Financial cycles,

systemic risks associated with housing credit

and macroprucential rules

for mortgage market

Jan Frait

Executive Director

Financial Stability Department

XIXth European Congress of the EFBSPrague, 18 October 2019

Credit to non-financial sector and its cyclicality

2

Source: Jordá, Schularick and Taylor (2016)

Bank credit to non-financial private sector in advanced economies

Share of housing loans and its cyclicality

3

Source: Jordá, Schularick and Taylor (2016)

Share of real estate lending on total lending to non-financial sectorin advanced economies

Share of housing loans in EU economies

4

Bank loans for house purchase

Source: ESRB Risk Dashboard

Share of housing loans on bank loans in CZ

5

Structure of private sector loans by purpose

(CZK bil.)

Source: CNB

0

500

1000

1500

2000

2500

3000

3/02 3/04 3/06 3/08 3/10 3/12 3/14 3/16 3/18

NFCs Consumer loans Housing loans

Share of real estate related loans in CZ

• 60% of loans are somehow related to real estatesector.

6

Concentration of bank credit to real estate

(CZK bn.; rhs: in %)

Source: CNB

40

45

50

55

60

0

500

1000

1500

2000

2500

3000

3/07 3/09 3/11 3/13 3/15 3/17 3/19

NFC - activities in real estate

Households - housing

Other loans to private NF sector

Leve of concentration

Size of housing loans in EEA economies

7

RRE loans as a percentage of GDP (2018)

Source: ESRB

Structure of housing loans in EEA economies

8

Household debt and loan characteristics (2018)

Source: ESRB

Cyclicality of RRE prices

9

Residential real estate prices (indices)

Notes: CZ 100=2008, US 100=2008, EA 100=2007, JP 100=2010 for land

Source: BIS

60

80

100

120

140

160

180

200

0

20

40

60

80

100

120

140

1975 1979 1983 1987 1991 1995 1999 2003 2007 2011 2015

CZ US EA JP (rhs)

RRE prices growth in EU

10Source: ESRB Risk Dashboard

Change in nominal residential property prices

• Czech house prices are still „best performer“ in the EU as to 3Y growth.

11

Overvaluation of RRE prices in the EU

• ECB methods view Czech housing as apparently overvalued which is not unique in the EU.

Source: ESRB Risk Dashboard

Over/undervaluation of residential property prices

RRE prices growth and valuation in CZ

12

Transaction prices of residential property

(2015 = 100; right-hand scale: %)

Source: CZSO

-10

-5

0

5

10

15

80,0

90,0

100,0

110,0

120,0

130,0

140,0

12/08 12/10 12/12 12/14 12/16 12/18

House Price Index

Year-on-year change in index (rhs)

Estimated overvaluation of apartment prices

(%)

Source: CNB

Note: The methodology of the indicators is described in detail in

Plašil, M., Andrle, M. (2019): Assessing House Price Sustainability ,

Thematic Article on Financial Stability 1/2019, CNB.

-10

0

10

20

30

40

50

60

70

12/08 12/10 12/12 12/14 12/16 12/18

Prudential approach Valuation approach

• CNB methods deliver similar results for overvaluation.

13

Systemic risks associated with housing loans

• Banking crises are regularly associated with aftereffects of overly optimistic and loose lending to property market.

• GFC started by mortgage-related issues in UK and US, followed by widespread defaults in IE, GR, NL, ES, HU.

• Costs were shifted from borrowers to creditors in cases where large number of families were at risk of lost of home (GR, IE, HU …).

• Public finance had to step in in some cases.

• Political representations responded by instituting macroprudential policy authorities that were given mandate to set tools that should „make sure that itnever happens again“.

14

Legal risks associated with housing financing

• Misconduct fines for banks in recent years to a large extent related to housing financing (including part of payment protection insurance, PPI).

Source: The UK’s slow-burn £50bn banking scandal, Financial Times, 11/09/2019

15

Competition and a collective behaviour

• When the music stops, in terms of liquidity, things will be complicated. But as long as the music is playing, you've got to get up and dance. We're still dancing.

• Former Citigroup CEO Chuck Prince, summer, 2007.

• There was no doubt that the activities of competitors influenced decisions made by AIB…in the area of mortgage lending, … in order to protect market share, AIB felt compelled to relax its underwriting standards,… over time, in response to aggressive competition.

• Jim O’Leary, former Independent Non-Executive Director of Anglo Irish Bank. Report of the Joint Committee of Inquiry into the Banking Crisis. https://inquiries.oireachtas.ie/banking/

• We entered the 100% mortgage market only reluctantly to protect our franchise.

• Brian Goggin, former Group Chief Executive of Bank Of Ireland, ditto.

