Measuring Credibility of Monetary Policy of the Central Bank of Azerbaijan Republic. Introduction

Transcription

1 1 Measuring Credibiliy of Moneary Policy of he Cenral Bank of Azerbaijan Republic Inroducion There is a general recogniion in academic and empirical lieraure ha he primary goal for moneary policy is price sabiliy 1. Oher objecives such as promoing growh and financial sabiliy are considered as secondary o his. The raionale is ha he bes way ha a cenral bank can promoe growh and financial sabiliy is by keeping inflaion low and sable. However, lack of credibiliy of moneary policy is one of he major challenges for Cenral Banks o keep inflaion low and sable in order o mainain financial sabiliy and promoe high growh. In ha conex, credibiliy becomes a cenral issue in moneary policy. In Azerbaijan, he primary goal of he Cenral Bank is o ensure he price sabiliy, while he financial sabiliy is considered as a key responsibiliy since he Cenral Bank is also responsible for prudenial regulaion. Moreover, he poliical economy conex in Azerbaijan makes i much harder for he Cenral Bank o be insulaed from oher objecives, including he promoion of oupu and employmen growh. Thus, he moneary policy of he Cenral Bank in Azerbaijan (hereinafer CBAR) is seen as keeping inflaion low and sable, mainaining financial sabiliy and promoing growh. Inflaionary process in Azerbaijan depends on many facors. In paricular, hey have been affeced by high consumer demand, foreign prices, moneary expansion as a resul of oil boom and herefore expansionary fiscal policy (IMF Counry Repor, 2013). As i is known, Azerbaijan is an oil exporing counry in ha he oil is he major facor fueling he economic growh. In he condiion of high economic growh caused by oil boom, inflaion began o increase sharply since In such an environmen, abiliy of he Cenral Bank in fighing agains high inflaion is limied, bu his policy auhoriy wih a goal of price sabiliy has o implemen relevan measures in curbing high price increases (Hasanov, 2011). Despie he global financial crisis decreased inflaion o a considerable exen since he end of 2008, inflaion began o rise again from he beginning of 2010 and hus become he primary objecive of he CBAR again. Moreover, since Azerbaijani economy faces 1 Bernanke, B., T. Laubach, F. Mishkin and A. Posen, Inflaion Targeing: Lessons from Inernaional Experience, Princeon, NJ: Princeon Universiy Press, 1999.

2 2 susainable appreciaion of exchange rae, which undermines compeiiveness of non-oil secor of economy, CBAR adops a fixed exchange rae regime for a long ime o preven nominal appreciaion of he naional currency in order o preserve he inernaional compeiiveness of he radable non-oil secor. The purpose of his research projec is o analyze he policy credibiliy of he moneary policy credibiliy of CBAR. For his purpose, we apply Kalman filer echniques following Agenor and Taylor (1993) mehodology o analyze analyzed wheher fixed exchange rae regime has affeced inflaion expecaions and acual inflaion raes for he period in Azerbaijan. The key idea of heir analysis was o observe credibiliy of moneary policy in inflaion persisence dynamics. Inflaionary processes in Azerbaijan have been affeced by high consumer demand, foreign prices, and expansionary fiscal policy as a resul of oil boom. Since he poliical economy conex in Azerbaijan direcs he Cenral Bank o pursue he promoion of oupu, i is hard for CBAR o precommi o price sabiliy objecive. In his manner, he research projec aemps o answer he quesion of wha is he moneary policy credibiliy of CBAR, ha is, how inflaion expecaions reac o he moneary policy of CBAR. Here, credibiliy is defined as he exen o which beliefs abou he curren and fuure course of economic policy are consisen wih he announced program by policymakers. I becomes clear ha, his research projec can be imporan research documen for CBAR o implemen opimal and adequae policy decisions o eliminae credibiliy shocks in order o aain is primary goal. Hypohesis The null hypohesis: sabilizaion policy of CBAR is consisen wih is declared objecives. The alernaive hypohesis: CBAR policies are heavily influenced by is discreionary poliical objecives. The economeric model will be used in order o es hese hypoheses. Lieraure Review I is generally recognized ha he source of credibiliy lies in he ime inconsisency of opimal policies (Kydland and Prseco 1977). Time-inconsisency in moneary policy arises from a policymaker s fuile aemp o simulae oupu above he naural level, resuling in an inflaionary bias under discreionary moneary policy. The sandard explanaion of his inflaionary bias is based

