INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA

Transcription

1 Volume 9, Issue 1, pp.71-89, 2016 ISSN X DOI /rebs INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA Carmen PINTILESCU *, Mircea ASANDULUI **, Elena-Daniela VIORICĂ ***, Dănuţ-Vasile JEMNA **** Absrac Based on monhl-recorded daa for he period relaed o oupu growh and inflaion, we use heeroskedasic models in order o esimae he nominal and real uncerain in Romania. Real uncerain is derived from oupu growh volaili and nominal uncerain is derived from inflaion volaili. Of he 12 possible hpoheses regarding causal relaionships beween oupu growh, inflaion, nominal uncerain and real uncerain, we consider 7 hpoheses for which we find srong heoreical argumens and empirical evidence in lieraure. In order o ensure he robusness of he resuls, he Granger-causali ess are performed for 4, 8 and 12 lags, which are hen used o es 7 economic hpoheses. Kewords: oupu growh, inflaion, uncerain, heeroskedasic models JEL Classificaion: E31, E50 1. INTRODUCTION The insurance of a susainable economic growh and macroeconomic sabili represens a major objecive for he economic polic of ever sae. The recen economic crisis deermined larger volailiies of he oupu growh and inflaion and accenuaed he uncerain regarding he evoluion of he mos * Carmen PINTILESCU, Alexandru Ioan Cuza Universi of Iasi, Romania, Facul of Economics and Business Adminisraion, carmen.pinilescu@uaic.ro ** Mircea ASANDULUI, Alexandru Ioan Cuza Universi of Iasi, Romania, Facul of Economics and Business Adminisraion, mircea.asandului@uaic.ro *** Elena-Daniela VIORICĂ, Alexandru Ioan Cuza Universi of Iasi, Romania, Facul of Economics and Business Adminisraion, dana.viorica@gmail.com **** Dănuţ-Vasile JEMNA, Alexandru Ioan Cuza Universi of Iasi, Romania, Facul of Economics and Business Adminisraion, danu.jemna@uaic.ro

2 72 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA imporan macroeconomic indicaors. In his conex, he sud of he relaionship among inflaion, oupu growh and heir uncerainies represens an imporan opic for he naional polic makers as well as for he specialiss. Saring from he 1990s, besides he implemenaion of economic reforms and he ransiion o marke econom, he Cenral and Eas European counries have experienced grea inflaionis phenomena. Among hese counries, Romania experienced one of he highes inflaion raes (256.1 % in 1993). The inconsisenc of he economic reform measures deermined a significan variabili of inflaion, 2005 being he firs ear in which a single digi inflaion (9 %) was recorded. Thus, in Romania s case, inflaion is considered o be one of he main phenomena of insabili of he economic environmen and he uncerain abou fuure inflaion can lead o uncerain abou oher economic variables. For Romania, as a member of he European Union, an imporan objecive of economic polic is o ensure real convergence wih developed economies and his can be achieved b regisering a high level of growh in conjuncion wih macroeconomic sabili. The economic crisis which sared in 2008 resuled in a significan reducion of he GDP growh, which emphasized he variabili and uncerain of ensuring susainable economic growh. In his economic conex, he analsis of he relaionship among inflaion, economic growh and heir uncerainies for Romania ma represen a foundaion for he formulaion of proposals ha would aim a he insurance of a susainable economic growh. In he presen paper, we analse he causaliies beween inflaion, oupu growh and heir uncerainies in Romania. We use monhl-recorded daa for he period relaed o oupu growh and inflaion o esimae heir uncerainies b differen heeroskedasic models. Inflaion is measured b means of he Consumer Price Index (CPI) and Oupu Growh is measured b he Indusrial Producion Index (IPI). The inflaionis phenomenon represened for Romania one of he main facors of insabili of he economic environmen. In compliance wih he sraeg of inflaion argeing, ha aimed, in he pre-adherence program o he European Union, o reach an inflaion level below 10 %, we spli he daa series in wo subsamples because in he firs par of he period we have ver high values for inflaion. We chose he breaking-poin b deermining he firs monh for which he inflaion average for he las 12 monhs is lower han 10 %. This condiion was

