HOUSEHOLDS LENDING MARKET IN THE ENLARGED EUROPE Debora Revoltella and Fabio Mucci copyright with the author New Europe Research ECFin Workshop on Housing and mortgage markets and the EU economy, Brussels, 21st November 2005
AGENDA Stylized facts Liquidity constraints and the role of credit Sustainability and risks of retail lending growth Conclusions 2
A SNAPSHOT OF HOUSEHOLD FINANCIAL BEHAVIOUR WITH SPECIAL FOCUS ON THE LIABILITY SIDE Year 2004 data Eurozone New Europe* US Gross Disposable income per capita, 15,871 (1) 3,178 23,617 Households financial wealth over GDP 182% (2) 53% Household liabilities over GDP 50% 12% Home ownership 70% (5) 69% 277% 124% 69% 3 Note: (1) As of 2003; excluding Ireland and Luxembourg; (2) Proxy including all eurozone countries with the exception of Luxembourg, Ireland, Greece and Austria; unquoted shares, other equity and other accounts have been excluded from total wealth (3) Eurozone 2001; (*) New Europe includes 8 CEECs new EU members, plus other candidate or EU approaching countries (I.e. Bulgaria, Romania, Croatia and Turkey). Malta, Lithuania, Slovenia and Cyprus are excluded. Source: UCI New Europe Research Network, ECB, OECD, Federal Reserves and US Census Bureau
EUROZONE VS NEW EUROPE: SAME PATTERN OF FAST GROWTH OF RETAIL LENDING MARKET, BUT FROM VERY DIFFERENT STARTING POINTS Household Indebtedness [as a percentage of GDP] Household debt growth in nominal terms [CAGR 00-04] Bank credit to individuals [ weight in tot. banking sector] 46% 50% 23% 38% 43% 45% 23% 7% 12% 7% 2000 2004 2000 2004 2000 2004 2000 2004 NE(10) Eurozone NE(10) Eurozone NE(10) Eurozone Very different market size and level of households indebtedness Similar trends in terms of accelerated growth Increasing relevance of retail lending over total bank lending (meaning increasing banking focus towards the retail sector) Source: UCI New Europe Research Network Database, based on National Central Banks and Eurostat 4
RETAIL LENDING DRIVEN BY THE MORTGAGE MARKET, STILL WITH FAST GROWTH IN CONSUMER LENDING Household Indebtedness by type of instruments [as a percentage of GDP] 50% 46% 9% 9% 12% 7% 7% Other loans (1) 7% 2% 3% 5% 29% 34% Consumer loans Mortgage 4% 4% 1% 2000 2004 NE(10) 2000 2004 Eurozone Average annual mortgage growth in NE equal to 43%, vs 8% in Eurozone Average annual consumer lending growth equal to 12% in NE, vs 2% in Eurozone Note: (1) In the New Europe, the aggregate other loans includes overdraft, revolving credit cards, other loans and financial leasing Source: UCI New Europe Research Network Database and ECB 5
MARKET STRUCTURE AND GROWTH BY COUNTRY Prevalence of consumer credit Prevalence of mortgages (Mortgage growth - Consumer loans growth)* 50% 40% Hungary 30% Poland Latvia 20% Bulgaria Estonia Austria Belgium 10% Turkey Romania Czech R. Spain Croatia Italy France Portugal 0% Germany Greece -10% The Netherlands Finland -20% -30% Slovakia -40% -50% 0% 20% 40% 60% 80% 100% Mortgage over total loans *CAGR 04-00 Prevalence of consumer credit 6 Prevalence of mortgages Note: for Slovakia, loans granted by non-bank financial institutions excluded from total individuals liabilities while for the Czech R. CAGR 04-01 Source: UCI New Europe Research Network Database and National Central Banks for Eurozone countries