3 3 on a wo-way ineracion beween policymakers and a raional public wihin he conex of expecaions-augmened Phillips curve in a game-heoreic approach. Under discreion, policymakers ry o creae inflaion surprises in order o push employmen above is naural level owards he higher desired level. Bu individuals undersand he empaion of policymakers and correcly forecas inflaion, neuralizing any effec of inflaion on employmen. As a consequence, employmen remains a is naural level bu inflaion bias arises. This is a dynamic inconsisency of opimal moneary policy under discreion (Kydland and Prseco 1977, Barro and Gordon 1983). A large lieraure developed o examine alernaive soluions o he inflaionary bias of imeinconsisen moneary policy under discreion (Rogoff (1985), Canzonery(1985), Flood & Isard (1989), Lohmann (1992), Walsh (1995), McCallum(1995,1997), Svensson (1997)). Rogoff (1985) analyzes he preferences of he cenral bank o solve he ime-inconsisency problem and suggesed ha he governmen should appoin a conservaive cenral banker who places greaer relaive weigh on he inflaion objecive han does sociey (he governmen) as a whole. In oher words, he governmen should delegae responsibiliy for moneary policy o an independen cenral bank. Independen cenral bank is able o se policy o minimize is own assessmen of social coss. Thus, he inflaion bias problem is solved hrough delegaion. Lohmann (1992) showed ha he governmen can do even beer if i appoins a weighconservaive cenral banker bu limis he cenral bank s independence. Rogoff s soluion highlighs a rade-off; cenral bank can reduce he inflaion bias bu only a he cos of disoring sabilizaion policy. According o he Lohmann (1992), if he aggregae supply shock urns ou o be oo large, he cenral banker should respond more acively o large shocks in order o reduce he deadweigh loss induced by his disorion. Bu in normal imes, he moneary auhoriy should follow he simple zero-inflaion rule. Walsh (1995) saes ha, he mos convenien way o deermine an opimal incenive srucure is o assume ha he governmen can offer he head of he cenral bank a sae-coningen wage conrac. Such a conrac allows one o derive explicily he manner in which he bank s incenives should depend on he sae of he economy, namely oupu flucuaions and inflaion variabiliy. Presening he cenral bank wih his incenive conrac achieves he dual objecives of eliminaing he inflaionary bias while sill ensuring opimal sabilizaion policy in response o he cenral bank s privae informaion abou he aggregae supply shock. Thus, a sae-coningen wage

4 4 conrac for he cenral banker allows cenral bank o eliminae he inflaionary bias while ensuring opimal sabilizaion policy. Neverheless, here are some oher reasons ha can cause credibiliy problem of policy auhoriy oher han ime inconsisency problem. If policymakers arge he normal level of employmen, an inflaionary bias arises if hey are uncerain abou economic condiions and are more sensiive o employmen below han above normal level. This view implies a posiive associaion beween inflaion and he variance of oupu shocks (Cukierman and Gerlach 2003). According o Cukierman and Gerlach (2003), in he absence of a Kydland Presco-Barro Gordon (hereinafer KPBG) ype inflaion bias, he use of moneary policy o sabilize shocks o he naural level of employmen may lead o an inflaion bias even if policymakers are saisfied wih he poenial level of employmen so ha he KPBG inflaion bias is non-exisen. Some uncerainy abou he fuure sae of he economy and asymmeric concerns abou posiive and negaive oupu gaps combine o creae an inflaion bias. This resul obains in spie of he fac ha he cenral bank s desired level of economic aciviy is equal o poenial oupu or normal employmen (Cukierman and Gerlach 2003:543). Tambakis (2004) saes ha, when he shor-erm Phillips curve is nonlinear and convex, he KPBG inflaion bias is posiive even when moneary policy credibly precommis o arge he naural rae, he cenral bank s preferences are quadraic and symmeric, and here is no fuure uncerainy. In a sochasic convex economy, average unemploymen is always above he naural rae and is increasing in inflaion variabiliy. I follows ha average inflaion and unemploymen move ogeher, he mean inflaion rae is higher when inflaion is more volaile. As a resul, credibiliy problem can arise no only from ime-inconsisency of opimal plans bu also from oher reasons, such as uncerainy abou he fuure sae of he economy even if cenral bank s desired level of economic aciviy is equal o poenial oupu and he feaure of shor-erm convex Phillips curve in he economy. Kremers (1990) analyzes Ireland s inflaion experience afer is paricipaion in European Moneary Sysem (EMS) and concludes ha he paricipaion in a fixed bu adjusable exchange rae sysem wih a group of low-inflaion counries caused inflaion expecaions o bring down in Ireland. According o he paper, a semi-fixed exchange rae policy may provide a source of discipline enhancing credibiliy of disinflaion and reduces is derimenal impac (i.e. oupu loss) on he economy.

