International Monetary Fund Washington, D.C.

Transcription

1 Inernaional Moneary Fund February IMF Counry Repor No. /5 Nigeria: Seleced Issues This Seleced Issues paper for Nigeria was prepared by a saff eam of he Inernaional Moneary Fund as background documenaion for he periodic consulaion wih he member counry. I is based on he informaion available a he ime i was compleed on January,. The views expressed in his documen are hose of he saff eam and do no necessarily reflec he views of he governmen of Nigeria or he Execuive Board of he IMF. The policy of publicaion of saff repors and oher documens by he IMF allows for he deleion of marke-sensiive informaion. To assis he IMF in evaluaing he publicaion policy, reader commens are invied and may be sen by o publicaionpolicy@imf.org. Copies of his repor are available o he public from Inernaional Moneary Fund Publicaion Services 7 19h Sree, N.W. Washingon, D.C. 31 Telephone: () 3 73 Telefax: () publicaions@imf.org Inerne: hp:// Price: $1. a copy Inernaional Moneary Fund Washingon, D.C.

2 INTERNATIONAL MONETARY FUND NIGERIA Seleced Issues Prepared by Chad Seinberg (AFR) and Sephen Swaray and Jennifer Moyo (boh MCM) Approved by he African Deparmen January, Conens Page I. Nigerian Inflaion: A Srucural Model for Policy Analysis...3 A. The Model... B. Assumpions on Equilibrium Values... C. Baseline Forecas...7 D. Risk Analysis... E. Conclusions...9 References...17 II. Banking Secor Sabiliy Following Consolidaion...3 A. Recen Developmens... B. Bank Soundness... C. Sress Tess...3 D. Emerging Risks and Vulnerabiliies...33 E. Banking Supervision...3 F. Conclusions...3 Tables II.1. Bank Groups... II.. Porfolio Qualiy by Group... II.3. Financial Soundness Indicaors for he Banking Secor, II.. Comparaor Counries: Financial Soundness Indicaors, II.5. Sensiiviy o Credi Risk...3 II.. Sensiiviy o Ineres Rae Risk...3 Figures I.1. Model Diagram... I.. Equilibrium Values...11 I.3. Baseline Scenario...13 I.. Fiscal Expansion Scenario...1

3 I.5. Naira Appreciaion Scenario...15 I.. High Food Price Scenario...1 Boxes II.1. Sress Tess: Mehodology and Resuls...31 Appendices I.A1. Complee Model Equaions...1 I.A. Calibraion...

4 3 I. NIGERIAN INFLATION: A STRUCTURAL MODEL FOR POLICY ANALYSIS 1 1. Macroeconomic condiions in Nigeria have improved in recen years. Inflaion has fallen o single digis. An oil-price-based fiscal rule has helped conain governmen spending o levels consisen wih macroeconomic sabiliy. Financial markes have developed; he mauriy srucure of he domesic bond marke is longer and he size and imporance of he inerbank foreign exchange marke have increased, resuling in a more flexible exchange rae. These developmens augur well for he implemenaion of moneary policy.. A he same ime, money demand has become increasingly unpredicable. Financial marke developmen and rapid financial deepening are affecing he relaionship beween money and prices. Meanwhile, increased foreign invesor ineres in Nigerian asses has spurred capial inflows. These developmens will likely coninue, furher complicaing forecass of money demand. 3. Agains his backdrop, he Cenral Bank of Nigeria (CBN) has announced ha i inends o adop inflaion argeing (IT). The auhoriies are working o pu he righ condiions in place before implemening he new regime. In examining experience wih IT, he Fund s board recommended he following condiions be in place before IT is implemened: (i) arge credibiliy (including cenral bank independence, arge specificaion, and a communicaion sraegy); (ii) a reasonably good undersanding of he inflaion process a he cenral bank; (iii) considerable exchange rae flexibiliy; and (iv) he absence of fiscal dominance.. This chaper describes a simple forecasing and policy analysis sysem (FPAS) for forecasing inflaion and assessing economic risks in Nigeria. The model is designed o suppor policy analysis in an inflaion-argeing regime, which principally aims o provide anchors for inflaion and inflaion expecaions. I should help inform he CBN s view of he inflaion process and help he bank srucure is communicaion sraegy. 5. This chaper is divided ino five secions. Secion A explains he model s srucure. Secion B discusses he key assumpions underlying he esimaed equilibrium values for inflaion, oupu, ineres raes, and he exchange rae. Secion C discusses Nigeria s curren policy mix and describes he model s baseline forecas. Secion D analyzes risks o he baseline forecas and illusraive examples of he impac of a more expansionary fiscal sance, a real appreciaion of he naira, and higher food prices. Secion E concludes. 1 Prepared by Chad Seinberg. This model calibraes and modifies a model developed by Berg, Karam, and Laxon (a, b), which has now been adoped for over counries wihin he IMF.

5 A. The Model. The model capures he link beween he policy insrumen (a shor-erm ineres rae) and he nexus of oupu, inflaion, and he exchange rae in a small-open economy. The model expresses key variables as deviaions-from-equilibrium, and does no ry o explain he underlying real equilibrium values. However, he model can address many policy issues ha arise rouinely in making decisions abou moneary policy acions and communicaing he reasons o he public. 7. The model provides a helpful organizing framework for esablishing near- and medium-erm baseline forecass. This approach encourages a more-srucured and ransparen discussion of curren policy. In generaing alernaive scenarios, he model is a valuable ool for assessing key risks o he baseline forecas and in analyzing he relaive imporance of various assumpions.. There are four core equaions: (i) an aggregae demand equaion (IS curve) ha relaes he level of real aciviy o expeced and pas real aciviy, he real ineres rae, he real exchange rae, and he level of foreign aciviy; (ii) a price-seing equaion (Phillips curve) ha relaes inflaion o pas and expeced inflaion, he oupu gap, he exchange rae and he relaive price of oil; (iii) an uncovered ineres pariy condiion for he exchange rae, wih some allowance for backward-looking expecaions; and (iv) a rule for seing he policy ineres rae as a funcion of he oupu gap and expeced inflaion. Appendix I.A1 provides he full lis of model equaions. Appendix I.A provides deails on calibraion. 9. Because Nigeria is an open economy, he model considers he impac of foreign shocks on he domesic environmen using he Unied Saes as a proxy for he world economy. 3 The equaions for he U.S. economy have a similar srucure o he Nigerian equaions bu exclude he open-economy linkages.. The moneary ransmission mechanism of he model is depiced in Figure I.1. The policy rae affecs he economy hrough an inernal and exernal channel. The inernal channel is he long run ineres rae, which firs affecs aggregae demand, hen oupu, and finally inflaion. The exernal channel is he exchange rae, which impacs inflaion direcly (hrough pass-hrough) and aggregae demand indirecly hrough is impac on foreign demand. Inflaion expecaions are influenced by he inflaion oucome and vice versa. The srucure of he economy deermines he srengh of he differen arrows in Figure I The model may no adequaely capure he srucure of Nigeria s economy in wo areas. Firs, weaher-relaed supply shocks heavily influence he CPI in Nigeria bu are no 3 The Unied Saes has he larges links o Nigeria s economy. However, if we exclude oil receips, by some measures Europe s links o he Nigerian economy are even sronger. Fuure versions of his model may hus use Europe as a proxy for he world s economy.