A CLUSTER ANALYSIS FOCUSED ON CYCLICAL FEATURES OF RETAIL LENDING MARKET PROVIDES FIRST SIGNALS OF CONVERGENCE 7 Bulgaria Romania Croatia Poland Turkey Czech R. Slovakia Hungary Estonia Latvia Finland Italy Greece Austria Germany Ireland Spain Portugal Belgium France Netherlands 0 5 10 15 20 Cluster based on: debt over GDP (2004) weight of mortgages in total loans (2004) real credit growth change in fin.wealth Gap 25 30 35 Source: UCI New Europe Research Network
ALTHOUGH A CLEAR GAP REMAINS WHEN TAKING INTO ACCOUNT THE DEGREE OF MARKET DEVELOPMENT IN THE TWO REGIONS Gap 0 10 20 30 40 50 60 70 80 90 Bulgaria Romania Croatia Czech R. Hungary Polond Estonia Latvia Slovakia Greece Turkey Austria Germany Portugalo Finland Spaina Italy Ireland Belgium France Netherlands Cluster based on: real GDP growth change in unemployment rate per capita GDP fin. wealth over GDP debt over GDP (2000) change in interest rates weight of mortgages in total loans real credit growth change in fin.wealth 100 8 Source: UCI New Europe Research Network
AGENDA Stylised facts Liquidity constraints and the role of credit Sustainability and risks of retail lending growth Conclusions 9
HIGHER VOLATILITY OF CONSUMPTION RELATIVE TO INCOME GROWTH PROVIDES FIRST EVIDENCE OF LIQUIDITY CONSTRAINTS, PARTICULARLY IN NE COUNTRIES Consumption and real GDP volatility and Correlation between consumption and credit Coeff. of variation of real consumption Coeff. of variation of real GDP Corr(cons, credit) Bulgaria 4.76% 3.39% 0.29 Czech Rep 2.64% 1.76% 0.06 Croatia 1.28% 1.14% 0.24 Estonia 3.00% 1.94% -0.15 (0.08 three lags) Hungary 1.85% 0.40% 0.65 Latria 1.80% 1.37% 0.2 Poland 1.18% 1.59% 0.21 Romania 4.83% 2.79% -0.37 (0.51 two lags) Slovakia 1.98% 1.01% -0.14 (0.11 one lag) Turkey 4.76% 3.90% 0.85 simple avg NE 2.81% 1.93% - Belgium 0.77% 0.75% 0.05 Germany 0.87% 0.76% -0.42 Spain 1.00% 0.54% 0.34 France 0.67% 0.72% 0.46 Greece 0.67% 1.00% - Italy 0.68% 0.71% -0.07 Netherlands 1.19% 0.92% -0.01 Portugal 1.09% 1.08% 0.08 simple avg EU 0.87% 0.81% - Note: own estimates based on UCI New Europe Research Database, ECB and Eurostat 10
ECONOMIC ANALYSIS OF CONSUMPTION GROWTH AND HOUSEHOLDS DEBT: A SIMPLE FRAMEWORK 3 GROUPS OF HOUSEHOLDS: (1-λ 1 -λ 2 ) not liquidity constrained - can choose their optimal consumption path, on the basis of permanent income, having free access to credit markets (λ 1 +λ 2 ) liquidity constrained, thus consumption is a function of current income Among those liquidity constrained, λ 2 can access credit markets Δc t = β + (1-λ 1 -λ 2 ) σ E t-1 r t-1 + (λ 1 +λ 2 ) Δy t + λ 2 ΔHC t CREDIT EQUATION ΔHC t = a bi 1 t-1 + c (i1 * Δy) t-1 CONSUMPTION EQUATION Δc t = β + (λ 1 +λ 2 ) Δy t + λ 2 (a-bi 1 t-1 + c (i1 * Δy) t-1 ) 11
EMPIRICAL ANALYSIS FOR OLD EUROPE PROVIDES EVIDENCE THAT IN ITALY, BELGIUM AND UK ABOUT 50% OF THE CONSUMERS ARE LIQUIDITY CONSTRAINED (DE BONDT 1999)* Developed credit market should allow a smoothing of consumption patterns of individuals, in line with the life cycle theory Excess sensitivity of consumption to current income is frequently interpreted as evidence of liquidity constraints, that is the inability or unwillingness to use capital markets to smooth consumption The evidence provided by De Bondt shows that in most EU countries 80% to 100% of all consumers are liquidity unconstrained with the only exceptions of Italy, United Kingdom and Belgium where about 50% of the consumers are liquidity constrained Moreover, the empirical analysis tend to confirm that variations in credit market conditions as reflected by changes in the so-called external finance premium (the wedge between cost of internal and external finance connected to informal asymmetries, cost of borrowers monitoring etc) significantly impact private consumption growth in the case of Germany, Italy and the Netherlands De Bondt (1999), Credit channels and consumption in Europe: empirical evidence, BIS wp no. 69 12
