How does Loan-o-Value Policy Srenghen Banks Resilience o Propery Price Shocks: Evidence from Hong Kong Eric Wong Research Deparmen Hong Kong Moneary Auhoriy Presenaion a he IMF-EBA Colloquium on New Froniers in Sress Tesing in London 2 March 2017 The views and analysis expressed in his presenaion are hose of he presener and do no necessarily represen he views of he Hong Kong Moneary Auhoriy 1
Research quesion and main findings Wha should be he opimal arge of LTV policy (i.e. limiing LTV caps) in pursuing banking sabiliy? Propery prices Household leverage Credi growh Main Findings: Very limied policy ransmission hrough propery markes The main policy impac is ransmied hrough household leverage raher han credi growh 2
Road map A brief overview of MPPs in HK Effeciveness of LTV policy from a hisorical perspecive The impac of LTV policy on propery prices The impac of LTV policy on household leverage (direc impac) and credi growh (indirec impac) Economeric evidence Acual scenario (wih LTV policy ighening) and Counerfacual scenario (wihou ighening) Defaul risk of wo scenarios under severe macro sress The conribuions of he direc and indirec impacs 3
A brief overview of MPPs in HK 1. By HKMA 1. LTV raio caps since 1991 2. Deb-serving raio (DSR) limis, and sress-esing DSRs 3. Maximum enor of 30 years for morgage loans 4. A risk-weigh floor of 15% for morgage loans for banks adoping IRB 5. Regulaory reserves 6. Sable funding requiremen since 2014 2. By HKSAR governmen 1. Special samp duies (SSD) since Nov 2010 2. Buyer samp duies (BSD) since Oc 2012 3. Double samp duies (DSD) since Feb 2013 4
Hisory of LTV policy in HK 5
Effeciveness of LTV policy in pas episode of propery marke downurn Index (Q3 1997 = 100) 100 90 80 70 60 50 Residenial propery price index (lhs) -66% % 5 4 3 40 30 20 10 0 Delinquency raio (rhs) 1996 1998 2000 2002 2004 2 1 0 Noe: Delinquency raios are compued based on hose pas due 30 days or more. Sources: HKMA and Raing and Valuaion Deparmen 6
Propery marke since 2009 How far he LTV policy is ransmied hrough he propery marke? 7
Lower LTV caps Transmission Channels of LTV policy Consrain borrowers Loan marke Credi demand Credi supply Direc effec ( leverage) Indirec effec ( credi growh) Impac on he credi cycle PD and LGD of borrowers Improve resilience Source: A simplified version of graph 3.3. from CGFS (2012), Operaionalising he selecion and applicaion of macroprudenial insrumens, CGFS Paper no. 48 8
Economeric evidence of he direc effec: How do LTV caps reduce he average LTV raio in he marke Long-run deerminans of he marke LTV raio LTV LTV 0.696* DSR ) [ 6.83] (0.605 0.326* LTVcap [14.3] [16.2] 0.285* Proreurn [2.91] 2 Adj. R 0.77 Conribuion of main facors o change in he marke LTV raio 1.436* Proyield [6.11] 9
LTV raios under he acual and counerfacual scenarios The marke LTV raio under he acual and counerfacual no policy scenarios 10
An economeric model for he indirec effec 1. Model characerisics a) Allowing for disequilibrium (sandard assumpion, Sigliz and Weiss, 1981) b) Demand and supply models (as LTV policy may affec boh he demand and supply of morgages) c) Relevan MPP variables are included in he equaions as conrol variables 2. Esimaion resuls a) Q D = f(ltv, r, special samp duies, DSR limi, unemploymen rae) + - - - b) Q S = f(ltv, RAROC(r), house price growh, deposi growh, DSR limi) + + + + - 11
Indirec effec: quanifying he impac on supply and demand Esimaed supply of morgage loans Esimaed demand for morgage loans 12