6 5 capured in he model. Insead, he impac of weaher on food prices is capured in he error erm, making his a key forecasing variable when rainfall deviaes from normal condiions. Second, he srengh of he ineres rae channel of moneary policy, while evolving, is sill relaively weak. There are sill difficulies in managing he overnigh inerbank ineres rae, and he link beween he moneary policy rae and lending and savings is no well defined. Over ime, however, as he financial sysem develops and global capial flows increase, his channel of moneary policy should srenghen. 1. In line wih modeling pracices in many policymaking insiuions, he parameer values in he model equaions were derived based on economic principles, he economeric evidence available, and an undersanding of how he economy funcions. The raionale for he coefficien choices are in he descripions of he four key equaions ha follow. (The appendices provide addiional deails). Aggregae demand equaion 13. Domesic oupu depends on he real ineres rae, he real exchange rae, and demand in he res of he world (represened by he Unied Saes). Dynamics are added hrough pas and fuure domesic oupu gaps. Parameers on lead and lag values of he oupu gap are se such ha he shocks ake wo years o work hrough he economy, based on he norm in emerging markes. The real exchange rae is assumed o have a larger impac on oupu han real ineres raes, because he ineres rae channel is sill evolving. U.S. oupu is assumed o have a low impac on domesic demand since U.S. business cycles ypically have lile relaion o condiions in Nigeria. Phillips curve equaion 1. Inflaion depends on expeced and lagged inflaion, he oupu gap, he exchange rae gap, and movemens in he real price of oil. Parameers on lead and lag values of inflaion are se such ha moneary policy has a low level of credibiliy, i.e., a high degree of inflaion ineria is assumed, since economic agens ypically view pas inflaion as a beer predicor of oucomes han he officially announced inflaion arge. Therefore, he parameer value is se such ha only large increases in he policy rae would impac curren inflaion. Exchange rae pass-hrough is assumed o be high, bu he level of impors in he CPI is relaively low. Because wage conracs and indexaion may be less prevalen in economies ha are growing rapidly, i is assumed ha inflaion responds o excess demand wihou significan lags.

7 Uncovered ineres pariy 15. The exchange rae and ineres raes are linked by an ineres-pariy condiion. This saes ha, wih inernaional capial mobiliy, he expeced reurn on a shor-erm invesmen should be he same a home as abroad, ignoring counry-risk premiums. If he exchange rae is expeced o depreciae, herefore, an invesor will demand o be compensaed in he form of a higher domesic ineres rae. For forward-looking exchange-rae expecaions, he model permis, bu does no impose, model-consisen ( raional ) expecaions. 1. The basic FPAS model has been modified o reflec he curren rend real appreciaion in Nigeria. In such an environmen, in addiion o he las period s value, agens add in he observed rae of rend appreciaion o forecas he value of he currency in he nex period. Moneary policy reacion funcion 17. The moneary policy reacion funcion mirrors he specificaion in many similar models. The policy insrumen is a shor-erm nominal ineres rae; he cenral bank ses i o anchor inflaion o a arge rae, π*. The cenral bank may also adjus is acions when oupu deviaes from equilibrium. Here i is assumed ha (i) given he low level of policy credibiliy, he moneary sance is aggressive when inflaion deviaes from he arge; (ii) he oupu gap is less imporan han inflaion in policy decisions; and (iii) he CBN prefers o smooh changes in he ineres rae. B. Assumpions on Equilibrium Values 1. The resuls of he model depend on he assumed equilibrium values. The model expresses each variable in erms of is deviaion from equilibrium, bu does no aemp o explain movemens in equilibrium real oupu, he real exchange rae, he real ineres rae, or in he inflaion arge. Esimaing hese equilibrium values is likely more imporan han geing he model s parameers accurae, as discussed in he previous secion. A key quesion, herefore, is how o choose hese values so ha hey accuraely reflec he Nigerian economy. Figure I. provides iniial esimaes; below are he key assumpions. Inflaion arge. I is assumed ha he moneary auhoriies have been seing he ineres rae according o a crierion ha is similar o inflaion argeing. The inflaion In several ransiion economy versions of his model, he rend appreciaion is also subraced from he Phillips curve equaion such ha in he seady sae here is a posiive rae of real appreciaion (e.g., Tiffin (7)). This would also imply ha he rend real appreciaion does no reduce inflaion. Fuure versions of his model may consider his exension.

