’s g ow h a e, e he US-bila e al eal exchange a e and ps = (τ −g ) he p ima y su plus. Imposing he s eady s a e condi ion o e he p e ious equa ion and sol ing o ps, we ob ain: (3) This condi ion s a es ha he p ima y su plus is enough o co e he ne , eal in e es cos o se icing deb (ne om he eal g ow h a e o GDP). In o he wo ds, he s eam o long- un p ima y su pluses—app op ia ely discoun ed—has o ully co e d, he speci ied le el o deb , in his case, he cu en deb le el (Bo ensz ein e al., 2010). Using his equa ion, we ind he le el o p ima y su plus equi ed o s abilize he deb - o- GDP a io a ound i s cu en alue o a gi en in e es a e and g ow h a e o he economy (see Table 2). A p ima y su plus o 0.05 pe cen o GDP is equi ed o main ain he deb - o-GDP a io a i s cu en le el (24.5 pe cen ). Table 2 con ains all he assump ions made o calcula e he equi ed ps. The las obse ed (o cu en ) le el o p ima y su plus (−0.32 pe cen , nega i e meaning a p ima y de ici ) implies a equi ed adjus men om cu en ps o 0.37 pe cen o GDP o each he 0.05 pe cen o GDP needed o s abilize deb a i s cu en le el. Since his calcula ion is based on he assump ions, desc ibed in he le panel o Table 2, abou he in e es a e, he long- un eal g ow h a e and o he a iables, i is ob ious ha he equi ed p ima y su plus will a y wi h changes in hese assump ions. The e o e, on he igh panel o he same able, some simula ions a e pe o med. The i s one calcula es wha would be he equi ed p ima y su plus i GDP g ows 15 one s anda d de ia ion below he assumed po en ial g ow h, ce e is pa ibus. In his case, in o de o s abilize deb - o-GDP a io in i s cu en le el, a p ima y su plus o 0.34 pe cen would be equi ed. The second simula ion assumes an a e age eal in e es a e 2 s anda d de ia ions highe o wha was p e iously assumed, inc easing he equi ed p ima y su plus om 0.05 pe cen o 1.07 pe cen o GDP. Addi ionally, i we assume an ini ial deb le el o 40 pe cen o GDP (ins ead o he cu en 24.5 pe cen ) he equi ed ps in his case would be o 0.08 pe cen . Finally, in a wo s - case scena io, in which all nega i e shocks we e o happen simul aneously, he p ima y su plus equi ed o s abilize deb is 2.24 pe cen o GDP.8 Table 2. Long-Run App oach The assump ions made in his app oach a e hose ha one would expec o be ul illed in a s eady s a e condi ion. The e o e, hey cons i u e a use ul benchma k, bu hey a e a he s ong assump ions. Ne e heless, a p ima y su plus o 0.05 pe cen o GDP equi ed o main ain he cu en le el o deb - o-GDP a io, i is easonably a ainable, i we obse e ha he economy has un su pluses up o 1.7 pe cen o GDP in he pas ; achie ing a su plus o 0.05 pe cen would no equi e majo adjus men s o iscal e o ms. Addi ionally, his equi ed p ima y su plus is close o 8 In o de p o ide and idea o how sensi i e a e hese esul s o changes o he unde lying assump ions, Table 7 o Appendix A.1 con ains a sensi i i y analysis in which a g id o possible p ima y su pluses is p esen ed, esponding o changes in he assumed eal ou pu g ow h a e, he a e age eal in e es a es and he eal dep ecia ion. Long- un App oach Simula ions S eady-S a e Le