LIQUIDITY CONSTRAINTS PLAY A KEY ROLE IN THE NE (CORICELLI, MUCCI AND REVOLTELLA 2005) By estimating both consumption and supply equations on quarterly data for a sample of 10 New Europe countries over the period 2000-2004, the analysis reveals that: the effect of household credit conditions on individual consumption is highly significant. As expected tightening of credit conditions have larger effects during a slow-down, rather than during an upturn; supply effects have been very relevant in explaining the dynamic of consumer credit during the sample period. Moreover, the estimation results suggest that the value of the share of liquidity-constrained households (λ 1 +λ 2 ) appears to be large at 53%. Among the liquidity constrained individuals, only 5% have access to credit. Important to note: the total share of households with access to consumer credit includes as well those households that are not liquidity-constrained. Coricelli, Mucci and Revoltella (forthcoming), The New Europe household lending market 13
PROSPECTS: SIMULATION OF DEMAND AND SUPPLY EFFECTS Int. rate Credit demand Credit supply Relaxation of supply conditions in the context of liquidity constraints behind accelerated growth of household lending in most NE and some old EU countries Interest rates decline potentially partly balanced by upward movements in the demand curve (with enlargement of clients pool) Credit 14
AGENDA Stylised facts Liquidity constraints and the role of credit Sustainability and risks of retail lending growth Conclusions 15
RELATIVE LOW INDEBTEDNESS VS THE US. IN THE NE CONCERNS FOR FAST GROWTH Household debt as a percentage of GDP (2004) 64% 46% 50%53% 16 Source: UCI New Europe Research Network Database, ECB, National Central Banks for Eurozone countries and Eurostat 5% 6% 10% 12%12% 13% 14% 19%20% 24% 25% 31% 32%35% 40%40% 40% 78% 124% Portugal Romania Turkey Slovakia NE(10) Bulgaria Poland Czech Rep. Latvia Hungary Estonia Italy Greece Croatia Belgium France Austria Finland Germany Eurozone Spain Netherlands US
AND HIGHER VULNERABILITY TO ADVERSE SHOCKS GIVEN THE LOW LEVEL OF WEALTH Data as of 2004 Household Debt over Fin. Wealth Per capita debt Estonia 83% Eurozone 13,286 Croatia 46% Croatia 1,939 Hungary 36% Hungary 1,603 Romania 31% Estonia 1,587 Latvia 29% (*) Czech R. 1,217 Bulgaria 28% Latvia 852 Eurozone 27% Poland 740 Poland 24% Slovakia 663 NE(10) 23% NE(10) 540 Slovakia 22% Bulgaria 300 Czech R. 21% Turkey 210 Turkey 13% Romania 143 Note: (*) As of end of 2003. Source: UCI New Europe Research Network Database, based on National Central Banks and Eurostat 17
ACCUMULATION OF CORRECTED NET FINANCIAL WEALTH PROVIDES A POSITIVE SIGNAL IN THE NEW EUROPE AND IN SOME EUROZONE COUNTRIES 350% Corrected (2) Net Financial Wealth 2000-2004 300% 250% 2000 2004 200% 150% 100% 50% 0% Bulgaria Croatia Czech R. Estonia Hungary Latvia(1) NE(10) Poland Romania Slovakia Turkey Eurozone Belgium Germany Spain France Italy Netherlands Portugal Finland 18 Note: (1) Latvia as of 2003; (2) Corrected Net Financial Wealth is calculated as Gross Financial Wealth minus the non-mortgage part of households debt. All divided by GDP. Source: UCI New Europe Research Network Database, Natianal Banks, ECB and Eurostat