3 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA 73 fulfilled for he November Therefore, he firs sub-sample comprises he values regisered unil Ocober 2004, while he second sub-sample comprises he values regisered for he period November 2004 November Of he 12 possible hpoheses regarding causal relaionships beween uncerainies, oupu growh and inflaion, we consider he ones for which we find srong heoreical argumens and empirical evidence in he lieraure. In order o ensure he robusness of he resuls, he Granger-causali ess are performed for 4, 8and 12lags, which are hen used o es he seleced hpoheses. The paper is srucured as follows: Secion 2 presens he hpoheses regarding he causali beween oupu growh, inflaion and heir uncerainies, as well as he empirical lieraure o dae; Secion 3 describes he mehodolog, namel he heeroskedasic models used for he esimaion of he condiional residual variances as measures of uncerain; Secion 4 presens he resuls, while Secion 5 concludes. 2. LITERATURE REVIEW The relaionship beween inflaion, oupu growh and heir uncerainies has been exensivel researched since he 1970s, following he publicaion of Milon Friedman s (1977) Nobel lecure. The mos discussed issue concerns he relaionship beween inflaion and is uncerain, which was firs analsed b Okun (1971), who found a posiive relaionship beween he inflaion rae and is variabili for 17 OECD counries. Ye he real conribuion would laer belong o Friedman s (1977) work on he real effecs of inflaion. The laer auhor saes ha an increase in inflaion leads o more uncerain, which ulimael leads o a decrease in oupu. The lieraure on he subjec is hus fairl vas and regards boh he heoreical foundaion for he exisence and direcion of hose relaionships and he empirical evidence found for he developed and ransiion economies (Grier and Perr, 2000; Neanidis and Savva, 2012; Harmann and Roesel, 2013, Gokas and Disbudak, 2014; Clemens and Galvao, 2014). In he presen paper, we analse 7 ou of he 12 possible hpoheses regarding he causal relaionship among inflaion, oupu growh and heir uncerainies (Grier and Perr, 2000) and we find srong heoreical as well as

4 74 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA empirical proof in various sudies (Grie e al. (2004), Founas e al (2006), Founas e al. (2007)). H1: Inflaion Granger-causes inflaion uncerain The firs hpohesis according o which inflaion Granger-causes inflaion uncerain is he mos invesigaed in he lieraure. Various sudies prove ha inflaion causes imporan coss for he econom, one of which is relaed o uncerain abou fuure inflaion raes due o he effecs i has on fuure invesmen and saving decisions of economic unis. Friedman (1977) and Ball (1992) found empirical evidence for a posiive relaionship beween inflaion and inflaion uncerain, while oher auhors, such as Pourgerami and Maskus (1987), Hungarian and Zilberfarb (1993), Davis and Kanago (2000) found an opposie causal relaionship. Seen as a cause or an effec of inflaion, he lieraure highlighs he negaive effecs of inflaion uncerain on economic variables like invesmen and growh. More recen sudies confirm he causal relaionship beween inflaion and inflaion uncerain (Grier and Perr, 1998; Founas, 2001; Hwang, 2001; Thornon, 2007, 2008). H2: Inflaion uncerain Granger-causes oupu growh In his Nobel lecure, Friedman (1977) argued ha an increase in inflaion will lead o more uncerain abou inflaion, leading o a decrease in oupu. Inflaion uncerain affecs resource allocaion hrough is effecs on ineres raes, he real cos of inpus and final goods prices. Conrar o Friedman, Dose and Sare (2000), using "cash-in-advance" pe of model, showed ha high levels of inflaion uncerain have a posiive impac on economic growh. Increased inflaion uncerain causes increased savings and hus increased invesmen and oupu. The relaionship beween inflaion and growh is confirmed b recen sudies b Hasanov and Oma (2011) and Khan e al. (2013) ha invesigaed he phenomenon in he counries of Cenral and Easern Europe.

5 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA 75 H3: Inflaion uncerain Granger-causes inflaion The hpohesis ha uncerain abou inflaion has an influence on inflaion has been esed in numerous sudies. Cukierman and Melzer (1986) found a posiive relaionship beween inflaion and inflaion uncerain, arguing ha when inflaion uncerain increases, monear auhoriies adop an opporunisic behaviour in order o simulae oupu growh b increasing he mone suppl, hereb generaing higher inflaion. In he opposie direcion, Holland (1995) shows a negaive relaionship beween inflaion and inflaion uncerain. Holland found ha, in condiions of high uncerain, monear auhoriies have a sabilizing behaviour manifesed b reducing he mone suppl in order o reduce he negaive effecs of inflaion uncerain on he econom (also known as he Fed hpohesis). The opporunisic or sabilizaion behaviour of he monear auhoriies depends on he level of independence of he cenral bank (Grier and Perr, 1998), and on he economic siuaion of a counr. Therefore, he exen o which monear polic affecs inflaion is difficul o predic. Recen sudies have confirmed his hpohesis (Grier and Perr, 1998; Apergis, 2004; Naraan e al., 2009; Jiranakul and Opiela, 2010). H4: Oupu growh uncerain Granger-causes inflaion In addiion o inflaion uncerain, i is generall acknowledged ha he real uncerain can affec he inflaion rae. Devereux (1989) and more recenl Cukierman and Gerlach (2003) idenified a posiive relaionship beween real uncerain and inflaion, while Cukierman and Melzer (1986) claimed he exisence of a negaive relaionship beween he wo variables. Using he Barro- Gordon model, Devereux (1989) showed ha he opimal amoun of wage indexaion is reduced when he he real uncerain increases causing he monear auhori o generae unexpeced inflaion for agens. Cukierman and Melzer (1986) saed ha a greaer uncerain abou economic growh reduces inflaion uncerain, herefore i reduces inflaion. More recen sudies conduced mosl in developed counries have shown a relaionship beween he growh and inflaion uncerainies (Founas and Karanasos, 2007).