el o Deb 2.3% Deb - o-GDP Ra io 24.46% A e age Real In e es Ra e 3.71% Requi ed P ima y Su plus (%GDP) 0.34% Real Dep ecia ion -1.92% Requi ed Adjus men (%GDP) 0.65% In la ion a e (GDP de lac o ) 4.00% Long- un g ow h a e 3.50% Alpha 57.14% 8.03% Es ima ed p ima y su plus o NFPS -0.32% Requi ed P ima y Su plus (%GDP) 1.07% Requi ed Adjus men (%GDP) 1.39% Requi ed P ima y Su plus (% GDP) 0.05% 40% Requi ed P ima y Su plus (%GDP) 0.08% Requi ed Adjus men (%GDP) 0.37% Requi ed Adjus men (%GDP) 0.40% A e age P ima y Su plus (1990-2018) -0.35% Requi ed P ima y Su plus (%GDP) 2.24% Requi ed Adjus men (%GDP) 2.56% A e age Real In e es Ra e (+2 s.d.) GDP g ow h (-1 s.d.) Deb /Y All Simula ions 16 a balanced budge , which implies ha , gi en ou assump ions, he cu en le el o deb is no a om he alue consis en wi h he no ion o equilib ium implied by he s eady s a e. 3.2 Sho -Run App oach This app oach is also based on equa ion (2), bu in con as o he long- un app oach, he ocus he e is on he sho - un deb dynamics wi h a cen al (baseline) scena io and disc e e s ess es s; o e o he mos likely pa h o deb de e minan s. To elabo a e he baseline, we use se e al sou ces o de e mine he mos likely pa h o he de e minan s o he deb - o-GDP a io in he sho o medium e ms. Fo he Real GDP g ow h a e, we use p ojec ions o Banco de Gua emala o he i s wo yea s and o subsequen yea s we used IMF’s p ojec ions epo ed in he Wo ld Economic Ou look (WEO), eleased in Ap il 2019. P ojec ions o he p ima y su plus also come om he WEO, Ap il elease. Bo h in e es a es, domes ic and o eign, a e se o he las obse ed alue o he i s i e yea s o he p ojec ion. The nominal exchange a e dep ecia ion and he in la ion a e a e aken om o ecas s o a mac o model de eloped wi hin he cen al bank. Table 3. Sho -Te m o Endogenous App oach: Baseline Baseline: no shocks 2019 2020 2021 2022 2023 2024 2025 Real GDP G ow h 3.7% 3.8% 3.7% 3.6% 3.6% 3.5% 3.5% Domes ic Nominal In e es Ra e 7.0% 7.0% 7.0% 7.0% 7.0% 10.0% 10.0% Nominal Exchange Ra e Dep ecia ion -0.4% -2.1% -1.5% -0.3% 0.7% 0.0% 0.0% P ima y Su plus -0.6% -1.0% -0.6% -0.5% -0.3% -0.3% -0.3% In la ion 4.0% 3.6% 3.6% 3.7% 3.7% 3.8% 3.9% Fo eign Nominal In e es Ra e 4.4% 4.4% 4.4% 4.4% 4.4% 5.0% 5.0% Deb /GDP a io 24.5% 24.9% 25.1% 25.2% 25.2% 25.6% 26.0% The sho - un dynamics o he deb - o-GDP a io a e epo ed in Table 3. Unde hese assump ions, deb dynamics appea o be s able. We expec he deb - o-GDP a io o inc ease 1.5 pe cen age poin s o GDP in he nex se en yea s, assuming ha he go e nmen will un a small p ima y de ici h oughou he p ojec ed ho izon; in his case, deb (as a sha e o GDP) shows a modes inc ease, going om 24.5 pe cen in 2019 o 26 pe cen in 2025.9 This baseline scena io conside s, wi h he pa h o he p ima y su plus, he expec ed o mos likely beha io o he iscal au ho i y. Ne e heless, i could be he case ha he iscal au ho i y decides o implemen an ac i e 9 A sensi i i y analysis o his esul s can be ound in Appendix A.2. 