Macroprudential policy conference Should macroprudential policy target real estate prices? 11-12 May 2017, Vilnius Getting ready to prevent and tame another house price bubble Tomas Garbaravičius Board Member 12 May 2017 Vilnius
Content Macroprudential policy and house prices Why are house prices important? Practical challenges for policymakers Macroprudential policy in Lithuania Location-based macroprudential policy Concluding remarks 2
Macroprudential policy and house prices Policymakers claim that they do not target house prices it is important to note that the FPC has made clear that its aim in using these tools is not to control house prices. Rather, it is to mitigate the risks that a cycle of rising house prices and overextension of credit can pose to financial and macroeconomic stability. It is not the FPC s role to control house prices. The tools have not been aimed at targeting property prices. Bank of England The overall aim of macroprudential policy financial system stability cannot be mapped into a clear single target. Reserve Bank of Australia These policies do not aim at targeting property prices but may help to dampen the amplitude of property price cycles and to prevent the collateral damage that other more blunt policies might cause. Hong Kong Monetary Authority 3
Macroprudential policy and house prices Policymakers claim that they do not target house prices but they react to house price changes (Norway, Denmark, Canada) 4
Macroprudential policy and house prices Policymakers claim that they do not target house prices but they react to house price changes (Norway, Denmark, Canada) and ex-post are evaluated and/or evaluate themselves by looking at what happened to house prices. 5
WHY ARE HOUSE PRICES IMPORTANT? 6
Why are house prices important? Macroprudential policy objective is to protect the real economy from the financial system by protecting the financial system from the economy. [consumption smoothing] [focus on resilience] High indebtedness & high asset prices = more risk Burst of a house price bubble affects the economy and aggregate demand through, among other things, wealth and confidence/expectations effects higher unemployment, reduced construction activity less credit amid credit losses for lenders Real estate-related crises are longer and more severe 7
Real estate-related financial crises are longer and more severe Major financial distress episodes tend to be associated with real estate price bubbles The cost of twin (credit and real estate) crises is larger Source: Jordà, Òscar & Schularick, Moritz & Taylor, Alan M., 2015. "Leveraged bubbles," Journal of Monetary Economics, Elsevier, vol. 76(S), pages 1-20. Recoveries from leveraged property busts take longer 8
Nexus between housing market, financial sector and the economy Property market Major household wealth component Major expenditure item over the life-cycle Contribution to economic activity Serves as collateral Affects all lending conditions Credit quality of balance sheets Households Real economy Financial sector Housing acquisition is the primary reason for household borrowing 9
Economy: contribution of construction activity to GDP Share of construction value added in euro area and Lithuania % of GDP 20 18 Lithuania 16 14 12 10 8 Euro area: other buildings 6 4 2 Euro area: dwellings 0 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 Source: Eurostat. 10
Financial sector: real estate-related lending has become a dominant form of credit Òscar Jordà & Moritz Schularick & Alan M. Taylor, 2016. "The great mortgaging: housing finance, crises and business cycles," Economic Policy, CEPR;CES;MSH, vol. 31(85), pages 107-152. 11
Households: residential real estate is the most important asset for households in Lithuania Lithuania: Value of different types of wealth Share of owner-occupied dwellings across Europe Value of housing stock Cash and deposits Non-listed equities Pension assets Listed equities & investment fund units Debt securities 1 3 7 34 34 87 % of GDP 0 20 40 60 80 100 Sources: Center of Registers, Eurostat and Bank of Lithuania calculations. Romania Hungary Lithuania Slovakia Estonia Bulgaria Spain Slovenia Poland Greece Portugal Italy Iceland Ireland Cyprus Latvia Belgium United Kingdom Luxembourg Norway Malta France Czech Republic Netherlands Denmark Austria Liechtenstein Germany Sweden Switzerland Source: Eurostat. Percentages 0 20 40 60 80 100 12