8 7 arge is assumed o be given by he rend of acual inflaion unil end-5. 5 Since 5 and he sar of Nigeria s reform program suppored by he Policy Suppor Insrumen, he arge is assumed o be.5 percen, he higher end of he auhoriies announced arge of single digis. Currenly, Nigeria s inflaion rae is below is inflaion arge a end-7. Equilibrium real ineres rae. The equilibrium real ineres rae is assumed o be 5 percen, which implies a risk premium of abou 3 basis poins. While lower han he risk premium of Nigerian sovereign bonds raded in (before he London Club buyback), his is broadly consisen wih issues of African sovereign deb in 7. The acual real ineres rae has flucuaed widely over he years, wih he curren level below he assumed equilibrium level. Equilibrium real exchange rae. Saff esimaes indicae he naira could be undervalued by up o 15 percen over he medium erm. I is unlikely ha his gap is adding o aggregae demand hrough he exernal secor. To close his gap, herefore, we assume a rend real appreciaion in he equilibrium value hrough ; afer ha, he equilibrium value is assumed o be consan. The hisorical equilibrium value is given by he rend value as calculaed using he LRX filer. Equilibrium non-oil oupu. The equilibrium value is calculaed using an HP filer on he log level of non-oil oupu. Poenial non-oil oupu growh is percen. This variable is he mos difficul o esimae because daa sources are weak: GDP is calculaed annually, and he growh series is relaively uniform. There are also few oher measures of economic aciviy wih high frequency. Using he HP filer, Nigeria is assumed o be marginally above poenial. The hisorical values show minimal deviaions from rend, wih he oupu gap never above 1 percenage poin. Swings in he gap in 3 relae o changes in agriculural oupu and are correlaed wih rends in he headline inflaion rae. C. Baseline Forecas 19. The model is helpful in preparing baseline forecass. The model is a relaively simple and racable represenaion of he economy; as such, he resuls are bes used o frame discussions abou he baseline forecas and o help shape medium- and long-erm forecass, where issues of consisency and dynamics are resolved as economic fundamenals converge o he esimaed equilibrium pah. 5 This is calculaed using he Laxon-Rose-Xie (LRX) filer (a more general version of he univariae Hodrick- Presco (HP) filer), which is described in deail in Appendix IV of Berg, Karam, and Laxon (b).

9 . The baseline forecas is a combinaion of judgmen in he near erm and a pure model forecas in he medium-erm. 7 forecass are judgmen. forecass are a combinaion of judgmen and model forecas, as model residuals are gradually phased ou over he four quarers. Forecass beyond 9 are a pure model forecas. 1. Nigeria s economy is growing robusly, aided by low ineres raes (Figure I.3). Oupu is above poenial, suppored by srong fiscal demand and lower ineres raes, which are simulaing privae secor credi growh. Shor-erm ineres raes are below he neural rae as Naira appreciaion pressures coninue and inflaion remains well below arge (aided by posiive weaher condiions ha helped lower he headline rae in and 7). Moneary condiions will add o aggregae demand as he oupu gap is projeced o average 1. and 3.7 percenage poins in and 9, respecively.. Noneheless, raes in he medium erm will need o rise o reurn he real ineres rae o is neural rae and conain excess aggregae demand. The rise in ineres raes should cause oupu o dip in he medium erm and push he oupu gap o slighly negaive levels in 11. D. Risk Analysis 3. This secion describes he resuls of shocks o he baseline assumpions. An increase in domesic demand afer a fiscal shock, an exogenous appreciaion of he naira, and a higher food price scenario are considered. The srengh of he model is in evaluaing risks o he forecas, appropriae responses o a variey of shocks, and dependencies of he forecas and policy recommendaions on various assumpions abou he economy (see Figure I.1). Fiscal expansion (Figure I.). An illusraive example of a fiscal expansion in, semming from addiional spending of 5 percenage poins of non-oil GDP a year, is considered. This is implemened hrough an add-on shock o he residual in he aggregae demand equaion. 5. The resuls show ha he addiional spending would likely exceed he economy s absorpive capaciy. Afer iniially acceleraing, growh slows sharply as he cenral bank raises ineres raes o keep inflaion in single digis. Higher ineres raes, in urn, squeeze credi and dampen privae secor aciviy, causing he naira o appreciae. The loss of compeiiveness furher consrains he privae secor. Exogenous appreciaion of he naira (Figure I.5). In his shock, i is assumed he naira appreciaes fifeen percen above rend for wo quarers. The ne effec in he wo quarers is ha he naira appreciaes o a level below N 1 per US$1.

10 9 7. This shock dampens inflaion and lowers foreign demand and radable oupu in he shor run. The policy response is o lower ineres raes and keep heir pah lower han under he baseline, hus muing he shock s impac on aggregae demand. High food prices (Figure I.). In his shock, i is assumed food prices are higher han projeced in he firs hree quarers of. This is implemened hrough an add-on shock o he residual in he Phillips curve equaion. High food prices are assumed o add percenage poins o he headline inflaion rae. 9. To keep headline inflaion wihin is arge range, he moneary auhoriies would need o raise ineres raes more rapidly han under he baseline scenario. The consequences of his policy are reduced oupu (and a smaller oupu gap) and a somewha more appreciaed naira. Over he medium-erm, he growh rae decreases more gradually (compared o he baseline), as he build up in excess demand is less pronounced in he shorrun. E. Conclusions 3. This chaper presens a simple srucural model of inflaion adaped for Nigeria based on he mehodology of Berg, Karam, and Laxon (a, b). The model embodies he principle ha moneary policy in an inflaing argeing framework aims o anchor inflaion and inflaionary expecaions. 31. This approach allows differen policy opions o be considered sysemaically in a baseline forecasing exercise. The chaper analyzes he impac of hree shocks on he baseline forecas a fiscal expansion, an unexpeced appreciaion of he naira, and higher food prices. 3. The developmen and calibraion of his model are ongoing. As we learn more abou he ransmission mechanism in Nigeria, he model will coninue o be refined. Given he imporance of oil in he Nigerian economy, a more deailed reamen of he price of oil on fiscal policy and he real equilibrium exchange rae could be considered. The auhoriies may choose no o respond o his supply shock if core, raher han headline, inflaion is included in he moneary policy reacion funcion.

