Bank balance shees, lending and he macroeconomy ea Zicchino Bank of England Join HKIMR and CCBS Workshop on Financial Markes, Financial Sabiliy, and Financial Fragiliy 29 November-2 December 2005
Wha is he relaionship beween moneary policy and financial sabiliy? Are moneary shocks amplified by he banking secor? How fricions in credi markes inerplay wih moneary policy shocks? Wha is he role played by prudenial regulaion in he ransmission of macroeconomic and moneary shocks o he banking secor? 2
Transmission of moneary (and macro) shocks The credi channel: i is based on he heoreical predicion ha he exernal finance premium should depend on borrowers financial posiion a) The balance shee channel sresses he poenial impac of changes in moneary policy on borrowers balance shees (such as borrowers ne worh, cash flow and liquid asses) b) The bank lending channel focuses on moneary policy affecing he exernal finance premium by shifing he supply of inermediaed credi 2. The bank capial channel: moneary policy and oher macro shocks can affec bank lending hrough heir impac on bank capial 3
a. The credi (or firms balance shee) channel 4
b. The bank lending channel According o he lending view, moneary policy affecs banks loan supply For example, a conracion in reserves leads banks o reduce loan supply, hereby raising he cos of capial o bankdependen borrowers The exisence of a bank lending channel does no require banks o be oally incapable of replacing los deposis (due o conracion in reserves). I is sufficien ha banks do no face a perfecly elasic demand for heir liabiliies 5
Terms of lending shows he ne percenage of small businesses who repored ha credi was harder o obain relaive o he previous quarer Business cycle peaks 6
2. The bank capial channel Moneary policy shocks Regulaory policy shocks Macroeconomic shocks We now focus on on he he impac of of macroeconomic condiions on on bank capial (Zicchino, [2005]) and he he poenial procyclicaliy of of he he Basel regulaory regime 7
Ouline of he res of he presenaion The basic heoreical model (Basel I) Exension o he model (Basel II) Resuls and implicaions for he procyclicaliy debae Empirical analysis: Bank weakness and bank loan supply (Zicchino, Nier [2005]) 8
The basic model (Basel I) Dynamic, Oligopolisic Banking Indusry A bank maximises he presen value of fuure profis subjec o a capial adequacy consrain q s Tier capial: and Tier 2 capial: b Two asses:, T as period s marke value of equiy and bonds deermine he capial consrain for he curren period: q s b θ 9
The basic model (Basel I) oan demand: = l0 lr l2m ε, ε, is a random exogenous shock o loan demand defined over [, ] M = ρ M M ε M, ε M, is a mean zero sochasic shock o macroeconomic condiions 0
[ ] [ ] = v E b s q b s q v Max θ λ τ π ( ) ( ) p D D T M T D D r D c D r c T r r δ π = 2 2 A bank s maximisaion problem subjec o: M l r l l, ε = 2 0 ( ) ( ) b b r s q d b s q π = ( ) b q D T = α
Sylised facs on loan losses δ( M ) Wrie-off/loan raio and UK (invered) oupu gap Per cen 6 (Invered) Oupu gap (HS) Wrie-off/loan raio (RHS) Per cen.6 4.4 _ 2 0-2 -4.2.0 0.8 0.6 0.4-6 0.2-8 978 80 82 84 86 88 90 92 94 96 98 2000 0.0 Sources: Bank of England; OECD Economic Oulook. 2
Sylised facs on loan losses δ( M ) Wrie-off/loan raio and UK GDP growh rae Per cen 2 Wrie-off/loan raio (RHS) (Invered) GDP growh rae (HS) _ 0 - Per cen.6.4.2.0-2 -3-4 -5 0.8 0.6 0.4 0.2-6 0.0 88 90 92 94 96 98 2000 02 Source: Bank of England. 3