5 5 Rodriguez F., M. Ibanez, J. Rodriguez and S. Rivero (2008), provide an overview of several sudies abou he credibiliy of hree of he weakes currencies ha paricipaed in he EMS (he Spanish pesea, he Poruguese escudo, and he Irish pound), as well as o presen a join analysis of credibiliy of weak and hard currencies. According o he paper, he Exchange Rae Mechanism (ERM) faciliaed he necessary process of disinflaion and ha i raised he coss of inflaion in he EMS. The aim of ERM was o provide discipline o macroeconomic policies ha would ensure ha fiscal deficis would no be compounded by moneary expansions. In hese maers, Germany emerged almos naurally as he anchor for exchange rae and inflaion expecaions. In fac, he ERM could be seen as an insiuional arrangemen which has enabled he member counries o borrow he repuaion of he Bundesbank by pegging heir exchange raes o he Deuschmark. In his way, weaker currencies achieved a greaer degree of credibiliy due o he exisence of repuaion effecs wihin he ERM. In oher words, EU member saes wih hisories of high inflaion used he EMS as a way of imporing he Bundesbank s ani-inflaionary credibiliy. Abiyev (2008) analyzes disinflaion credibiliy of The Cenral Bank of Turkey. In his sudy, i is sressed ha here are some reasons can cause credibiliy problem oher han ime inconsisency. On he oher hand, opimal policy rule is defined as a policy rule which minimizes long-run inflaion-oupu variabiliy. In he sudy, he model of assessing credibiliy is based on wo assumpions: firs, inflaion is an inerial process. Inerial inflaion is inversely relaed o he policy credibiliy. Second, a policy regime change affecs oupu and inflaion. Thus oupu-price relaion would change when a policy regime change occurred. Wih he use of mulivariae GARCH, he volailiy rade-off beween inflaion and growh shocks, ime-varying inflaion-oupu correlaion coefficiens, inflaion uncerainy and oupu volailiy esimaed and hus he effec of disinflaion policies on ime-varying inflaion-oupu correlaion coefficiens, inflaion uncerainy and oupu volailiy had been evaluaed. I is shown hen inflaion argeing has increased he credibiliy of The Cenral Bank of Turkey o reduce inflaion o he low levels. Increased credibiliy eliminaed he inerial componen in he public s inflaion expecaions and hus conribued o he real side of he economy. Mehodology The effeciveness of sabilizaion program depends on he role of credibiliy facors. When policies implemened lack credibiliy, privae agens would evenually recognize ha he governmen effors o mainain a consisen se of policies over ime would no be successful and he