6 76 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA H5: Oupu growh uncerain Granger-causes oupu growh Numerous sudies have idenified a relaionship beween real uncerain and oupu growh, bu he sign of his effec is ambiguous. Mirman (1971) argues ha high levels of economic growh uncerain causes an increase in he savings rae and hus i leads o higher economic growh rae. Afer suding he impac on invesmens in riskier echnologies, Black (1987) shows ha here is a posiive relaionship beween he real uncerain and he oupu growh if he expeced reurn is large enough o offse he risks associaed wih he invesmens. On he oher hand, Pindck (1991) showed ha beween oupu growh and oupu growh uncerain here is a negaive relaionship deermined b he fac ha higher uncerain abou fuure profis generaed b hese invesmens leads o a dela or cancellaion of invesmens, leading furher o a lower level of economic growh. Recen sudies have confirmed he influence of oupu growh on real uncerain, bu wih he same ambigui concerning he sign of his relaionship (Henr and Olekalns, 2002; Blackburn and Pelloni, 2004; Grier e al., 2004; Founas e al., 2004; Founas and Karanasos, 2007; Naraan, 2009). H6: Oupu growh Granger-causes inflaion uncerain Regarding he impac of oupu growh on inflaion uncerain, Bruner (1993) demonsraed he exisence of a negaive relaionship beween he wo. On he oher hand, Hungarian and Zilberfarb (1993) found ha high levels of economic growh can lead o low levels of inflaion, resuling in a reduced inflaion uncerain. More recenl, Grier and Perr (2000), Elder (2004) and Founas e al (2006) provide evidence ha nominal uncerain is a negaive deerminan of growh in almos all he developed counries considered in heir sudies. H7: Oupu growh Granger-causes oupu growh uncerain The leas esed hpohesis in he lieraure is he assumpion on he influence of oupu growh on real uncerain. A negaive relaionship beween he wo can be explained b he fac ha, in he shor erm, susained economic growh leads o an increase in inflaion. Based on Friedman's heor (1977), his will increase inflaion uncerain. This, paired wih Talor's heor (1979) ha here is a rade-off beween inflaion uncerain and oupu growh uncerain, will resul

7 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA 77 in a lower oupu growh uncerain. A posiive relaionship beween oupu growh and real uncerain can be explained b he fac ha, when he economic growh rae decreases, he monear polic s response makes he inflaion rae become more uncerain, hus meaning ha i will be characerized b a higher uncerain (Brunner, 1993). A posiive relaionship beween he oupu growh and he real uncerain can be explained b he fac ha, when he economic growh rae decreases, he monear polic s response makes he inflaion rae highl uncerain (Brunner, 1993). Furhermore, under he Talor effec, we can conclude ha in his case he uncerain of oupu growh will increase, hus highlighing he posiive direcion of he relaionship. The debaes over he relaionships beween inflaion, oupu growh and heir uncerainies, and respecivel over he empirical esimaions of hese casual relaionships are ver numerous in he lieraure. However, here is no an empirical unanimi in he resuls of he sudies, which can be explained b he high degree of cross-counr heerogenei (Greenspan, 2004) or b he acual specific challenges ha cause fundamenal changes in he monear policies (Harmann and Roesel, 2014). For Romania, we found few sudies concenraing on he analsis of he causal relaionship beween inflaion, oupu growh and heir uncerainies. Jemna e al (2014) and Viorica e al (2014) analse onl 2 hpoheses regarding he causali beween inflaion and inflaion uncerain. In a larger sud regarding he European emerging economies, Pinilescu e al (2014) ake ino consideraion, 5 of he 7 hpohesis esed in he presen paper, which uses wo subsamples (Ocober 1990-Ocober 2004 and November 2004-November 2014), for a beer daa homogenei. 3. DATA AND METHODOLOGY Following previous sudies b Gillman and Harris (2008), Kewebar and Nenovsk (2013) and Jemna e al (2014), we use he Consumer Price Index (CPI) o measure he inflaion, and he Indusrial Producion Index (IPI) o measure he oupu growh, as was previousl done b Founas, Karanasos and Kim (2006), Hasanov and Oma (2011), Pinilescu e al (2004). The daa considered (Januar 1990-November 2014) was rerieved from he IMF indicaors, using he Inernaional Financial Saisics (IFS). Inflaion (π) is