17 policy wi h a pa icula objec i e in mind. Fo example, in Table 4 we calcula e he p ima y su plus needed o s abilize he deb - o-GDP a io a i s cu en le el o 24.5 pe cen . This goal could be a ained e en i he go e nmen uns a small p ima y de ici (a nega i e ps) du ing he en i e p ojec ion ho izon, a p ima y su plus o -0.41 pe cen o GDP on a e age. This seems highly easible a e no ing ha he coun y has un a p ima y su plus o −0.35 pe cen o GDP on a e age o he las 19 yea s. Table 4. Sho -Te m App oach: S abilizing Deb - o-GDP Ra io a I s Cu en Le el Ac i e policy: PS needed o s abilize deb 2019 2020 2021 2022 2023 2024 2025 Real GDP G ow h 3.7% 3.8% 3.7% 3.6% 3.6% 3.5% 3.5% Domes ic Nominal In e es Ra e 7.0% 7.0% 7.0% 7.0% 7.0% 10.0% 10.0% Nominal Exchange Ra e Dep ecia ion -0.4% -2.1% -1.5% -0.3% 0.7% 0.0% 0.0% P ima y Su plus -0.6% -0.5% -0.5% -0.4% -0.3% -0.3% -0.3% In la ion 4.0% 3.6% 3.6% 3.7% 3.7% 3.8% 3.9% Fo eign Nominal In e es Ra e 4.4% 4.4% 4.4% 4.4% 4.4% 5.0% 5.0% Deb /GDP a io 24.5% 24.5% 24.5% 24.5% 24.5% 24.5% 24.5% We also simula e a scena io in which he go e nmen delibe a ely inc eases he p ima y su plus in one pe cen age poin o GDP in o de o inc ease in es men . To accoun o gene al equilib ium e ec s and he ac ha public in es men may ha e complemen a i ies wi h p i a e in es men , a ec ing posi i ely he g ow h a e o he economy, we adjus he expec ed u u e pa h o he eal GDP g ow h a e (using es ima es o he elas ici y be ween public in es men and ou pu g ow h) o accoun o inc eases in he g ow h a e a ising om highe le els o public spending. Table 5 con ains he esul o his exe cise, whe e we can obse ed ha deb - o-GDP a io inc eases, bu no d ama ically (5.7 pe cen age poin s o GDP), e en when he p ima y de ici inc eases by one pe cen age poin o GDP e e y single yea along he se en-yea ho izon.10 10 A g aphic depic ion o deb dynamics can be ound in Figu e 14 in Appendix A.2. 18 Table 5. Sho -Te m App oach: Inc easing P ima y De ici by 1 Pe cen Ac i e policy: Inc ease p. de ici by 1% 2019 2020 2021 2022 2023 2024 2025 Real GDP G ow h 3.7% 3.9% 4.0% 4.1% 4.0% 4.1% 4.2% Domes ic Nominal In e es Ra e 7.0% 7.0% 7.0% 7.0% 7.0% 10.0% 10.0% Nominal Exchange Ra e Dep ecia ion -0.4% -2.1% -1.5% -0.3% 0.7% 0.0% 0.0% P ima y Su plus -1.6% -2.0% -1.6% -1.5% -1.3% -1.3% -1.3% In la ion 4.0% 3.6% 3.6% 3.7% 3.7% 3.8% 3.9% Fo eign Nominal In e es Ra e 4.4% 4.4% 4.4% 4.4% 4.4% 5.0% 5.0% Deb /GDP a io 25.5% 26.9% 27.9% 28.8% 29.7% 30.5% 31.2% 3.3 Fan Cha s Un il now, deb sus ainabili y has been analyzed elying on medium and long- e m simula ions o he deb - o-GDP a io gi en speci ic mac oeconomic o ecas s and iscal policy assump ions wi hou any conside a ion o he unce ain y su ounding hese assump ions (besides he basic sensi i i y analysis p esen ed). This Fan Cha app oach is pa o wha is known as a p obabilis ic app oach o unce ain y analysis. In his con ex , unce ain ies a e cha ac e ized by he p obabili ies associa ed wi h e en s o ou comes o a se o a iables ha a ec he dynamics o public deb . This app oach ende s a p obabili y dis ibu ion o he deb - o-GDP a io. Ra he han simply p ojec ing one cen al scena io, his app oach inco po a es he s uc u e o andom shocks hi ing he domes ic economy o ob ain