4 A bank maximises is value by choosing he oan Rae, Deposis, and Capial: from balance shee ideniy from loan demand ( ) ( ) ( ) ( ) ( ) ( ) 0 0 0 = = = = = = b s q E v b s q v df b s q v m E D D v df b s q v m E r r v,,,,... ε τ π ε τ π : T :
5 Time line ( ) b s q θ,,,, b q s T D r ( ) b s q θ 2 2,,,, b s q T D r
The model soluion oan rae r MC MR D Bank oans 6
7 ( ) j j j M j M j M M l l M = δ ρ ρ 2 2 2 Resuls Effec on bank loans and loan rae of a shock o M Case I (he capial consrain is non-binding) Banks loans increase as a resul of an improvemen in macroeconomic aciviy (he loan rae can eiher increase or decrease depending on he parameers of he loan demand funcion and he loan defaul rae) ( ) j j j M j M j M M l l M r = δ ρ ρ 2 2 2
oan rae Effec on bank loans and loan rae of a shock o M Case I: he capial consrain is non-binding r 2 r MC MC 2 MR MR 2 D D 2 2 Bank oans 8
Resuls Effec on bank loans and loan rae of a shock o M Case II (he capial consrain is binding) * * q s b = = 0 since M θ r M = l l 2 9
oan supply and loan rae afer a posiive shock o M oan rae Case II: he capial consrain is binding r 2 r MC MC 2 MR MR 2 D = 2 D 2 Bank oans 20
Resuls Effec on bank capial of a shock o M Banks increase (decrease) heir capial holdings in response o a posiive (negaive) shock in he curren level of macroeconomic aciviy (M ) 2
Basel II model Cyclical paern of asse risk-weighs Sandardised Approach: evidence of a cyclical paern in credi raings se by raing agencies (Nickell, Perraudin and Varoo, 2000) Inernal Raing Based Approach: mos inernal raing sysems use a shor-erm horizon o measure risk. For example, borrowers probabiliy of defaul is ofen deermined over a one-year period (Basel Commiee on Banking Supervision, 2000) 22
Basel II model The capial consrain is now: where ( ) w M M < 0 θ q w M ( ) s b w T T, In conras o he previous case, under he assumpion of a risk-sensiive capial consrain, banks can eiher raise or lower heir capial holdings afer a posiive shock o M Their choice depends on he effec ha a change in M has on he likelihood of he capial consrain binding nex period 23
ε Effec on capial depends on he sign of, M * where ε, is he criical shock o loan demand a ime such ha he capial consrain jus binds *, If ε *, M > 0 capial decreases In In conras o o he he case of of consan risk-weighs, a posiive shock o o curren macroeconomic condiions does no necessarily induce a bank o o increase he he amoun of of capial i i holds 24
Resuls Effec on bank loans and loan rae of a shock o M Case I (he capial consrain is non-binding): same effec as before Case II (he capial consrain is binding) M = q θ s b [ w( M )] 2 ( ) w M M r M = l l 2 q θ s b [ w( M )] 2 ( ) w M M 25
oan supply and loan rae afer a posiive shock o M oan rae Case II: he capial consrain is binding r I B I r II B II r MC MC MR MR D D = I II Bank oans 26
Summary able: responses o a posiive shock o macroeconomic condiions M oans oan rae Bank capial Basel I, non-binding capial consrain Basel I, binding capial consrain? = Basel II, non-binding capial consrain Basel II, binding capial consrain? ower han in he Basel I model Undeermined, lower han in he Basel I model Undeermined, lower han in he Basel I model 27
Summary of he resuls If loan risk-weighs depend on M, we inroduce anoher channel hrough which curren macroeconomic condiions affec banks lending: a capial raio channel Under Basel II rules, a macroeconomic shock will affec he risk-weighs so ha he capial consrain becomes looser (if he shock is posiive) Therefore banks may no be willing o mainain he same level of capial during periods of high economic aciviy as under Basel I 28