8 78 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA measured b he annualized monhl difference of he log CPI [π = ln(cpi /CPI -1) x 1200]. The oupu growh ( ) is considered hrough he annualized monhl difference in he log of he IPI [ =ln(ipi /IPI -1) x 1200]. When discussing he uncerain of he inflaion and he oupu growh, oher means can be used for heir esimaion. One of he simples mehod was proposed b Hafer (1986) and Davis and Kanago (2000), who use sandard deviaion o esimae uncerain. Anoher manner is inroduced b Johnson (2002), who esimaes he uncerain b he error generaed b a simple forecas model. A ver large par of he lieraure proposes he esimaion of he uncerain, wih good resuls, b means of he condiional variance esimaed hrough a heeroskedasic model. From his famil of models, auhors like Evans (1991), Võrk (2000) or Neanidis and Savva (2012) use models ha are simpler, eas o conrol and o manipulae, such as he GARCH model. A more complex mode, EGARCH, is used in sudies like Asghar e al (2011), Baharumshah e al (2011). Grier e al (2004), Ajevskis (2007), Khan (2010) or Pinilescu e al (2014) use GARCH-in-mean models for he esimaion of he uncerain. A differen approach from he previousl presened one is used b Berumen e al (2011), where he esimae uncerain hrough a model of he pe sochasic volaili in mean (SVM). In he presen analsis, for esimaing uncerain we ake ino consideraion more heeroskedasic models (ARCH, GARCH, EGARCH, GARCH-M, PARCH), ha were in accordance wih he lieraure and recen research papers. From he esimaed models, we will choose he bes fied model based on he values of he informaion crieria Akaike (AIC) Schwarz (SIC) and Hannan-Quinn(HQ). The firs par of he analsis consiss in esing he saionari of he series aken ino consideraion. If we do no confirm he saionari hpohesis for one of he series, hen we have o saionaries he series hrough one of he radiional procedures, such as he creaion of he series of firs-order differences. We will use he Augmened Dicke-Fuller (ADF), Phillips-Perron and Kwiakowski-Phillips- Schmid-Shin (KPSS) ess in order o es he saionari. Then, according o he mehodolog proposed b Founas and Karanasos (2007), based on he saionar series, we will esimae a bi-variae VAR model which compues he condiional means of he inflaion rae and he oupu growh.

9 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA 79 The scope of his mehod is o idenif he opimum lag lengh of heir correlaion. The number of lags is chosen using he Akaike (AIC) and Schwarz (SIC) informaion crieria. Afer obaining he opimum lag lengh, we esimae he uncerainies for he wo considered variables, namel inflaion and oupu growh. B means of Granger causali ess, we search for he exisence of a causali relaionship beween hese wo variables. As saed before, we measure he uncerainies hrough he condiional variance of hese variances esimaed based on he heeroskedasic models. We consider five heeroskedasic models for esimaing uncerain, namel: ARCH, GARCH, EGARCH, GARCH-M and PARCH. The model used for esimaing he condiional variance, which measures he inflaion uncerain, will be chosen based on he Akike and Schwarz informaion crieria applied o he lowes absolue values. The saring poin for man subsequen developmens is he ARCH model (Engle, 1982), which was highl debaed and analsed (Bollerslev, Chou and Kroner (1992), Bera and Higgins (1993), Bollerslev, Engle and Nelson (1994) or Diebold and Lopez (1995)). Unlike he previous models, he ARCH model does no rel on he pas sandard deviaion, bu on he condiional variances, noed wih 2 h ( h ). In he ARCH process of order q, he condiional variance akes he following form: h q j1 2 j j (1) where ω > 0 and α j 0 o ensure he condiions ha h should alwas be posiive. The GARCH model (Bollerslev, 1986) offers he possibili o measure uncerain for inflaion and oupu growh, including he lagged condiional variances as auoregressive erms. For GARCH (1,1) he model has he specificaion: Y ' (2) X

10 80 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA where X is a kx1 vecor of independen variables, β is a kx1 vecor of regression coefficiens, and ε is he residual, respecing he condiion ~ N 0,. h h is he condiional variance, which is esimaed b he equaion: h h (3) EGARCH (1,1) proposed b Nelson (1991) has he same specificaion and he condiional variance is obained b he equaion: where. h ln h 0 ln (4) 1 h GARCH-M (1,1) (Engle, Lilien and Robins, 1987) model has he same condiional variance as GARCH (1,1), bu he condiional mean depends on is own condiional variance. This model has he specificaion: ' Y X h (5) Anoher developmen of he GARCH models was achieved b Talor (1986) and Schwer (1989), who inroduced he sandard deviaion GARCH model, where he sandard deviaion raher han he variance is modelled. This model, along wih several ohers, was generalized b Ding e al. (1993) wih he Power ARCH specificaion. In he Power ARCH model, he power parameer δ of he sandard deviaion can be esimaed raher han imposed, and he opional γ parameers are added o capure asmmer of up o order r: h q j1 h j j p i1 1 i i (6) The Bollerslev (1986) model ses δ = 2, γ = 0, and he Talor (1986) model ses δ = 1 and γ = 0. Empirical esimaes indicae he power erm is sample dependen and values of near 1 are common in he case of sock daa (Ding e al. 1993), while for foreign exchange daa he power erm varies beween 1 and 2 (McKenzie and Michell, 2002). Afer choosing he bes fied model, we generae he condiional variance series and we will use hese series as he uncerainies of he wo macroeconomic variables. The nex sep is o es he Granger causali for 4, 8 and 12 lags. The