Defaul risk analysis 1. Acual scenario: acual daa of LTV and new morgage loans 2. Counerfacual scenario: based on he esimaed LTV and new morgage loans assuming no LTV policy ighening 3. Comparing he defaul risk under he wo scenarios wih he following macro shocks 1. Morgage ineres raes increase by 300 basis poins 2. Propery prices drop by 60% 3. Household income decreases by 20% 4. Unemploymen rae increases o 8.5% From 2013Q1 o 2014Q4 4. Analysis he conribuion of he direc and indirec impacs 13
Esimaed LTV raios for seleced vinage monhs under he acual scenario wih propery price shocks 1.4 1.2 1 Dec-12 Jun-12 Jan-12 Jan-11 Jan-10 Jan-09 Jan-08 Jan-07 Assuming propery prices drop by 60% Dec-12 1.25 1.16 Jun-12 0.8 0.6 0.4 0.2 0 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 14
Esimaed LTV raios for seleced vinage monhs under he counerfacual scenario wih propery price shocks 1.6 1.4 1.2 Dec-12 Jun-12 Jan-12 Jan-11 Jan-10 Jan-09 Jan-08 Jan-07 Assuming propery prices drop by 60% Dec-12 1.40 1.30 Jun-12 1 0.8 0.6 0.4 0.2 0 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 15
Conribuion of direc and indirec effecs P NE k1 N k Lk, I( lvk, 1) k1 N k L k, P NE = Proporion of morgage loans wih negaive equiy (NE) N k = Number of morgage loans underwrien a ime k L k, = Average loan amoun ousanding a ime for hose ha were underwrien a ime k I(, 1) lvk = Indicaor funcion for LTV raio > 1 (i.e NE) a ime for hose underwrien a ime k The direc impac will be capured by a smaller number of ime poins wih NE The indirec impac will be capured by a lower during he upcycle of propery markes. N L k k, 16
Linking he NE indicaor o defaul risk NE DSR ln( Ploan ) 0 1P 2( U (1 U ) * P ) Ploan = Delinquency and rescheduled raios for morgages a U = Unemploymen rae P DSR = Proporion of morgage loans wih DSR > 0.6 17
Decomposiion of he direc and indirec impacs 18
References: Wong, T. C., Fong, T., Li, K. F., & Choi, H. (2011). Loan-o-Value Raio as a Macroprudenial Tool - Hong Kong's Experience and Cross-Counry Evidence, HKMA Working Paper (no. 01/2011). Wong, T.C., Tsang, A., & Kong, S. (2016). How Does Loan-To-Value Policy Srenghen Banks Resilience o Propery Price Shocks Evidence from Hong Kong. Inernaional Real Esae Review, 19(1) 120-149. 19
Appendices 20
Indirec effec: An empirical sudy DD SS D 0 α 1 X 1 2r S 0 β 1 X 1 2r Q min DD, SS r 1 DD SS S D Q 0 D α 1 X 1 r 2 r 1 D where r 1 r 0, if r 1, 1 r oherwise Q 0 S β 1 X 1 r 2 r 1 S where r 1 r 0, if r 1, 1 oherwise r 21
22 Indirec effec: Model specificaion Demand equaion: where Supply equaion: where D r CNY U DSR DSR SSD r LTV ROE r LTV ROE LTV Q 1 7 6 5 4 3 2 1 0 12 10 ) ) * (, ( ), ( ) * ( 1 1 ) * * ( r LTV GPR LTV E r L GPR V ROE S r DSR DSR CD PPG r RAROC LTV Q 1 6 5 4 3 2 1 0 12 10 ) ( k c r RAROC ) ) *( 1 (
Indirec effec: Esimaion resul 23
Indirec effec: Esimaion resul 24
Conribuion of direc and indirec effecs P DSR k1 N k Lk, I( DSRk, 0.6) k1 N k L k, P NE = Proporion of morgage loans wih DSR > 0.6 N k = Number of morgage loans underwrien a ime k L k, = Average loan amoun ousanding a ime for hose ha were underwrien a ime k I(, 0.6) DSRk = Indicaor funcion for DSR raio > 0.6 a ime for hose ha were underwrien a ime k The indirec impac will be capured by a lower during he upcycle of propery markes. N L k k, 25
Conribuion of direc and indirec effecs Esimaed sressed delinquency raio for morgage loans under he acual and counerfacual scenarios 26