Empirical Evidence: Bank weakness and bank loan supply (Zicchino and Nier, [2005]) Peek and Rosengren (995): capial crunch for New England banks during a period in which capial regulaion was acively enforced (990-9); Empirical sudies of he impac of changes in capial regulaion on banks lending behaviour ouside he US are rare: Gambacora & Misrulli (2003); Barajas, Chami & Cosimano (2005); 29
Hypohesis developmen If a reduced volume of loans during recessions is he resul solely of lower loan demand or of a reducion in loan supply due o a lower level of savings, hen all banks should reac by reallocaing heir porfolios in he same way, independen of heir specific characerisics Bu if we observe a differen behaviour among bank wih differen characerisics, his would be evidence of a supply-induced resricion of credi 30
Hypoheses (H) Is he loan supply of banks wih weaker balance shees lower han ha of sronger banks? (H2) Is he effec of loan losses on bank loan supply sronger for poorly capialised banks? (H3) Is he effec of bank capial on lending sronger during periods of weak economic growh? 3
Variables an daa The sample used in his research conains daa on more han 600 lised banks from 3 counries from 993 o 2000 Table I. Summary Saisics Mean Sd. Dev. Min Max Obs. No. of banks oan growh 0.095 0.89-0.583 0.962 423 735 RoE 0.08 0.6-0.798 0.554 48 736 Provisions 0.009 0.04-0.004 0.32 4444 676 Capial 0.098 0.3 0.03.929 493 78 GDP growh 0.03 0.023-0.069 0.095 493 78 32
Table II. Average of Provision Raio Year Argenina Ausralia Ausria Bel gium Brazil Canada Chile Finland France Germany Greece Hong Kong Indonesia Ireland Israel Ialy 993 0.08 0.007.. 0.008 0.005 0.00 0.02 0.04. 0.0 0.003 0.009 0.008 0.006 0.08 994 0.094 0.0036.. 0.095 0.0069 0.0049 0.075 0.04 0.0067 0.023 0.0050 0.00 0.0034 0.0035 0.028 995 0.033 0.0022.. 0.0654 0.0048 0.0028 0.050 0.023 0.005 0.0078 0.0053 0.0078 0.0027 0.0052 0.027 996 0.035 0.005 0.0080. 0.062 0.0038 0.0053 0.0058 0.003 0.0047 0.000 0.0049 0.007 0.0023 0.0047 0.07 997 0.020 0.007 0.007 0.0024 0.0449 0.0036 0.0072 0.006 0.054 0.0057 0.03 0.0082 0.025 0.006 0.0050 0.0092 998 0.024 0.0034 0.0059 0.000 0.0463 0.0035 0.046-0.0004 0.0076 0.0048 0.03 0.0244 0.0500 0.008 0.005 0.0082 999 0.0293 0.003 0.0074 0.0036 0.0437 0.0039 0.088 0.0003 0.0098 0.005 0.020 0.0267 0.0520 0.003 0.0036 0.0070 2000 0.0268 0.007. 0.0024 0.0365 0.0044 0.00 0.0003 0.0068 0.004 0.0095 0.044 0.027 0.0028 0.0040 0.0064 Toal 0.026 0.0029 0.0080 0.002 0.043 0.0044 0.000 0.0065 0.04 0.0050 0.0 0.05 0.077 0.0027 0.0046 0.003 Obs 40 85 4 2 94 58 54 9 47 35 42 92 82 3 87 276 No of banks 7 3 20 8 9 3 22 9 9 2 4 5 2 42 Year Japan Korea Malaysia Neherlands Norway Poland Porugal Singapore Spain Sweden Swizerland Taiwan Thai land Turkey Unied Kingdom USA 993 0.003 0.0069 0.009 0.0053 0.004 0.0252 0.083 0.003 0.0237 0.022 0.0044 0.0044 0.0049 0.0047 0.027 0.0059 994 0.0039 0.002 0.005 0.0057 0.0037 0.0389 0.046 0.0035 0.02 0.022 0.0099 0.0046 0.0053 0.0269 0.0072 0.0034 995 0.0097 0.0085 0.0068 0.005 0.0020 0.020 0.030 0.0023 0.000 0.0076 0.0035 0.0035 0.0040 0.0408 0.0052 0.0039 996 0.0073 0.006 0.007 0.0044 0.005 0.076 0.020 0.0036 0.006 0.0033 0.0063 0.0045 0.0042 0.09 0.005 0.004 997 0.023 0.028 0.060 0.0026 0.004 0.0065 0.0220 0.037 0.0036 0.009 0.0052 0.0069 0.0308 0.0074 0.0054 0.0044 998 0.056 0.0467 0.0509 0.0034 0.0029 0.05 0.0098 0.024 0.0033 0.009 0.0054 0.024 0.079 0.076 0.0054 0.005 999 0.009 0.0457 0.0204 0.0020 0.0029 0.04 0.06 0.064 0.003-0.000 0.0047 0.07 0.0855 0.0226 0.0052 0.0049 2000 0.0092 0.0534 0.035 0.005 0.0029 0.023 0.00 0.0037 0.0053 0.0008 0.0054 0.052 0.0262 0.0252 0.0043 0.0049 Toal 0.0090 0.0307 0.09 0.0036 0.0032 0.083 0.039 0.02 0.0082 0.0059 0.0053 0.0084 0.0334 0.093 0.006 0.0046 Obs 690 6 47 32 98 70 5 30 46 38 63 89 57 6 8 472 No of banks 00 3 8 5 4 3 7 6 9 6 4 28 3 3 26 204 less han median beween median and 75h percenile greaer han 75h percenile 33