11 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA 81 causali ess will be applied for he 7 relaionships beween inflaion, oupu growh and heir uncerainies, wih he corresponding number of lags. The resuls of he ess will poin ou he pe of correlaion beween he variables. For he saisical significan correlaions we will esimae a VAR model, in order o deermine he sign of he causali. 4. EMPIRICAL RESULTS The empirical analsis performed in our paper concerns he causali beween inflaion, oupu growh, inflaion uncerain and oupu growh in Romania. Because he analsed period ( ) is characerized a is beginning b he sruggle of he polic makers o creae he insiuional framework for a free econom, in he 1990 s we encouner ver high inflaion raes. In order o capure he specifici of ha decade, we worked wih 2 sub-samples. The breaking poin was chosen in compliance wih an earlier analsis (Jemna e al, 2014). Therefore, he wo sub-samples used are Ocober 1990 Ocober 2004 and November 2004 November Afer 2001, he monear polic sraeg of he Naional Bank of Romania was direc inflaion argeing. We chose he breaking poin considering ha November 2004 was he firs monh o have an inflaion average for he las 12 monhs lower han 10. The breaking poin highlighed a level of sead-sae inflaion uncerain which reflecs he uncerain when inflaion is a is seadsae level and here are no shocks o he ssem (Caporale e al, 2010). This breaking poin was esed b means of he Chow es. The resuls indicaed ha November 2004 represens a breaking poin in he inflaion series. As saed previousl, he inflaion is measured b he annualized monhl difference of he log CPI and he oupu growh b he annualized monhl difference of he log of he IPI. The summar saisics of hese wo variables are provided in Table 1. Table 1: Summar saisics for inflaion and oupu growh Macroeconomic indicaor Sub-sample 1 Ocober Ocober 2004 Sub-sample 2 November 2004 November 2014 Mean Sd.dev. Jarque-Bera Mean Sd.dev. Jarque-Bera Inflaion (CPI) Oupu growh (IPI)

12 82 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA Table 1 highlighs ver high values of inflaion and ver low values for he oupu growh in he firs sub-sample. The firs decade is represened b he period when Romania begun he ransiion owards he marke econom ssem and he price liberalizaion was one of he firs economic measures o be adoped. Following he measures of economic polic adoped in he 1990s, he earl inflaion in Romania exceeded 200 %. The reformaion and resrucuring process of he economic ssem deermined imporan oupu drops in he earl sages of ransiion. The second sub-sample is characerized b a more sable and significanl lower inflaion deermined b an increased sabili of he macroeconomic environmen. The oupu growh has a low mean value and is characerized b a high variabili, enhanced b he recen economic crisis. a. Tesing for saionari We sar b esing he saionari of daa series using he Augmened Dicke-Fuller (ADF) and Phillips-Perron (PP) ess, for which he null hpohesis is he non-saionari hpohesis and he Kwiakowski-Phillips-Schmid-Shin (KPSS) es, for which he null hpohesis is he saionari one. The resuls of he saisical esing are presened in Table 2. Table 2: Uni roo ess Macroeconomic Sub-sample 1 Sub-sample 2 indicaor Ocober Ocober 2004 November 2004 November 2014 ADF PP KPSS ADF PP KPSS Inflaion (CPI) Oupu growh (IPI) Noe: A consan and 12 lagged difference erms are used for he Augmened Dicke-Fuller es. The MacKinnon criical value for he rejecion of he uni roo null hpohesis a he 1 % significance level is The KPSS criical values for he rejecion of he uni roo null hpohesis a he 1 %, 5 % and 10 % significance levels are 0.739, and 0.347, respecivel. The applied ess highligh ha he analsed variables are saionar for he 2 sub-samples. Because he saionari condiion is saisfied for he 2 variables, for boh periods, we coninue our analsis b appling VAR and Granger-causali es mehods in order o idenif he adequae number of lags for each variable, as well as he lead-lag ineracions beween variables.

13 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA 83 b. Modelling of he inflaion, oupu growh and heir uncerainies Using he Likelihood Raio (LR), Akaike (AIC), Schwarz (SIC) and Hannan-Quinn (HQ) informaion crieria, we idenif a maximum of 3 lags in he firs period and 12 lags in he second period for each relaion beween inflaion and oupu growh. Using VAR Granger Causali, we idenified he dependen variable which will be used o esimae he five specified models (ARCH, GARCH, EGARCH, GARCH-M and PARCH). The bes fied model is chosen using informaion crieria. The esimaed equaions for inflaion, oupu growh and heir uncerainies are presened in he Table 3. Table 3: Inflaion, oupu growh, inflaion uncerain and oupu growh uncerain equaions Sub-sample 1: Ocober Ocober 2004 Inflaion equaion Oupu growh equaion Inflaion uncerain equaion h , 1, h, Oupu growh h uncerain equaion 6.41, 1 h, 1 Sub-sample 2: November 2004 November 2014 Inflaion equaion Oupu growh equaion Inflaion uncerain h equaion 3.48, 1 h, 1 Oupu growh h uncerain equaion 3.89, 1 h, Noes: Absolue -saisics are given below he esimaions. For measuring uncerainies, we generae he condiional variance series given b he chosen heeroskedasic model. Then we es he 7 economic