a comple e dis ibu ion o possible pa hs o deb - o-GDP a io, based on he dynamics p o ided by a Vec o Au o-Reg ession (VAR) econome ic model o ex e nal o ecas wi h o wi hou co ela ed e o s. The use o an cha s is a common p ac ice in he s udy o isk managemen in mone a y policy, as i is use ul o g aphically illus a ing he unce ain y su ounding in la ion o ecas s; om he e, i has been g adually ex ended o he analysis o deb sus ainabili y (A izala e al., 2008). The e a e se e al me hodologies o gene a e he Fan Cha s; he e we p esen esul s using he ex e nal o ecas s me hodology ollowing A izala e al. (2008), whe e p ojec ions o each o he a iables included in o he deb equa ion (equa ion (2)) a e made acco ding o: whe e: 19 and whe e is a ec o o ex e nal p ojec ions included in he deb equa ion and coming om he same sou ces as in he baseline scena io used in Sec ion 3.2, ητ is a ec o o simula ed e o s, wi h a iance Ω �, ha is es ima ed wi h he a iance-co a iance ma ix o he esiduals om an es ima ed VAR-model. Figu e 10 illus a es he esul s o compu ing he Fan Cha using his me hodology. I is clea om he cha ha he mos likely pa h o deb - o- a io is o con inue on a s able pa h along he p ojec ion ho izon. The p obabili y o exceeding he 30 pe cen o GDP ma k is slim. Ac ually, one ad an age o his me hodology is ha i allows us o calcula e he equency wi h which he deb - o-GDP a io su passes di e en h eshold alues in any o he p ojec ed yea s; in o he wo ds, we can calcula e he p obabili y ha he deb su passes o achie es ce ain h esholds. Fo example, acco ding o his es ima ion, su passing he 30 pe cen o GDP h eshold a o emen ioned has a p obabili y o 3 pe cen . Figu e 10. Ex e nal P ojec ions: Co ela ed E o s Ano he a ian o his ex e nal o ecas me hodology is one ha uses unco ela ed e o s. I simula es e o s ha ha e a a iance equal o he sample a iance o each o he se ies. Fan Cha s can addi ionally be cons uc ed using he poin es ima es coming om he VAR model, which cons i u e he cen al endency o he o ecas s o each o he a iables included in he simula ion. The simula ed pa hs a e he esul o he sum o he poin es ima e om he VAR and 20 one o he gene a ed e o s using he VAR-gene a ed a iance-co a iance ma ix. Bo h addi ional me hodologies we e applied, and esul s can be ound in Appendix B. 3.4 Mendoza-O iedo 5p oach Unde his app oach, Mendoza and O iedo (2004) assume ha he e is a go e nmen which is highly a e se o he isk o de aul on i s so e eign deb . This leads he go e nmen o espec a “Na u al Deb Limi ” (NDL), which ep esen s a c edible commi men o be able o epay e en in a iscal c isis. In u n, a “Fiscal C isis” is de ined as a long sequence o ad e se shocks o iscal e enues whe e public ou lays adjus o a ole able minimum. In his con ex , i he ac ual le el o deb emains highe han he NDL h eshold, he go e nmen aces a posi i e p obabili y o de aul . The c edible- epaymen -commi men alue o he deb ( he NDL) sa is ies he ollowing condi ion: (4) whe e d∗ he h eshold alue o GDP a io; min is he lowes possible ealiza ion o he ax e enue o GDP; emin ep esen s he minimum le el o go e nmen