11 Figure I.1. Model Diagram Inflaion Expecaion Longer Term Ineres Raes Inflaion Aggregae Demand Oupu Gap Policy Rae Exchange Rae Shocks o Inflaion: Indirec Taxes Energy Prices Shocks o Poenial Oupu Shocks o Aggregae Demand Foreign Demand Commodiy Prices Fiscal Policy Financial Shocks: Foreign Ineres Raes Porfolio Shifs Shocks o he Economy Source: Laxon and Sco ()

12 11 Figure I.. Equilibrium Values Inflaion arge Headline Inflaion (Percen) Year-on-year inflaion Real Ineres Rae (Percen) Equilibrium real ineres rae Real ineres rae Real Exchange Rae (Percen) Real exchange rae Equilibrium real exchange rae

13 1 Figure I.. Equilibrium Values (coninued). Non-oil GDP (Log) Log of Equlibrium Real O upu Log of real oupu Oupu Gap (Percen)

14 13 Figure I.3. Baseline Scenario Non-Oil GDP Growh (Percen) Poenial oupu Headline Inflaion (Percen) Inflaion arge Neural rae Nominal Ineres Rae (Percen) Nominal Exchange Rae (Naira/US$)

15 1 Figure I.. Fiscal Expansion Scenario Non-Oil GDP Growh (Percen) Shock growh Baseline growh Headline Inflaion (Percen) Shock inflaion 1 Baseline inflaion Shock ineres rae Nominal Ineres Rae (Percen) Baseline ineres rae Shock exchange rae Nominal Exchange Rae (Naira/US$) Baseline exchange rae

16 15 Figure I.5. Naira Appreciaion Scenario Shock growh Non-Oil GDP Growh (Percen) Baseline inflaion Headline Inflaion (Percen) Shock inflaion Baseline ineres rae Nominal Ineres Rae (Percen) Shock ineres rae Nominal Exchange Rae (Naira/US$) Shock exchange rae Baseline exchange rae Baseline growh

17 1 Figure I.. High Food Price Scenario Non-Oil GDP Growh (Percen) Shock growh Baseline growh Shock inflaion Headline Inflaion (Percen) 1 Baseline inflaion Shock ineres rae Nominal Ineres Rae (Percen) Baseline ineres rae Nominal Exchange Rae (Naira/US$) Shock exchange rae Baseline exchange rae

18 17 References Berg, A., P. Karam, and D. Laxon, a, A Pracical Model-Based Approach o Moneary Policy Analysis Overview, IMF Working Paper (Washingon: Inernaional Moneary Fund)., b, Pracical Model-Based Moneary Policy Analysis A How-o Guide, IMF Working Paper (Washingon: Inernaional Moneary Fund). Epsein, N., P. Karam, D. Laxon, and D. Rose,, A Simple Forecasing and Policy Analysis Sysem for Israel: Srucure and Applicaions, in Israel: Seleced Issues, Counry Repor No. /11 (Washingon: Inernaional Moneary Fund). Harjes, T. and L. Ricci, 7, A Quaniaive Analysis of Inflaion Dynamics in Souh Africa, in Souh Africa: Seleced Issues, Counry Repor No. 5/35 (Washingon: Inernaional Moneary Fund). Inernaional Moneary Fund, 1999, Inflaion Targeing Implicaions for IMF Condiionaliy, IMF Policy Paper (Washingon: Inernaional Moneary Fund).,, Inflaion Targeing and he IMF, IMF Policy Paper (Washingon: Inernaional Moneary Fund). Laxon, Douglas and Alasdair Sco,, On Developing a Srucured Forecasing and Policy Analysis Sysem Designed o Suppor Inflaion-Forecas-Targeing, in Inflaion Targeing Experiences: England, Finland, Poland, Mexico, Brazil, Chile, (Ankara: The Cenral Bank of Turkey), pp. -3. Tiffin, A., 7, Modeling Moneary Policy in Romania, in Romania: Seleced Issues, Counry Repor No. 7/ (Washingon: Inernaional Moneary Fund).

19 1 Appendix I.A1. Complee Model Equaions The following are he equaions used for he Nigerian economy. The equaions for he U.S. economy have a similar srucure bu exclude he open-economy linkages. Behavioral equaions π ygap = β ygap + β ygap β RRgap + β zgap + β ygap + ε US ygap ld + 1 lag 1 RRgap 1 zgap 1 USygap π [ z z ] + α π + α π ε = α πldπ + + (1 α πld ) π 1 + α ygap 1 + α z 1 rpoil, 1 rpoil, 1 ygap + [ 1 ] π = α π + (1 α ) π + α ygap + α z z + α [ π π ] + ε π c c, c, πld c, + c, πld c, 1 c, y 1 cz, c,3 1 c, 1 RS = γ RS + (1 γ ) * ( RR + π + γ π π + γ ygap ) + ε * * RS RSlag 1 RSlag π + + ygap ( ) ( ) z = δ z + 1 δ z z z RR RR ρ + /+ ε * * US * z z + 1 z 1 1 Seady-sae and equilibrium equaions * * * * y 1 = rpoil rpoil, + [ y y ] g υ π ε * * * g = (1 λ *) + λ * g g 1 + ε g g g * * * π = (1 *) + * 1 + π λ π λ π ε π * * * RR = (1 λ *) + λ * RR RR 1 + ε RR RR RR _ * * * z = + λ * + λ * z z 1 + ε z une *var_ une (1 une )*[(1 ) z z ] * * * * * US ρ = z z+ 1 + RR RR ( ) ( ) π Ideniies ygap = ( y y * ) π RRgap = RR RR * * zgap = z - z [ log( CPI ) log( CPI )] = 1 π = [log( CPI ) log( CPI )] RR = RS π + 1 Variable definiions US z = *log( S * CPI / CPI )

20 19 ygap y * y g * oupu gap, percenage poins log of real GDP log of poenial real GDP growh rae of poenial GDP, quarer/quarer a annual rae, percenage poins g seady-sae growh rae of poenial GDP, quarer/quarer a annual rae, percenage poins π cpi inflaion, quarerly a annualized rae, percenage poins π c core cpi inflaion, quarerly a annualized rae, percenage poins π * arge inflaion rae, annualized rae, percenage poins π four-quarer change in he CPI, annualized rae, percenage poins π Seady-sae inflaion arge, annualized rae, percenage poins π rpoil, change in he relaive price of oil, quarerly a annualized rae, percenage poins π rpoil, four-quarer (moving average) change in he relaive price of oil, percenage poins CPI level of he domesic consumer price index CPI US level of he U.S. consumer price index RS nominal ineres rae, percenage poins RRgap real ineres rae gap, percenage poins RR real ineres rae, percenage poins RR * equilibrium real ineres rae, percenage poins RR seady-sae equilibrium ineres rae, percenage poins RR US U.S. real ineres rae, percenage poins RR *US equilibrium U.S. real ineres rae, percenage poins zgap real exchange rae gap, percenage poins z log of he real exchange (an increase implies a depreciaion) S nominal exchange rae, value of foreign currency in local currency log of he equilibrium real exchange rae (an increase implies a depreciaion) z * z * ρ log of he seady sae equilibrium exchange rae equilibrium risk premium, percenage poins