14 84 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA hpoheses using he Granger-causali mehodolog. The resuls of Grangercausali ess beween inflaion, oupu growh and heir uncerainies are presened in Table 4. In order o ensure he robusness of he resuls, he Grangercausali ess are performed for 4, 8 and 12lags. Table 3: Granger-causali ess beween inflaion, oupu growh and heir uncerainies Tesed hpohesis H 1: Inflaion Granger-Causes inflaion uncerain H 2: Inflaion uncerain Granger- Causes oupu growh H 3: Inflaion uncerain Granger- Causes inflaion H 4: Oupu growh uncerain Granger- Causes inflaion H 5: Oupu growh uncerain Granger- Causes oupu growh H 6: Oupu growh Granger-Causes inflaion uncerain H 7: Oupu growh Granger-Causes oupu growh uncerain Sub-sample 1 Ocober Ocober 2004 Sub-sample 2 November 2004 November lags 8 lags 12 lags 4 lags 8 lags 12 lags * * 76.86* 19.19* 12.83* 8.14* (+) (+) (+) (+) (+) (+) 2.92* 2.26* 1.97* * 1.23 (-) (-) (-) (-) 4.31* 2.44* 2.76* 2.28* 1.90* 1.37 (-) (-) (-) (-) (+) * (-) * 1.88* (+) (+) 3.31* * (-) (-) Noe: Figures are F-saisics. +(-) indicaes wheher he sum of he lagged coefficiens of he causing variable is posiive/negaive * denoes he significance a he 0.05 level. For he saisicall significan causali relaions, we deermined he sign of he relaionship b means of VAR. The Friedman-Ball hpohesis is he onl hpohesis ha is full confirmed for boh sub-samples. The second Friedman hpohesis is full validaed for he firs period and pariall validaed for he second one. A negaive causali beween inflaion uncerain and inflaion is obained for he period before 2004, while for he second period he resuls are inconsisen. Srong empirical evidence was also found for he posiive causali beween oupu growh and inflaion uncerain afer 2004, while for he negaive relaion beween oupu growh and oupu growh uncerain we found evidences for he firs period. 5. CONCLUSIONS I is widel acceped ha an econom characerized b low inflaion is beneficial for he growh and sabili of he macroeconomic environmen.

15 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA 85 Imporan variaion of inflaion and economic growh generaes an increase in uncerain abou he fuure evoluion of hese indicaors, which ma affec he decision-making process of business and consumers. In his paper, using monhl-recorded daa for Romania (for he period )7 hpoheses were esed regarding he causal relaionships beween inflaion, oupu growh and heir uncerainies, for which srong heoreical and empirical evidence was found in lieraure. Using real economic crieria and a breaking poin es, we buil wo subsamples of daa (Ocober 1990-Ocober 2004 and November 2004-November 2014). Empirical sud resuls confirmed he Friedman-Ball hpohesis for boh periods. In he firs sub-sample, he price liberalizaion measures adoped in Romania in he earl 1990s caused ver high values for inflaion, which led o an increase of uncerain abou fuure inflaion raes. Increased uncerain abou he inflaion generaed high risks on long-erm invesmens, on ineres raes and on oher economic variables, generall affecing he economic acivi. Thus he negaive effec of inflaion uncerain on economic growh emerged ver clearl in Romania, in a ime characerized b a high macroeconomic volaili. In 2001, he Naional Bank of Romania has se inflaion argeing as a priori of monear polic sraeg, specificall implemened in Thus, he second sub-sample corresponds o a period in which Romania was characerized b a lower variaion of inflaion, which led o a reducion in uncerain abou inflaion. These resuls, consisen wih hose of Founas e al (2004) and Harmann and Roesel (2013), emphasize he fac ha in Romania s case, he effecs of inflaion on oupu growh have been sronger han in oher counries, due o he fac ha polic makers had o srenghen heir effors in order o keep inflaion low. Keeping a low inflaion rae and assuring a sable macroeconomic environmen suppors he Romanian Naional Bank s sraeg aimed a ensuring a real convergence wih he European Union. Presenl, measures adoped b he auhoriies for mainaining a reduced level of inflaion have accenuaed he appariion of a phenomena never encounered before in he Romanian econom, namel deflaion. Even hough, in he shor erm, he deflaionis shocks ma have posiive effecs, in he long erm, his can generae negaive effecs (decreased producion, increased unemplomen rae, sagnaion of he economic growh).