PRACTICAL CHALLENGES FOR POLICYMAKERS 13
Some examples of practical challenges Timing and target specification when to act, early warning models implementation and transmission lags target indicator(s) for ex-post assessment Choice of instruments and calibration models for transmission mechanism interaction with other policies (monetary policy, taxation, structural supply-side factors) choice of instruments; maybe non-macroprudential tools could do a better job, but there is an inaction bias? Communication What if measures prove insufficient? NB: Rental costs are part of consumer price index: 6% in EA and 1% in LT 14
Some indicators may help to detect imbalances in the residential RE sector: credit-to-gdp gap Percentages 10 Household credit-to-gdp gap (calculated similarly to the standard approach of Basel definition credit-to-gdp gap ) 7 countries: DK, HU, LT, NL, ES, SE, UK 5 0 5 10 15 Interquartile range OTHER Interquartile range RE CRISIS Median OTHER Median RE CRISIS 17 countries: AT, BE, BG, HR, CY, CZ, EE, FI, FR, DE, GR, IT, MT, PL, PT, RO, SK 20 2000 2002 2004 2006 2008 2010 2012 2014 2016 Sources: BIS, Eurostat, ECB and Bank of Lithuania calculations. 15
Some indicators may help to detect imbalances in the residential RE sector: price-to-income gap Percentages 70 Price-to-income gap (deviation from the sample average of respective country) 60 50 40 30 20 10 0 10 countries: DK, HU, IE, LV, LT, NL, SI, ES, SE, UK Interquartile range OTHER Interquartile range RE CRISIS Median OTHER Median RE CRISIS 10 13 countries: AT, BE, HR, CZ, FI, FR, DE, GR, IT, MT, PL, PT, SK 20 2000 2002 2004 2006 2008 2010 2012 2014 2016 Sources: OECD and Bank of Lithuania calculations. 16
Which indicators to monitor or even target? Nominal house price growth Real house price growth Price-to-income ratio Price-to-rent ratio Real house price deviations from trend Deviations from fundamental prices estimated using econometric models: user cost of housing error correction models DSGE Simple to explain Hard to explain 17
Price-to-income ratio tends to be driven by changes in house prices Volatility of house prices and wages in Europe Q2 2000 Q2 2016 Changes in the price-toincome ratio tend to be driven by changes in house prices, since wages are less volatile 18
Price-to-income ratio: Lithuania Index: 2010=100 180 Index: 2010=100 180 160 House prices 160 140 140 120 120 100 100 80 60 Average wage 80 60 Price-to-income ratio 40 40 20 20 0 2000 2003 2006 2009 2012 2015 Source: Bank of Lithuania calculations. 0 2000 2003 2006 2009 2012 2015 Source: Bank of Lithuania calculations. 19
Price-to-income ratio: Hong Kong Source: McDonald C., "When is macroprudential policy effective?", 2014, Hong Kong Census and Statistics Department, Bank for International Settlements. 20
MACROPRUDENTIAL POLICY IN LITHUANIA 21
Borrower-based requirements in Lithuania Macroprudential measures in 2011 i) 85% LTV cap ii) 40% DSTI cap iii) loan maturity cap of 40 years Motivation i) introduction of borrower-based requirements for preventive reasons ii) broad support by banks in the aftermath of the crisis Changes in 2015 i) 50% DSTI cap with 5% interest rate test ii) 60% DSTI exception for 5% of new loans iii) loan maturity cap of 30 years Motivation i) with rates at record lows, DSTI and maturity cap do not prevent from excessive indebtedness - interest rate test is needed ii) explicit goal not to staunch household credit recovery 22
Motivation behind macroprudential decisions New housing loans and Responsible Lending Standards (RLS) Macroprudential measures in 2011 i) introduction of borrower-based requirements for preventive reasons ii) broad support by banks in the aftermath of the crisis Changes in 2015 i) with rates at record lows, DSTI and maturity cap do not prevent from excessive indebtedness - interest rate test is needed ii) explicit goal not to staunch household credit recovery 23