21 Appendix I.A. Calibraion This appendix oulines he assumpions in he model. Aggregae demand: ygap = β ygap + β ygap β RRgap + β zgap + β ygap + ε US ygap ld + 1 lag 1 RRgap 1 zgap 1 USygap Table I.A.1. Aggregage Demand Equaion Parameer Lower range Upper range Nigeria U.S. Bea_ld Bea_lag Bea_RRgap.5. Sum. Sum.3 Bea_zgap.... Bea_USygap The oupu gap ends o exhibi subsanial ineria (high β lag ), which is normally lower in developing han in developed counries, while he effec from lead oupu β ) is usually limied. ( ld The effec of ineres raes is crucial for he moneary ransmission mechanism, as a larger β would imply a more effecive moneary policy. RRgap The effecs of exchange raes ( β zgap ) and foreign oupu ( β USygap ) end o be larger in more open economies. Significan lags in he ransmission of moneary policy imply ha he sum of β RRgap and β zgap should be smaller han β lag. Phillips curve and core Phillips curve: π π [ z z ] + α π rpoil + α π rpoil ε, 1, π c [ 1 ] = α πldπ + + (1 α πld ) π 1 + α ygap 1 + α z 1 1 ygap + π = α π + (1 α ) π + α ygap + α z z + α [ π π ] + ε c, c, πld c, + c, πld c, 1 c, y 1 cz, c,3 1 c, 1

22 1 Table I.A.. Phillips and Core Phillips Curve Parameer Lower range Upper range Nigeria U.S. Phillips curve Alpha_pie_ld > 1... Alpha_ygap Alpha_z Alpha_ *(1/3).1*(1/3) Alpha_ *(1/3).1*(1/3) Core phillips curve Alpha_c_pie_ld > 1... Alpha_c_ygap Alpha_c_z Alpha_c_ απld deermines he imporance of forward- (and backward-) looking componens in inflaion expecaions. For example, a lower α πld makes i more difficul for he moneary auhoriies o change inflaionary paerns. α ygap characerizes he relaion beween he oupu gap and inflaion. I increases, for example, wih he number of firms ha adjus prices every period. The larger α ygap, he less oupu responds o price level flucuaions. Hence, he larger α ygap is, he smaller he sacrifice raio would be (i.e., he cumulaive loss in oupu as a percen of poenial oupu necessary o permanenly lower inflaion by 1 percenage poin). α z relaes direcly o he weigh of impored goods in he CPI baske and he passhrough of foreign-currency prices (and hence he nominal exchange rae) on o he domesic-currency prices of impors. α and α1 relae o he weigh of oil relaed producs in he CPI baske and he passhrough o prices. α c,3 allows for some feedback from headline inflaion o core inflaion. Moneary policy reacion funcion: RS = γ RS + (1 γ ) * ( RR + π + γ π π + γ ygap ) + ε * * RS RSlag 1 RSlag π + + ygap

23 Table I.A.3. Moneary Policy Rule Parameer Lower range Upper range Nigeria U.S. Gamma_pie > 5... Gamma_RSlag Gamma_ygap > A key parameer in his funcion is γ π, which capures he degree of aggressiveness of he moneary auhoriies. Hence, a higher value for γ π implies ha he auhoriies will respond o a given shock wih a larger change in ineres raes. This ends o fronload he oupu coss bu is appropriae for economies wih lower credibiliy. The parameer γ RSlag measures he aversion of he auhoriies o aler he ineres raes, so ha a higher coefficien implies moneary reacion o a given shock is relaively slow. Given he high degree of uncerainy abou he oupu gap and subsanial real-ime measuremen errors in oupu, he parameer on he oupu gap ( γ ygap ) is usually small. Uncovered ineres pariy US * [ ] z z = E z+ 1 RR RR ρ /+ ε * * [ + 1] = δ + 1+ ( 1 δ ) 1+ ( 1) where E Z zz z z z z Table I.A.. Exchange Rae Parameer Lower range Upper range Nigeria U.S. Dela_z > The parameer δ (wih <δ<1) deermines he degree o which exchange rae expecaions are forward looking as opposed o backward looking. A value closer o 1 implies expecaions are much more forward looking and will hus promp a much greaer exchange rae response o anicipaed changes in fundamenals.

24 3 II. BANKING SECTOR STABILITY FOLLOWING CONSOLIDATION 1 1. Nigeria s financial secor has been ransformed by he banking consolidaion exercise. The consolidaion program required banks o raise heir minimum capial o N 5 billion (abou US$19 million) a end-5 from N 1 billion (US$ 7 million) a end-3. The primary objecive of he increased capializaion was o encourage he banking sysem o consolidae hrough mergers and acquisiions, as a large number of banks were unsound or marginal performers.. The banking secor has grown in various respecs even hough he number of banks was reduced from 9 o. The produc base is broadening and compeiion is inensifying. The increase in capial levels and he need o boos shareholder reurns are promping banks o move ino new areas of operaion, including cross-secor and crossborder operaions, and o exploi he formerly unapped reail secor. As a resul, credi o he privae secor and deposis in he banking sysem have broadly doubled in nominal erms beween end-5 and Sepember 7. Some banks are apping inernaional capial markes o fund heir aciviies, including hrough public offerings, issues of eurobonds, shares on he London Sock Exchange, and Global Deposiory Receips (GDR). 3. This rapidly changing environmen enhances economic prospecs bu poses challenges for financial secor sabiliy. Increased compeiion among banks could benefi consumers by increasing he secor s produc range, improving is service, and expanding access o beer priced services. I could also lead o improved mobilizaion and applicaion of resources, spurring economic growh. However, if no properly managed, he risks associaed wih hese developmens could lead o bank losses and hreaen he sabiliy of he financial sysem.. This chaper assesses he sabiliy of he banking secor in ligh of hese developmens. I assesses he risks and vulnerabiliies facing he banking sysem and assesses he impac of various shocks on banking sysem sabiliy. I also evaluaes he acions being aken o miigae hese risks and vulnerabiliies. 5. The chaper is srucured as follows. Secion A reviews he recen developmens in he financial secor. Secion B evaluaes he soundness of banks in ligh of hese developmens. Secion C examines he resilience of he banks o various quanifiable shocks and repors he resul of sress ess. Secion D considers possible emerging risks and vulnerabiliies o he banking sysem. Secion E discusses he adequacy of banking supervision and oher policy acions being aken o miigae risks. Secion F offers some conclusions and policy recommendaions. 1 Prepared by Sephen Swaray and Jennifer Moyo.