6 6 announced disinflaion program would no be susainable in he long-run. On he oher hand, when he public is confiden in he abiliy of policymakers o carry ou a newly announced sabilizaion program, he inflaionary process will be reduced by changing inflaionary expecaions. Since inflaionary expecaions have a significan effec on curren wage and price decisions, a reducion in he acual inflaion may resul. This model was applied by Agenor and Taylor (1993) o sudy he credibiliy effec of he Cruzado Plan implemened in Brazil in 1986 ha consised of a general freeze of prices, he ariffs for public services, and exchange rae. Moreover, he governmen inervenion removed wages indexaion. The procedure followed by he auhors was based on he exisence of he official foreign exchange rae and he black marke or parallel exchange rae marke in Brazil. The auhors consideraions were based on he exchange rae markes. They consider ha if a disinflaionary program loses credibiliy, agens anicipae he acceleraion in he inflaion rae. Considering fundamenal differences beween he Brazilian economic siuaion and he Azerbaijan economic siuaion, he model had o be modified in order o capure he dynamics of he Azerbaijan economy under invesigaion. This paper ess he size and he diffusion effec of he credibiliy facors of he fixed exchange rae based sabilizaion program implemened in Azerbaijan since 2000, by means of a wo-sep procedure which focuses on he analysis of he ime-series properies of he parallel marke premium and he inflaion process. In he model ha we apply in his projec we propose a differen approach o measure he credibiliy of sabilizaion policies of CBAR. The model is buil on wo assumpions: The firs assumpion is ha inflaion is, because of inerial forces, a serially correlaed process. Expecaions play a key role in breaking down inflaionary ineria. Moreover, he degree of inflaion persisen is assumed o be inversely relaed o he degree of policy credibiliy The second assumpion is relaed wih he definiion of an appropriae proxy ha is able o measure he degree of credibiliy of a program, or generally an economic policy ha can sabilize he level of inflaion. This proxy is given by ha par of he parallel marke premium, which is orhogonal o movemens in he 'marke fundamenals'. Parallel marke premium is defined as difference of ineres raes is based on he assumpion ha i is highly sensiive o marke expecaions regarding governmen policies, and herefore, is considered a good variable for examining credibiliy effecs. The idea is o consider how he spread given by he change in he wo ineres raes reacs o

7 7 an announced sabilizaion program. Hence, we assume ha he foreign exchange ineres rae (FX ineres rae) is highly sensiive o he change in credibiliy. Since he parallel marke premium is influenced by he privae agens expecaions regarding curren and fuure economic policies, he analysis of is behavior is crucial o he sudy of credibiliy facors which are affeced by changes in expecaions. Hence, we assume ha he FX ineres rae is highly sensiive o he change in credibiliy. The marke premium, ha is he difference beween ineres rae in domesic currency deposis and ineres rae on FX deposis, is an endogenous variable. This means ha is level will also reflec he behavior of marke fundamenals, such as domesic deb posiion, pas domesic inflaion, foreign prices ec. If movemens in he parallel premium canno be explained by movemens in he marke fundamenals, such as domesic deb posiion, pas domesic inflaion, foreign prices ec., hen he remaining variaion mus be due o variaion in he perceived degree of policy credibiliy. In oher words, all he variaions in ha par of he parallel marke will be srongly influenced by he level of repuaion and credibiliy gained by he auhoriies, and ha par can be used as an index of credibiliy. The model can be divided ino wo pars. The firs par consiss of decomposing he parallel marke premium ( ) ino wo componens: he 'fundamenal componen' and 'non fundamenal componen' which will be he basis for a measure of credibiliy. The fundamenal componen of parallel marke premium reflecs he behavior of marke fundamenals, such as budge expendiure, pas domesic inflaion, foreign prices ec. Non fundamenal componen of parallel marke premium, which is orhogonal o movemens in he fundamenal componen, will be heavily influenced by percepions of policy credibiliy so ha i can be used as an index of credibiliy. If we denoe parallel marke premium ( ) and predeermined fundamenal facors ( z ), he general form of he model is as follows: a( L) b( L) z u (1) a( L) 1 a k 1 L a2l a3l a k L (2) b( L) b k 1 L b2 L b3l b k L (3) Where a represens a scalar coefficien, he b denoes conformable coefficien vecors, L is he lag operaor, and u he residual process. Afer rearranging we ge:

8 8 a ak k b1 z 1... bk zk u (4) Equaion (4) can be inerpreed as he firs equaion in a vecor auoregressive sysem for parallel marke premium. The residuals from he equaion (4) we derive our credibiliy variable ( c ). I is assumed o be he complemen of he error erm u : c (5) u In he second par of he model, using a Kalman filer approach, a backward-looking process for inflaion wih parameers varying wih non-fundamenal componen of he parallel marke premium are esimaed. Harvey (1989) argues ha he key o handling srucural ime series models is he sae space form, wih he sae of he sysem represening he various unobserved componens such as rends and seasonal. Once in he sae space form, he Kalman filer provides he means of updaing he sae as new observaions become available. Assuming ha inflaion,, is given by a AR(1) process, he sysem o be esimaed is given by 1, (6) 1 (7) c Where (6) is he measuremen equaion wih ransiion equaion. and are ( nx 1) vecors and (7) is he and ( mx 1) are vecors and is a c (mx1) marix. I is assumed ha all errors are normally disribued and E( )=0 for all. Equaions (6) and (7) represen a sae space form. Finally, following he heory of Agenor and Taylor (1993); he Kalman filer recursions can hen be applied o yield opimal esimae+es of he sae variable sequence. The resuling esimae of should be negaive: he higher credibiliy is, he lower he inerial effec on inflaion. The coefficien of he equaion (7) should be smaller afer he accomplishmen of a credible disinflaion program. In he analysis he following daa will be used: CPI inflaion in Azerbaijan, exchange rae USD/AZN, GDP, shor-erm and long-erm deposi ineres-rae differenials and inflaion differenial (comparison wih USA). The period is , eiher monhly or quarerly.

9 9 References Abiyev V. Analysing The Credibiliy Problem of Disinflaion Policies in Turkey, PhD Disseraion, Universiy of Haceepe, Ankara, 2008 Abiyev V. & Funda Telaar, Dynamic Analysis of he Trade-off beween Inflaion Variabiliy and Oupu Volailiy in he Turkish Economy, TISK AKADEMI, 5(9), pp:83-111, 2010 Agenor, P. R. & M. P. Taylor, Analyzing Credibiliy in High Inflaion Counries: A New Approach, The Economic Journal, 103, , 1993 Backus, D. & J. Driffil, Inflaion and Repuaion, American Economic Review, 75, , 1985 Ball, L. & S. Cecchei, Inflaion and Uncerainy a Shor and Long Horizons, Brookings Papers on Economic Aciviy, 1990(1), , 1990 Barro R. J. & D. B. Gordon, Rules, Discreion and Repuaion in a Model of Moneary Policy, Journal of Moneary Economics, 12, , 1983 Blackburn, K. & M. Chrisensen, Moneary Policy and Policy Credibiliy, Jounal of Economic Lieraure, 27, March, 1 45, 1989 Blinder, Alan S., Cenral Banking in Theory and Pracice, Cambridge, MA, MIT Press, 1998 Canzonery M., Moneary Policy Games and he Role of Privae Informaion, American Economic Review, 75, , 1985 Cukierman, Alex & Sefan Gerlach, The Inflaion Bias Revisied: Theory and Some Inernaional Evidence, The Mancheser School, 71(5), , 2003 Debelle, Guy & Sanley Fischer, How Independen Should A Cenral Bank Be?, CEPR Publicaion No: 392, 1994 Fischer, Sanley, Rules versus Discreion in Moneary Policy, B.M. Friedman and F.H. Hahn (Eds.), Handbook of Moneary Economics, New York: Elsevier Science Publishers B.V., , 1990 Flood, P. R. & P. Isard, Moneary Policy Sraegies, IMF Saff Papers, 36(3), , 1989 Hasanov F., Analyzing price level in a booming economy: he case of Azerbaijan, EERC working Paper Series, Working paper No:11/02E, 2011

10 10 Kremers Jeroen J. M., "Gaining Policy Credibiliy for a Disinflaion: Ireland's Experience in he EMS", IMF Saff Paper No 37, 1990 Kydland, F. & E. Presco, Rules Raher Than Discreion: The Inconsisency of Opimal Plans, Journal of Poliical Economy, 85, , 1977 Lohmann, S., Opimal Commimen in Moneary Policy: Credibiliy versus Flexibiliy, American Economic Review, 82, , 1992 McCallum, Bene T., The Presen and Fuure of Moneary Policy Rules, NBER Working Paper No:7916, 2000 McCallum, Bene T., Crucial Issues Concerning Cenral Bank Independence, Journal of Moneary Economics, 39(1), , 1997 Persson, T. & G. Tabellini, Moneary and Fiscal Policy, Cambridge: The MIT Press, 1994 Rodriguez, F., Ibanez, M., Rodriguez J. and Rivero S., The Credibiliy of European Moneary Sysem: A Review, Cuadernos de Economia. Vol. 31, Num. 86, mayo-agoso, 2008 Rogoff, Kenneh, The Opimal Degree of Commimen o an Inermediae Moneary Targe, Quarerly Journal of Economics, 100, , 1985 Tambakis D. N., Inflaion Bias wih a Convex Shor-Run Phillips Curve and No Time- Inconsisency, Pembroke College, Cambridge and Cambridge Endowmen for Research in Finance, 2004 Taylor, J. B. Discreion versus Policy Rules in Pracice, Carnegie-Rocheser Conferences Series on Public Policy, 39, , 1993 Walsh, C., Opimal Conracs for Cenral Bankers, American Economic Review, 85, , 1995