New concern: buy-to-let activity 24
NEW TREND AND CHALLENGE: LOCATION-BASED MACROPRU POLICY 25
Capital cities are special Share of population living in the capital city, 2014 Change in share, 2010-2014 Estonia Latvia Hungary Slovenia Finland Lithuania Belgium Bulgaria Czech Republic Switzerland Spain Slovakia Romania Poland Italy Germany Netherlands Percentages Percentage points Estonia Latvia Hungary Slovenia Finland Lithuania Belgium Bulgaria Czech Republic Switzerland Spain Slovakia Romania Poland Italy Germany Netherlands Source: Eurostat. 0 10 20 30 40 50-1 0 1 2 3 26
and urbanisation is set to continue Europe Asia North America Lithuania 100% North America 90% 80% Rural population 70% 60% 50% 40% 30% Urban population 20% 10% 0% 1950 1970 1990 2010 2030 2050 Source: United Nations. 27
Official statistics may not fully reflect demographic trends in Vilnius and other cities Thousands of inhabitants 70 60 50 40 30 Not registered in Vilnius Registered in Vilnius 14% of inhabitants are not officially registered in Vilnius, but they are registered with primary health care service providers in Vilnius 20 10 0 0-4 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 >85 Source: Municipality of Vilnius. Age group 28
Largely similar house price dynamics in Vilnius and the rest of Lithuania Index: 2010=100 200 House prices is Lithuania Vilnius 160 120 80 Non-Vilnius 40 0 1998 2001 2004 2007 2010 2013 2016 Source: Bank of Lithuania calculations. Some preliminary evidence that in about 1/3 of pairwise cases house prices in Vilnius Granger-cause house prices in other largest Lithuanian cities, whereas in the remaining 2/3 of pairwise cases there is a two-way Granger causality. 29
In Lithuania, mortgage lending activity is concentrated in Vilnius 30
But housing market in Vilnius does not stand out with respect to per capita activity Housing transactions per 1000 inhabitants Housing units / 1000 inhabitants 30 30 30 30 25 25 25 25 20 20 20 20 15 15 15 15 10 10 10 10 5 5 5 5 0 0 0 2004 2008 2012 2016 2004 2008 2012 2016 0 2004 2008 2012 2016 VILNIUS KAUNAS KLAIPĖDA Sources: State Enterprise Centre of Registries and bank of Lithuania calculations. 2004 2008 2012 2016 REST OF THE COUNTRY 31
and in relation to total housing stock Housing transactions to total housing stock ratio Transactions to total housing stock ratio, percentages 7 7 7 7 6 6 6 6 5 5 5 5 4 4 4 4 3 3 3 3 2 2 2 2 1 1 1 1 0 0 2004 2008 2012 2016 0 2004 2008 2012 2016 0 2004 2008 2012 2016 VILNIUS KAUNAS KLAIPĖDA Sources: State Enterprise Centre of Registries and bank of Lithuania calculations. 2004 2008 2012 2016 REST OF THE COUNTRY 32
Examples of location-based macroprudential policy Until July 2014, LTV of 50% in Seoul and 60% in the rest of Korea (From 2014 universal LTV of 70% and DSTI of 50%) From October 2016 higher risk weights in Vancouver and recently also in Toronto (transaction tax for foreign buyers) From January 2017, LTV of less than 60% for second home buyers in Oslo. Speed limit: 8 percent for Oslo (10 percent for the rest of Norway) Using postal code or other geographical criteria might be arbitrary an alternative could be to use price/ m 2 thresholds 33
ARE WE READY FOR ANOTHER HOUSE PRICE BUBBLE? 34
Hypothetical scenario for not so distant future in Lithuania House prices become a serious concern house price-to-income ratio has been rising for a number of years, accompanied by strong household credit growth, housing assets are the major household wealth component, and substantial potential for reduced consumption in the event of a price correction but at the same time banks are well capitalised, countercyclical capital buffer is in place, no serious supply-side constraints or distortionary tax deductions, individual household indebtedness seems manageable thanks to borrower-based requirements, and aggregate household indebtedness is still very low by international standards To act or not to act? If yes, how? If all this applies largely to Vilnius? 35