25 A. Recen Developmens. Nigeria s financial sysem is dominaed by privae domesically owned commercial banks. Commercial banks accoun for 9 percen of he financial sysem s oal asses. The recen consolidaion in he banking secor reduced he number of banks o from 9; 19 of he new banks arose from mergers and acquisiions, and banks raised addiional capial or had already me he revised capial requiremen. Only 1 banks accouning for.5 percen of deposis failed o mee he higher minimum capial requiremen and are a differen sages of he liquidaion process by he Naional Deposi Insurance Corporaion (NDIC). 3 Excep for foreign-owned banks, banks shareholdings are diverse wih no relaed pary having a conrolling ineres. Direc and indirec governmen ownership is limied o percen. 7. Banks can be divided ino four groups for analyical purposes (Table II.1). Group 1 comprises he firs generaion banks, which were he larges radiional banks ha achieved he capial hreshold mosly on heir own and may have also consolidaed longesablished affiliaes and acquired one or wo smaller banks. Group 1 banks, herefore, have significan advanages in erms of franchise and a large resource base. Group consiues banks ha achieved he capial hreshold by merging hrough volunary parnerships. Group 3 comprises banks ha achieved he capial hreshold hrough four or more banks parnering ou of necessiy. Group is made up of banks wih majoriy or wholly foreign Table II.1. Bank Groups 1 Group 1 Group Group 3 Group Firs Bank, Guarany Trus, UBA, Union, Inerconinenal, Oceanic, Zenih Access, Diamond, Ecobank, ETB, Fideliy, IBTC-Charered, Plainum-Habib, WEMA, Afribank FCMB, Firs Inland, Skye, Spring, Serling, Uniy NIB, Sanbic, Sandard Charered 1 Bank grouping based on iniial bank srengh, number of banks ha merged, and ownership. ownership.. The nonbank financial secor accouns for some percen of he financial sysem s asses. The pension funds accoun for half of his oal, wih various oher insiuions accouning for he remainder which consis of 3 insurance companies, 75 communiy banks (which are being ransformed ino microfinance insiuions), 7 microfinance banks, 91 primary morgage insiuions, a sock exchange, a commodiy Two banks have merged since he iniial consolidaion. 3 Some banks ha failed o mee he higher minimum capial requiremen challenged he NDIC s peiion o liquidae hem. To dae, approval for liquidaion has been graned o 11 banks; 3 banks remain in cour.

26 5 exchange and 3 bureaux de change. In addiion, he sysem has acive money, foreign exchange, and capial markes. 9. Since consolidaion, he banks have acquired diverse cross-secor and crossborder linkages. Universal banking groups have emerged, wih more han half he banks having insurance and securiies subsidiaries. As par of a regional expansion drive, abou half he banks have cross-border operaions in oher counries, including he Gambia, Benin, Ghana, Sierra Leone, Souh Africa, he U.K., and he Unied Saes. While sill small, hese operaions are growing and banks are planning o expand hem.. Since consolidaion, banks balance shees have expanded considerably. From end-5 o Sepember 7, he sysem s balance shee has expanded by 1 percen, underpinned by increasing public confidence; deposis grew by 9 percen over he same ime period. Credi o he privae secor has also grown considerably in absolue erms. However, deposis and privae secor credi are sill relaively small given he size of he economy and compared wih ha in peer counries. 11. Wih banks under pressure o use funds raised from increased capializaion, credi o he privae secor has surged. Banks are moving ino new areas, including he reail secor. Several banks have increased heir branch nework, wih bank branches sanding a abou, a end-june, up from abou 3, jus afer consolidaion. Oher banks have begun o pursue lending opporuniies in new areas, such as infrasrucure, consrucion, oil, and gas financing. The need o ap he reail marke and increase service delivery o small and medium-sized enerprises, in paricular, has spurred banks o develop ailored producs, upgrade IT sysems, and expand ATM infrasrucure. 1. Despie he already high capial levels, banks coninue o augmen heir capial base, aided by significan inernaional capial flows. Increased capial is being raised hrough public offerings, issues of eurobonds, shares on he London Sock Exchange, and GDRs. One bank, for example, has been able o raise close o $1. billion from foreign invesors. The increased inernaional invesor ineres is a direc resul of he improved confidence in he banking secor. Among banks, he capializaion drive has been moivaed by hree facors. Firs, he Cenral Bank of Nigeria (CBN s) encouraged he formaion of larger banks including hrough he requiremen ha a bank have N 13 billion in capial o manage foreign reserves. Subsequenly, he CBN has removed his incenive. Second, some banks are posiioning hemselves o aggressively ake over oher banks in he even of furher bank consolidaion; an even some see as likely. Third, he increased capial will enable banks o lend for larger projecs.

27 13. Inernaional capial inflows are an imporan feaure of Nigeria s new financial sysem. Besides invesing in banks, inernaional invesors are increasingly invesing in governmen securiies, boh direcly and indirecly hrough derivaives. Alhough complee daa on he volume of hese invesmens is no available, i is esimaed ha nonresidens holding of governmen securiies is subsanial, amouning o N 575 billion (US$. billion) as of Augus 7. By asse class, nonresidens held almos percen of Federal Governmen of Nigeria (FGN) bonds ousanding, and abou percen of reasury bills and bonds. 1. Sronger capial flows o banks have increased aciviy in he inerbank foreign exchange marke, while he inerbank money marke, alhough srenghened, remains relaively weak. In he foreign exchange marke, new insrumens are beginning o emerge, including swaps and forward conracs, promping banks o srenghen heir reasury capaciies, mosly hrough aracing saff from abroad or hrough collaboraion and parnerships wih inernaional banks. In he money marke, he srenghened capial base of banks has parly helped he inerbank money marke o develop by raising confidence in he qualiy of poenial counerparies. This greaer confidence, in urn, has helped he cenral bank manage liquidiy. Sill, here is a endency for banks o simulaneously hold excess liquidiy; herefore, deerring efficien marke clearing. B. Bank Soundness 15. The banking secor is broadly sable. Banks are well capialized, liquid, and profiable. The disress in,5 Deposi Growh Srucure he banking sysem, (Billions of naira) before consolidaion, 3,5 when a large number of Foreign currency Time and savings banks were unsound or 3, Demand marginal performers, has,5 eased; he sysemwide, balance shee has almos doubled since December 1,5 5. Confidence has 1, reurned wih growh in 5 deposis nearly doubling since December 5. Dec. 5 Dec. Sep. 7 Source: Cenral Bank of Nigeria (CBN).,5, 3,5 3,,5, 1,5 1, 5