11 Auhor Fuad Mammadov. Currenly I work as an economic analys a Academy for Invesigaion and Applicaion of Invenions and Discoveries. I pursued Maser of Ars degree in Policy Economics in he Williams College in Williamsown MA, USA from 2012 o June, I used o work a he Insiue for Scienific Research on Economic Reforms, he Minisry of Economic Developmen of Azerbaijan as senior economis a Macroeconomic Research and Susainable Developmen Deparmens during I have also eaching experience on macroeconomics and inernaional rade in Azerbaijan universiies. I have aended a number of courses, seminars and workshops on economerics, which ook place in abroad, and Azerbaijan. My main research fields are macroeconomics and moneary policy. Appendix 1 Table A2.1: Iems of he projec budge Price in USD Books Moneary Theory and Policy (2010) by Carl Walsh Moneary Policy Sraegy (2009) by Frederic Mishkin Forecasing, Srucural Time Series Models and he Kalman Filer by Andrew C. Harvey Cenral Bank Independence, Targes and Credibiliy: Poliical and Economic Aspecs of Delegaion Arrangemens for Moneary Policy By Francesco Lippi Cenral Bank Sraegy, Credibiliy, and Independence: Theory and Evidence by Alex Cukierman Macroeconomics 6h Ediion, by Olivier Blanchard Time Series Analysis by Sae Space Mehods by J.Durbin Time Series Analysis by James Hamilon Compuer and Sofware Microsof Office Professional 2013 Produc Key Card - 1PC/1User EViews Enerprise Ediion 8 (CD & Manuals) Hardware (lapop) Wages Fuad Mammadov Toal Price of his book from hp:// Walsh/dp/ /ref=sr_1_3?ie=UTF8&qid= &sr=8-3&keywords=moneary+policy, $ $ Price of his book from hp:// ($ )

12 12 3 Price of his book hp:// Kalman/dp/ /ref=sr_1_1?ie=UTF8&qid= &sr=8-1&keywords=kalman+filer+economics, $ $ Price of his book from hp:// Credibiliy/dp/ /ref=sr_1_7?ie=UTF8&qid= &sr=8-7&keywords=cenral+bank+independence, $ $ Price of his book from hp:// Independence/dp/ /ref=sr_1_8?ie=UTF8&qid= &sr=8-8&keywords=cenral+bank+independence, $36.00+$ Price of his book from hp:// Blanchard/dp/ /ref=sr_1_fkmr0_2?s=books&ie=UTF8&qid= &sr=1-2- fkmr0&keywords=macroeconomics+updaed+blanchard , $ $ Price of his book from hp:// Science/dp/ X/ref=sr_1_2?ie=UTF8&qid= &sr=8-2&keywords=sae+space, $ $ Price of his book from hp:// Hamilon/dp/ /ref=sr_1_2?ie=UTF8&qid= &sr=8-2&keywords=ime+series+economerics $ Price of Microsof Office from hp://office.microsof.com/en-us/home-and-business/ 10 Academic Price of EViews Enerprise Ediion 8 from hp:// 11 Price of Lapop from hp://compuex.ws/?l=produc&idca=43&prod= Fuad Mammadov s wage over he projec period of 9 monhs. I is calculaed as: 30 USD per hour x 33 hours in a monh x 9 monhs Table A2.2: Research imeable 1 Lieraure review 2 monhs 2 Daa collecion and processing 2 monhs 3 Model esimaion and analyses 3 monhs 4 Working ou of he paper 2 monhs