16 86 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA Moreover, he implemenaion in Romania, as of 1 s June 2015, of he measure regarding he decreased Value Added Tax on food and non-alcoholic drinks and he new Fiscal Code (as of 1 s Januar 2016), will lead o a coninuous decrease ing in he inflaion rae. In order o conrol deflaion, he polic makers should adop a mixure of monear and fiscal policies which would mainain price sabili and assure a susainable economic growh. REFERENCES 1. Apergis, N., Inflaion, oupu growh, volaili and causali: evidence from panel daa and he G7 counries. Economics Leers, 83(2), Asghar, A., Ahmad, K., Ullah, S., Zaman, B., and Rashid, M. T., The Relaionship beween Inflaion and Inflaion Uncerain: A Case Sud for SAARC Region Counries. Inernaional Research Journal of Finance and Economics, 66, Ajevskis, V., Inflaion and Inflaion Uncerain in Lavia. Working Papers 2007/04, Lavijas Banka. 4. Baharumshah, A.Z., Hasanov, A. and S. Founas., Inflaion and inflaion uncerain: Evidence from wo Transiion Economies. Discussion Paper Series 2011/05, Deparmen of Economics, Universi of Macedonia. 5. Ball, L., Wh Does High Inflaion Raise Inflaion Uncerain?. Journal of Monear Economics, 29, 3, Bera, A. K. and Higgins, M. L., ARCH models: properies, esimaion and esing. Journal of economic surves, 7(4), Berumen, M., Yalcin, Y. and J. Yildirim., The inflaion and inflaion uncerain relaionship for Turke: a dnamic framework. Empirical Economics, 41, 2, Black, F., Business Ccles and equilibrium. Basil Blackwell. 9. Blackburn, K. and Pelloni, A., On he relaionship beween growh and volaili. Economics Leers, 83(1), Bollerslev, T., Engle, R. F. and Nelson, D. B., ARCH models. Handbook of economerics, 4, Bollerslev, T., Generalized Auoregressive Condiional Heeroskedasici. Journal of Economerics, 31, 3, Bollerslev, T., Chou, R. Y. and Kroner, K. F., ARCH modeling in finance: A review of he heor and empirical evidence. Journal of economerics, 52(1), Brunner, A. D. and Hess, G. D., Are higher levels of inflaion less predicable? A sae-dependen condiional heeroskedasici approach. Journal of Business & Economic Saisics, 11(2), Caporale, G. M., Onorane, L. and Paesani, P., Inflaion and Inflaion Uncerain in Euro Area. Caegor 7, Monear Polic and Inernaional Finance (CESIFO Working Paper No. 2720). 15. Clemens, M.P. and Galvão, A.B., Measuring Macroeconomic Uncerain: US Inflaion and Oupu Growh, ICMA Cenre Discussion Papers in Finance icma-dp , Henle Business School, Reading Universi

17 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA Cukierman, A. and Melzer, A., A Theor of Ambigui, Credibili, and Inflaion Under Discreion and Asmmeric Informaion. Economerica, 54, 5, Cukierman, A. and Gerlach, S., The inflaion bias revisied: heor and some inernaional evidence. Mancheser School, Universi of Mancheser, 71, Davis, G. and Kanago, B., The Level and Uncerain of Inflaion: Resuls from OECD Forecass. Economic Inquir, 38, 1, Diebold, F. X. and Lopez, J. A., Modeling volaili dnamics. Springer Neherlands, Devereux, M., A Posiive Theor of Inflaion and Inflaion Variance. Economic Inquir, 27, 1, Ding, Z., Granger, C. W. and Engle, R. F., A long memor proper of sock marke reurns and a new model. Journal of empirical finance, 1(1), Dose, M. and Sare, P., Inflaion Uncerain and Growh in a Cash-in-Advance Econom. Journal of Monear Economics, 45, 3, Elder, J., Anoher perspecive on he effecs of inflaion uncerain. Journal of Mone, Credi and Banking, Engle, R.F., Lilien, D.M. and Robins, R. P., Esimaing Time Varing Risk Premia in he Term Srucure: The ARCH-M Model. Economerica, 55, Evans, M., Discovering he Link Beween Inflaion Raes and Inflaion Uncerain. Journal of Mone, Credi and Banking, 23, 2, Founas, S., Karanasos, M. and Kim, J., Inflaion Uncerain, Oupu Growh Uncerain and Macroeconomic Performance. Oxford Bullein of Economics and Saisics, 68, 3, Founas, S. and Karanasos, M., Inflaion, Oupu Growh, and Nominal and Real Uncerain: Empirical Evidence for he G7. Journal of Inernaional Mone and Finance, 26, 2, Friedman, M., Nobel Lecure: Inflaion and Unemplomen. Journal of Poliical Econom, 85, 3, Gillman, M. and Harris, M.N., The Effec of Inflaion on Growh: Evidence from a Panel of Transiion Counries. Cardiff Economics Working Paper E2008/25. Cardiff Business School, Cardiff, UK. 30. Gökaş, P. and Dişbudak, C., Modelling Inflaion Uncerain wih Srucural Breaks Case of Turke ( ). Mahemaical Problems in Engineering. 31. Greenspan, A., Risk and uncerain in monear polic. American Economic Review, Grier, K.B. and Perr, M.J., On inflaion and inflaion uncerain in he G7 counries. Journal of Inernaional Mone and Finance, 17, 4, Grier, K.B. and Perr, M.J., The effecs of real and nominal uncerain on inflaion and oupu growh: some garch-m evidence. Journal of Applied Economerics, 15, 1, Grier, K.B., Henr, O.T., Olekalns, N. and Shields, K., The asmmeric effecs of uncerain on inflaion and oupu growh. Journal of Applied Economerics, 19, Hafer, R. W., Inflaion uncerain and a es of he Friedman hpohesis. Journal of Macroeconomics, 8, 3,