28 7 Capial adequacy 1. Banks are generally well-capialized wih an overall capial adequacy raio of 1. percen (Table II.3), considerably above ha in oher counries in he region (Table II.). The generally high capial adequacy is due largely o he consolidaion process, which required banks o raise heir capial base; some raised capial beyond regulaory requiremens. Three banks, accouning for abou percen of oal bank asses, were sill marginally undercapialized in June 7. While he increase in he absolue capial level could suppor larger exposures, i underscores why prudenial limis mus be sricly enforced. Given he CBN s exposure limi, which caps lending o a single borrower a percen, banks can make much larger loans (in absolue erms) o a single eniy. The resuling poenial for larger exposures and increased risk needs close monioring. Asse Qualiy 17. The qualiy of he loan porfolio has improved since consolidaion (Table II.), chiefly because a large proporion of nonperforming loans (NPLs) were wrien off when 1 banks closed in. These wrie-offs brough he aggregae NPL-o-oal loans raio o. percen in 7, from 1.1 percen in 5. Toal loans also grew considerably, causing he raio o decline, while he consolidaed banks worked o recover ousanding loans. Furher, mos banks ook more pruden measures o ensure ha he new loans performed beer; bu he impac of such measures on NPLs will become eviden only over he nex 1 o monhs. 1. NPLs have also declined in he individual bank groups. However, because of legacy loans from previously weak banks, he decline in he NPL-o-oal loans raio in Group 3 was lower han would be expeced for a growing loan book (Table II.). This paern in NPLs needs o be moniored closely. 19. The paern of provisioning followed ha of NPLs. In aggregae, afer significan provisioning in he period immediaely afer consolidaion, provisioning levels have since declined. The provisions-o-npl raio declined from. percen as of March, o. percen a June 7. The raio declined mainly because he loans wihin he pool of NPLs migraed upward. This paern was also refleced in he changes in Group 1 and Group 3 raios. Group and Group raios iniially increased in December, hen declined in June 7.

29 Table II.. Porfolio Qualiy by Group Number Mar- Dec- Jun-7 Mar- Dec- Jun-7 of banks Nonperforming loans (NPL) 1 Provisions/ NPLs Group Group Group Group Ne of provisions relaive o capial. Earnings and Profiabiliy. Banking consolidaion has resuled in a reducion in ne ineres rae margins. These have sabilized a around 3. percen, agains. percen for he world on average and percen in sub-saharan Africa. This drop largely reflecs reduced margins in large corporae businesses, reduced reasury bill raes, and an increase in loan-loss provisioning. 1. Banking secor profiabiliy indices have been dampened owing o declining margins and recen capial raising. The reurn on asses (ROA) and on equiy (ROE) were almos halved from December o June 7; he ROA sands a 1. percen and he ROE a 13. percen, considerably below raes in comparaor counries (Table II.). The decline in profiabiliy indices can be aribued o he increase in shareholder funds ha were used mainly o acquire IT and oher faciliies, such as new branches, o ensure newly consolidaed banks could compee. In addiion, some banks, primarily in Group 3, faced pos-consolidaion adminisraive challenges, and heir effors were focused on addressing hese challenges raher han on improving profiabiliy. Looking ahead, banks will be under pressure o raise profiabiliy o mee shareholder expecaions.. Performance, however, varies across he banks. Generally, he firs generaion banks (he larges radiional banks ha achieved he capial hreshold more or less on heir own) and he banks ha formed hrough volunary parnerships appear o fare beer han banks where four or more banks came ogeher. Liquidiy 3. The banking secor s holding of liquid asses has increased since consolidaion. Compared wih he percen prudenial regulaion, he average liquidiy raio increased from 1.1 percen a end-december 5 o. percen a end-december. The raio of loans and advances o deposis over ha period increased from 5.1 o 75. percen.

30 9 Table II. 3. Financial Soundness Indicaors for he Banking Secor, 7 (Percen, unless oherwise indicaed) Capial Adequacy Dec- Dec-5 Dec- Jun-7 Regulaory capial o risk-weighed asses Regulaory Tier I capial o risk-weighed asses Capial (ne worh) o asses Asse qualiy and composiion NPLs o gross loans NPLs ne of provisions o capial Secoral disribuion of loans 1 Manufacuring Trade and services Energy and minerals Agriculure Consrucion and propery Households Governmen Oher Earnings and profiabiliy ROA ROE Ineres margin o gross income Nonineres expenses o gross income Personnel expenses o nonineres expenses Trading and fee income o oal income Liquidiy Liquid asses o oal asses Liquid asses o shor erm liabiliies 1 Cusomer deposis o oal (non-inerbank) loans Foreign exchange liabiliies o oal liabiliies Source: Cenral Bank of Nigeria (CBN). 1 Included in he "core se" of financial soundness indicaors. NPL - non performing loans; ROA - reurn on asses; ROE - reurn on equiies.

31 3 Table II.. Comparaor Counries: Financial Soundness Indicaors, 7 (Percen, unless oherwise indicaed) CAR Non Reurn Reurn Performing on Asses on Equiies Loans (NPL) 1 (ROA) (ROE) Brazil Indonesia Malaysia Nigeria Souh Africa Sources: Cenral Bank of Nigeria (CBN), and IMF, GFSR. 1 NPLs o gross loans for Nigeria and Malaysia NPLs o oal loans for all ohers. CAR, NPL for 5, ROA and ROE for March. 3 CAR Sep. ; NPLs include compromised asses raio, resrucured loans, and foreclosed asses for he larges 1 banks. ROE is based on he larges 1 banks. ROA is before ax. 5 All enries are for March 7. C. Sress Tess. Sress ess using individual bank daa a end-june were conduced o assess he impac of full provisioning for NPLs, credi risk, and ineres rae risk on he asse qualiy of banks (Tables II.5 and II.). Box 1 describes he shocks and oulines he resuls. 5. Overall, banks appear resilien o he shocks measured. Even he effecs of credi risk, he larges quanifiable risk banks face, appear manageable. Only a small number of banks, accouning for less han 5 percen of asses, are undercapialized afer NPLs increase by percen. Two small banks accouning for less han 5 percen of sysem asses become insolven. Ineres rae risk also appears well-conained.