expendi u e o GDP a io ha can be sus ained i he coun y we e o en e a iscal c isis (in which = min). The Mendoza and O iedo App oach (MO) migh be he mos sui able me hodology o analyzing deb dynamics in Gua emala o wo easons. Fi s , his me hodology ocuses on e enue unce ain y; as men ioned abo e, one o he main iscal issues in his coun y is he inabili y o collec axes. Second, he p oblem p esen ed by he au ho s, ha o a go e nmen which is highly a e se o he isk o a collapse in i s iscal ou lays (which leads he go e nmen o espec he NDL), is akin o he appa en public-deb a e sion pe cei ed in Gua emala. This model equi es in o ma ion abou he ola ili y and pe sis ence o go e nmen e enues, he a e age le els o e enue and expendi u e, he size o he po en ial adjus men in expendi u e in he e en o eaching a c isis s a e, he wo ld eal in e es a e and he economy s eady s a e g ow h a e. The alues used o all hese a iables a e de ailed in Table 6. The a e age le el o e enues and expendi u es as sha es o GDP a e calcula ed om his o ical da a. Pe sis ence o go e nmen e enues a e app oxima ed by he au o eg essi e coe icien o he e enue- o-GDP cyclical componen ( om an HP il e ). One impo an aspec o he model is he 21 po en ial adjus men he economy can endu e in case o a iscal c isis. In case o he minimum e enue, we assume ha i will be wo s anda d de ia ions below i s mean (8.8 pe cen o GDP). Fo he maximum expendi u e adjus men , we ake wo di e en assump ions, he i s one is ha expendi u es can all wo s anda d de ia ions below i s mean in case o c isis, which is ai ly s anda d. Bu also, since his app oach is qui e sensi i e o he adjus men o go e nmen expendi u e, we assume al e na i ely ha he minimum le el o expendi u es ha can be a ained is equal o he minimum le el o expendi u es obse ed in ecen his o y (i.e., 8.1 pe cen o GDP). Table 6. Assump ions o he MO Model Assump ions Values Al e na i e Real in e es a e 1.05 S eady-S a e GDP G ow h 1.035 A e age le els o non-in e es expendi u es 11.5% Maximum Expendi u e Adjus men (2.1 s.d.) 3.2% Adjus ed Go e nmen Expendi u e 8.3% 8.1% A e age le els o e enues 11.2% Vola ili y o he go e nmen e enues 1.2% Pe sis ence o Tax e enue 48.3% Minimum Le els o e enues (-2 s.d.) 8.8% Ini ial Le el o Deb 24.5% Na u al Deb Limi 37.7% 46.6% Numbe o epe i ions o he simula ions 500 Acco ding o his app oach, he e is s ill oom o Gua emala o inc ease i s deb , as he es ima ed NDL (in bo h cases) is easonably highe han he cu en le els. Gi en he sensi i i y o di e en assump ions on e enue ola ili y and expendi u e adjus men , Figu e 11 depic s how he NDL will inc ease as expendi u e becomes mo e lexible, and he same happens as e enue ola ili y dec eases. On he con a y, as expendi u e becomes less lexible and ola ili y o e enues inc eases, he NDL will be lowe .11 11 See Table 9 in Appendix A.3 o da a on he sensi i i y o di e en assump ions on e enue ola ili y and expendi u e adjus men . 28 Appendix B. Addi ional Fan Cha s Figu e 15. Ex e nal P ojec ions: Unco ela ed E o s Figu e 16. VAR-Model P ojec ions 29 Re e ences Alesina, A., and P. Giuliano. 