NO EXCESSIVE GROWTH IN HOUSING PRICES SO FAR DESPITE THE POSITIVE CORRELATION BETWEEN DEMAND FOR MORTGAGES AND PROPERTY PRICES Mortgage and House Prices Levels in the Eurozone Mortgage and House Prices Levels in the New Europe 16 s e c i r p e s u o H 14 12 10 8 6 4 2 0-2 -4 HPRIC E = 0.36 HLOAN + 3.229 0 10 20 30 40 Mortgage loans Housing loans s e r ic p e s u o H 100 80 60 40 20 Mortgage loans y = 1.19x + 32.3 0-10 -5 0 5 10 15 20 Housing market exposure represents a source of vulnerability in the context of fast expansion of debt and rising house prices Much debate concerns single eurozone countries In the NE, despite the increasing demand for housing connected to expectations of rising incomes in the catching-up with EU, better provision of housing credit and current demographic trends we do not see signals of potential real estate bubbles given the still limited supply of new homes Source: UCI New Europe Research Network estimates on Eurozone countries and Mihaljek D. (2005) Rapid expansion of credit in South Eastern Europe: a cause for concern?, BIS for NE countries 19
IN THE NE SOME CONCERNS ON THE MACRO SIDE 6,000 3,000 0-3,000-6,000 Romania Turkey Latvia 20% 4.0 20% 1.5 trade balance/gdp - RX credit cycle 1.0 2.0 factor2 trade balance/gdp - RX credit cycle factor2 2000q01 2000q03 2001q01 2001q03 2002q01 2002q03 2003q01 2003q03 2004q01 2004q03 0% -20% 0.0-2.0-4.0 2000q01 2000q03 2001q01 2001q03 2002q01 2002q03 2003q01 2003q03 2004q01 2004q03 0% -20% 0.5 0.0-0.5-1.0-1.5 trade balance/gdp - RX credit cycle factor2 2000q01 2000q03 2001q01 2001q03 2002q01 2002q03 2003q01 2003q03 2004q01 2004q03 40% 20% 0% -20% -40% 8.0 6.0 4.0 2.0 0.0-2.0-4.0-6.0 Estonia trade balance/gdp - RX credit cycle factor2 55% 30% 5% -20% -45% 20.0 10.0 0.0-10.0-20.0-30.0 Croatia trade balance/gdp - RX credit cycle factor2 30% 15% 0% -15% -30% -45% 0.4 0.2 0.0-0.2 Bulgaria trade balance/gdp - RX credit cycle factor2 60% 30% 0% -30% 2000q01 2000q03 2001q01 2001q03 2002q01 2002q03 2003q01 2003q03 2004q01 2004q03 2000q01 2000q03 2001q01 2001q03 2002q01 2002q03 2003q01 2003q03 2004q01 2004q03 2000q01 2000q03 2001q01 2001q03 2002q01 2002q03 2003q01 2003q03 2004q01 2004q03 20 Note: the series on real credits to individuals and trade balance have been seasonality adjusted using the census X11 additive method. The trend have been extracted using Hodrick and Prescott (HP) filter with a smoothing parameter of 1600. Factor 2 represents in all graphs the absolute threshold used to identify credit booms episodes (1.75 times std of the cycle). Consumption booms have been identified using the same criteria and are represented in all graphs by grey areas. Sources: Trade data: IMF- DOTS based on country s declarations of exports and imports and Eurostat. Credit Data: UniCredit New Europe Research Network database.
AGENDA Stylised facts Liquidity constraints and the role of credit Sustainability and risks of retail lending growth Conclusions 21
CONCLUSIONS Similar trends in the retail lending market, although NE starts from a much lower base. Fast growth, driven by the mortgage market, with decrease of liquidity constraints and improvement in supply conditions Evidence of liquidity constraints in some old EU and most of NE countries. Relaxation of constraints and increasing access to credit markets explain most of the current accelerated growth of household credit No major concerns for sustainability at the enlarged European level, but a number of issues to monitor Micro/Macro issues in few NE countries Real estate market pressures in the Eurozone 22