32 31 Box II.1: Sress Tess: Mehodology and Resuls Sress ess were conduced o assess he impac of boh credi and ineres rae shocks. The analysis is based on end-june 7 daa for individual banks. Credi risk is assessed using hree shocks (Table II.5). Firs, an allowance is made for full provisioning. Second, an increase in he NPLs raios by 5, 5 and percen, possibly reflecing adverse macroeconomic developmens is simulaed. Third, a migraion of 5 and 5 percen of he loans from performing o non-performing loan saus, assuming he same disribuion of he new nonperforming loans among he subsandard, doubful, and loss caegories, is assessed. The effecs of credi risk are manageable. The effec of a correcion for under-provisioning of NPLs is minimal. Exra provisioning o provide percen coverage of repored NPLs has a limied impac on banks capial adequacy. This resul is no surprising since he banks have been provisioning more for he NPLs. This shock resuls in four banks (consiuing 3 percen of asses) wih capial adequacy raios (CARs) less han he sauory percen. Furher, given ha before he shock hree banks already had CARs below percen, he effec of he shock is limied. Indeed, for only one addiional small bank (banking asses comprising abou.7 percen of oal banking asses) does he raio fall below percen. The effec of he second caegory of shocks simulaing an increase in NPLs also appears limied: wo banks accouning for only 3.5 percen of sysem asses become insolven when NPLs double. The effec of he hird se of ess relaing o he migraion of loans o nonperforming saus is also limied: five banks accouning for percen of sysem asses become undercapialized. No banks become insolven. Ineres rae risk is assessed using he radiional mauriy-gap analysis (Table II.). Shocks are calibraed for boh an increase and decrease in he ineres rae by percenage poins. The ineres rae change is applied on he one-year cumulaive gap beween asses and liabiliies afer having accouned for he zero-ineres demand deposis, and considering only he flow of funds. This es is based only on he flow impac from a gap beween ineres-sensiive asses and liabiliies. The effecs of ineres risk are also manageable. An increase in ineres raes by percenage poins benefis he banking secor, while a similar decrease in ineres raes resuls in five banks (7 percen of banking sysem asses) becoming significanly undercapialized (wih hree banks having a CAR of abou 5 percen). The beneficial effec on he Nigerian secor of an ineres rae increase can be aribued o he low-cos demand deposis in Nigeria ha arac a zero ineres rae, while mos ineres-bearing liabiliies are very shor erm (hree monhs or less). On he asse side, oal credis are mainly shor erm (hree monhs or less). A percenage poin increase in ineres raes raise all banks capial adequacy raio (including hose ha were marginally undercapialized) o levels above he prudenial requiremen of percen. The benefis, however, of an ineres rae increase may be offse parly by he second-round effecs of deerioraion in asse qualiy because of higher ineres raes. Accouning for he sock impac on he aggregae bank porfolio resuling from he repricing of bonds afer a change in ineres raes does no aler he resul. The capial adequacy raio remains high a.3 percen.

33 3 Table II.5: Sensiiviy o Credi Risk (Regulaory Capial o Risk-weighed Asses, percen) NPLs Increase Migraion o NPLs of Before Provisioning 5% of 5% of he Shock Brough o Performing Performing Required Levels 5% 5% % Loans Loans All banks Group Group Group Group Banks wih CAR < % Share of asses Banks wih CAR < 1.. Share of asses Sources: Cenral Bank of Nigeria (CBN) daa, and IMF saff esimaes. Table II.. Sensiiviy o Ineres Rae Risk (Regulaory Capial o Risk-weighed Asses, percen) Ineres Rae Change % Increase Before % % plus Bonds he Shock Increase Decrease Repricing 1 All banks Group Group Group Group Banks wih CAR < % Share of asses.7 7. Banks wih CAR < Share of asses Sources: Cenral Bank of Nigeria (CBN), and IMF saff esimaes. 1 Uses oal governmen bonds held by banks and assumes a sysem average mauriy of. years

34 33 D. Emerging Risks and Vulnerabiliies. While he banking secor is sable, prospecive developmens pose risks ha need o be managed. These include he following: Large banks are operaing across borders and across secors wihou having risk-based and consolidaed supervision fully implemened. Wihou consolidaed supervision and home-hos supervision arrangemens, he risks assumed by hese banks could go undeeced. Furher, cross-border operaions, paricularly in neighboring counries ha face poliical insabiliy, expose banks o hese counries sovereign risk. Developmens may have oupaced banks risk managemen capaciies. The skills and human resources needed o manage complex insiuions and assess risks may have lagged he pace of developmen of financial insiuions and producs. Banks are working o augmen heir capaciies in hese areas, bu such effors are expeced o bear frui only in he medium erm. Banks under pressure o deliver shareholder reurns migh be overly eager o inves capial and deposis in risky aciviies. As compeiion has increased, banks have moved ino new and riskier areas of operaion, including he reail secor, and have invesed in long-erm producs like railway bonds and in long-erm infrasrucure invesmens. If no adequaely moniored, such aciviies could resul in bank losses and hreaen overall sabiliy. Foreign invesmen in domesic reasury bills and bonds could presen imporan challenges should invesor senimen change. Given he banking secor s increasing share on he sock exchange, banking secor insabiliy could quickly spread o he capial markes. Banks now make up close o percen of he sock exchange s marke capializaion. This heavy concenraion increases he risk ha bank insabiliy could cause a sock marke crashes, dashing confidence in he banking sysem. Poenial macroeconomic imbalances could add o financial secor risk. Excessive spending a he federal or sae level leading could impair economic performance and undermine financial secor sabiliy. Nigeria s economy also remains relaively undiversified and heavily dependen on oil, exposing i o erms of rade shocks.