2015. “Cul u e and Ins i u ions.” Jou nal o Economic Li e a u e 53(4): 898–944. A izala, F. e al. 2008. “Deb Sus ainabili y Fan Cha s: Combining Mul i a ia e Reg ession Analysis and Ex e nal Fo ecas s.” Washing on, DC, Uni ed S a es: In e -Ame ican De elopmen Bank. www.edua doca allo.com/FC_05_05_2009_EC_FA.pd Ba aglini, M., and S. Coa e. 2008. “A Dynamic Theo y o Public Spending, Taxa ion, and Deb .” Ame ican Economic Re iew 98(1): 201–36. Bisin, A., and T. Ve die . 2017. “On he Join E olu ion o Cul u e and Ins i u ions.” NBER Wo king Pape 23375. Camb idge, Uni ed S a es: Na ional Bu eau o Economic Resea ch. Blancha d, O.J. 1990. “Sugges ions o a New Se o Fiscal Indica o s.” OECD Economics Depa men Wo king Pape 79. Pa is, F ance: O ganisa ion o Economic Co-ope a ion and De elopmen . Blancha d, O.J. e al. 1990. “The Sus ainabili y o Fiscal Policy: New Answe s o an Old Ques ion.” OECD Economic S udies 15: 7-36 Bohn, H. 2007. “A e S a iona i y and Coin eg a ion Res ic ions Really Necessa y o he In e empo al Budge Cons ain ?” Jou nal o Mone a y Economics 54(7):1837–1847. Bo ensz ein, E. e al. 2010. “Templa e o Deb Sus ainabili y: A Use Manual.” Technical No e IDB-TN-105. Washing on, DC, Uni ed S a es: In e -Ame ican De elopmen Bank. Bui e , W.H. 1985. “A Guide o Public Sec o Deb and De ici s.” Economic Policy 1(1):13–61. Ga in, M., and R. Pe o i. 1997. “Fiscal Policy in La in Ame ica.” In: B.S. Be nanke and J.J. Ro embe g, edi o s. NBER Mac oeconomics Annual. Volume 12. Camb idge, Uni ed S a es: Na ional Bu eau o Economic Resea ch and MIT P ess. Ghosh, A.R. e al. 2013. “Fiscal Fa igue, Fiscal Space and Deb Sus ainabili y in Ad anced Economies.” Economic Jou nal 123(566): F4–F30. Izquie do, A., e al., edi o s. 2018. Mejo Gas o pa a Mejo es Vidas: Cómo Amé ica La ina y el Ca ibe Puede Hace Más con Menos. De elopmen in he Ame icas epo . Washing on, DC, Uni ed S a es: In e Ame ican De elopmen Bank. K ogs up, S., and C. Wyplosz. 2010. “A Common Pool Theo y o Sup ana ional De ici Ceilings.” Eu opean Economic Re iew 54(2): 269–278. 30 Mendoza, E.G., and J.D. Os y. 2008. “In e na ional E idence on Fiscal Sol ency: Is Fiscal Policy ‘Responsible’”? Jou nal o Mone a y Economics 55(6): 1081–1093. Mendoza, E.G., and P.M. O iedo. 2004. “Public Deb , Fiscal Sol ency and Mac oeconomic Unce ain y in La in Ame ica: The Cases o B azil, Colombia, Cos a Rica, and Mexico.” NBER Wo king Pape 10637. Camb idge, Uni ed S a es: Na ional Bu eau o Economic Resea ch. Richa dson, G. 2008. “The Rela ionship be ween Cul u e and Tax E asion ac oss Coun ies: Addi ional E idence and Ex ensions.” Jou nal o In e na ional Accoun ing, Audi ing and Taxa ion 17(2): 67–78. Tauchen, G. 1986. “Fini e S a e Ma ko -Chain App oxima ions o Uni a ia e and Vec o Au o eg essions.” Economics Le e s 20(2): 177–181. Velasco, A. 2000. “Deb s and De ici s wi h F agmen ed Fiscal Policymaking.” Jou nal o Public Economics 76(1): 105–125. Weingas , B.R., K.A. Shepsle and C. Johnsen. 1981. “The Poli ical Economy o Bene i s and Cos s: A Neoclassical App oach o Dis ibu i e Poli ics.” Jou nal o Poli ical Economy 89(4): 642–664.