18 88 Carmen PINTILESCU, Mircea ASANDULUI, Elena-Daniela VIORICĂ, Dănuţ-Vasile JEMNA 36. Harmann, M. and Roesel, J., Inflaion, oupu and uncerain in he era of inflaion argeing A muli-econom view on causal linkages. Journal of inernaional mone and finance, 37, Hasanov, M. and Oma, T., The Relaionship Beween Inflaion, Oupu Growh and Their Uncerainies: Evidence from Seleced CEE Counries. Emerging Markes Finance and Trade, 47, Henr, O. T. and Olekalns, N, (2002. The effec of recessions on he relaionship beween oupu variabili and growh. Souhern Economic Journal, Holland, S., Inflaion and Uncerain: Tess for Temporal Ordering. Journal of Mone,Credi, and Banking, 27, 3, Jemna, D. V., Pinilescu, C., Viorica, E. D. and Asandului, M., Inflaion and inflaion uncerain in Romania. Economic compuaion and economic cberneics sudies and research, 48,1, Jiranakul, K. and Opiela, T. P., Inflaion and inflaion uncerain in he ASEAN-5 economies. Journal of Asian Economics, 21, 2, Johnson, D., The effec of inflaion argeing on he behavior of expeced inflaion: evidence from an 11 counr panel. Journal of Monear Economics, 49, 8, Khan, M., Inflaion and Srucural Oupu Growh Variabili in Bulgaria. Working Papers Series. Universie D' Orleans, France. 44. Khan, M., Kebewar, M. and Nenovsk, N., Inflaion Uncerain, Oupu Growh Uncerain and Macroeconomic Performance: Comparing Alernaive Exchange Rae Regimes in Easern Europe. Working Papers Series. Universie D' Orleans, France. 45. McKenzie, M. and Michell, H., Generalized asmmeric power ARCH modelling of exchange rae volaili. Applied Financial Economics, 12, 8, Mirman, L. J., Uncerain and Opimal Consumpion Decisions. Economerica, 39, 1, Naraan, P. K., Naraan, S. and Smh, R., Undersanding he inflaion oupu nexus for China. China Economic Review, 20, 1, Neanidis, K. and Savva, C., Macroeconomic uncerain, inflaion and growh: Regime-dependen effecs in he G7. Journal of Macroeconomics, 35, Nelson, D. B., Condiional heeroskedasici in asse reurns: A new approach. Economerica, 59, Okun, A., The Mirage of Sead Inflaion. Brookings Papers on Economic Acivi, 2, Pourgerami, A. and Maskus, K.E., The effecs of inflaion on he predicabili of price changes in Lain America: Some esimaes and polic implicaions. World Developmen, 15, 2, Pindck, R., Irreversibili, uncerain and invesmen. Journal of Economic Lieraure, 29, Pinilescu, C., Jemna, D. V., Viorică, E. D. and Asandului, M., Inflaion, Oupu Growh, and Their Uncerainies: Empirical Evidence for a Causal Relaionship from European Emerging Economies. Emerging Markes Finance and Trade, 50, 4, Schwer, G. W., Wh does sock marke volaili change over ime?. The journal of finance, 44,5, Talor, J. B., Esimaion and conrol of a macroeconomic model wih raional expecaions. Economerica: Journal of he Economeric Socie,

19 INVESTIGATION ON THE CAUSAL RELATIONSHIP BETWEEN INFLATION, OUTPUT GROWTH AND THEIR UNCERTAINTIES IN ROMANIA Talor, S., Modelling financial ime series. John Wile & Sons. 57. Thornon, J., The Relaionship beween Inflaion and Inflaion Uncerain in Emerging Marke Economies. Souhern Economic Journal, 73, Ungar, M. and Zilberfarb, B., Inflaion and is unpredicabili: heor and empirical evidence. Journal of Mone, Credi and Banking, 25, Viorica, D., Jemna, D., Pinilescu, C., and Asandului, M., The Relaionship beween Inflaion and Inflaion Uncerain. Empirical Evidence for he Newes EU Counries. PloS one, 9, 3, e Võrk, A., Inflaion Uncerain and is Impac on Economic Acivi in Esonia. Transformaion of Economic and Poliical Ssems in he Balic Sea Region II, Seleced Suden Papers,