Main resuls Table : Deerminans of loan growh () (2) (3) Dependen variables oan growh oan growh oan growh RoE 0.25*** 0.23*** 0.42*** Capial 0.005-0.09** 0.*** Provisions -.84*** -2.85*** -.53*** GDP 0.43*** 0.43*** 0.4*** agged Capial*Provisions 6.23*** GDP*Capial -0.5** Year dummies yes Yes yes Counry dummies yes Yes no No. of obs. 3460 3453 3460 Goodness of fi (a) 0.25 0.25 0.8 *** denoes saisical significance a he one percen level, ** a he five percen level and * a he en percen level. (a) The goodness of fi is calculaed as he square of he correlaion coefficien beween he acual and he fied value of he loan growh. 34
Exension: differences across counry and ime Table 2: Deerminans of loan growh: crisis versus non-crisis counries () banking crisis (2) no banking crisis (3) banking crisis (4) no banking crisis Dependen variables oan growh oan growh oan growh oan growh RoE 0.2*** 0.27*** 0.8*** 0.25*** Capial -0.3 0.03-0.44*** -0.06 Provisions -.69*** -2.44*** -3.33*** -3.98*** GDP 0.53*** 0.76*** 0.5*** 0.8*** agged Capial*Provisions 3.33*** 6.86** Year dummies yes Yes yes yes Counry dummies yes Yes yes yes No. of obs. 040 2420 036 247 Goodness of fi 0.4 0.5 0.42 0.5 *** denoes saisical significance a he one percen level, ** a he five percen level and * a he en percen level. 35
Financial marke developmen and srucure () (2) (3) (4) (5) (6) Dependen variable: oan growh RoE 0.38*** 0.39*** 0.42*** 0.42*** 0.37*** 0.38*** Capial 0.06* 0.05 0.05 0.05 0.09 0.09** Provisions -.76*** -2.6*** -.3*** -2.5*** -.7*** -.27*** GDP 0.20*** 0.2*** 0.36*** 0.35*** 0.28*** 0.28*** Financial developmen -0.09*** -0.0*** Financial developmen*provisions 0.77 Bank concenraion -0.08*** -0.09*** Bank concenraion*provisions.45 Foreign banks 0.7*** 0.8*** Foreign banks*provisions -.67 Year dummies yes yes yes yes yes yes Counry dummies no no no no no no No. of obs. 3460 3460 342 342 323 323 Goodness of fi 0.9 0.9 0.9 0.9 0.9 0.9 36
Bank weakness and he effec of moneary policy Dependen variable: oan growh () (2) (3) (4) (5) (6) RoE 0.25*** 0.25*** 0.25*** 0.25*** 0.24*** 0.25*** Capial 0.006 0.006 0.005 0.006 0.009 0.004 Provisions -.87*** -.89*** -.85*** -.36*** -3.5*** -.9*** GDP 0.45*** 0.45*** 0.45*** 0.42*** 0.43*** 0.46*** r -0.02*** -0.005 r - 0.02*** 0.0 r 0-0.03-0.04* Provisions* r -2.32*** Provisions* r -.74*** Provisions* r 0 0.94 Year dummies yes Yes yes Yes yes yes Counry dummies yes Yes yes Yes yes yes No. of obs. 3436 3436 3436 3436 3436 3436 Goodness of fi 0.25 0.25 0.25 0.26 0.26 0.26 37
Summary of empirical resuls Procyclicaliy of bank lending is a widespread phenomenon and by no means confined o specific episodes We documen ha: shocks o capial resul in a reducion of bank loan supply his effec is sronger in crisis imes moneary policy is more effecive when banks have weak balance shees 38
Concluding remarks Credi marke fricions play a role in he ransmission of moneary (and macroeconomic) shocks Banks balance shee posiion can amplify moneary and regulaory policy shocks as well as macroeconomic shocks Risk-sensiive capial requiremens may amplify he procyclicaliy of bank lending 39