16

• Banks valued loan sales skills above credit analysis and risk… Responding to increased competition and pressure for increased earnings, banks set aggressivetargets for profit growth… this implied a partial change in business model and strategy without the corresponding necessary strengthening of governance,procedures and practices. This was accepted partly because future economic developments were trusted to remain benign in Ireland as they already had been for several years.

• The common issues (prior to the crisis) identified included: … the slow slide from lower-risk to higher-risk lending, from cash flow-lending to asset-backedlending and from small to large to enormous loan amounts…

• Nyberg Report, Misjudging Risk: Causes of the Systemic Banking Crisis in Ireland, 2011. One of the reports forthe Joint Committee of Inquiry into the Banking Crisis.

Competition and a collective behaviour

17

Macroprudential tools in the EU

• Increasing number of national macroprudentialauthorities use macroprudential mortgage limits.

• Other regulations such as amortization rules, maturity caps, affordability tests, tests for higher lending rates are used as well.

Source: ESRB

Use of LTV/DTI/DSTI in the EU economies

18

ESRB warnings and recommendations

• The European Systemic Risk Board (ESRB) published on 23 September 2019 a set of country-specific warnings and recommendations on medium-term vulnerabilities in the residential real estate sector.

• Systemic risks were identified and warnings sent to authorities in CZ, DE, FR, IC and NO.

• Insufficient response to 2016 Warnings and thus Recommendations to do more sent to authorities in BE, DK, LU, NL, FI and SE.

• CZ: … From a macroprudential perspective, the ESRB considers the main vulnerabilities to be the high overvaluation of house prices coupled with the high growth in housing credit and loose lending, against the backdrop of the non-legally-binding nature of the borrower-based measures currently available and in place.

19

Fast credit growth in some EU economies

Source: ESRB Risk Dashboard

Annual growth rates of MFIs loans to households

20

• Policy is pursued through set of recommendations: • defines prudent lending procedures and standards,

• comply or expect supervisory action principle,

• incremental approach to building up the framework,

• 1st issue in June 2015, applicable immediately,

• modified in 2016, 2017 a 2018 always in June, adjustment periods provided,

• till 1H2017 applicable to banks, branches of foreign banks and credit unions; now to any provider of housing loans or unsecured loans.

• To assess the degree of easing of lending standards, the CNB conducts twice a year a survey about characteristics of each new mortgage loan.• since 2018 the providers also report DTI and DSTI

levels, purchase prices and valuations of property purchased (not a full coverage so far).

CNB‘s approch to mortgage lending

• Household debt

21

Policy development in CZ

2015 2016 2017Low and sustainable, but growing.

Low and sustainable, but growing.

Low and sustainable, but growing.

Indicates upturn and potential divergence from fundamentals.

May not be fully grounded in fundamentals.

Pre-emptive policy

measure targeting

collateral stretch

necessary.

Signs of some

easing (volume and

share of loans with

higher LTV/LTI

ratios).

Highly relaxed,

some institutions

may be taking on

increased risks.

Above the level consistent with the fundamentals.

Policy must be modified mainly by lowering the maximum LTV levels.

Tightening of

LTV in place but

persisting

material share of

loans with higher

LTI/LSTI.

• Property prices

• Lending standards

LTV

LTI/LSTI

• ConclusionPolicy must be modified with respect to DTI/DSTI levels.

Source: CNB

Note: The spiral is derived on the basis of apartment price growth

and the amount of new loans for house purchase in relation to the

level of wages.

Spiral between apartment price growth and new loans for

house purchase in relation to the level of wages

(x-axis: y-o-y growth in apartment transaction prices in %;

y-axis: new loans in relation to wages)

22

Risk of spiralling between propertyprices and housing credit

• CNB‘s policy objective is to prevent a loop between property price growth and growth in loans posing a threat to the financial sector’s stability (mainly to banks).

LTV distribution of new loans

(x-axis: LTV in %; y-axis: share of loans in %)

Source: CNB

0

5

10

15

20

25

30

35

40

45

< 50 50–60 60–70 70–80 80–90 90–100 > 100

2015 H2 2016 H1 2016 H2 2017 H1 2017 H2

LTV distribution of new loans

(x-axis: LTV in %; y-axis: amout of loans in bil. CZK)

1.17E+11

Source: CNB

0

5

10

15

20

25

30

35

40

45

50

< 50 50–60 60–70 70–80 80–90 90–100 > 100

2014 H1 2014 H223

• Collateral standards nominally tightened, loans with restricted LTV shifted into highest allowed LTV buckets.

Known data at the time of 1st decision Rear mirror view 2017

Banking system stretch – lendingstandards: LTV

Decision Valid from to Max LTV Limit in % 10/2016 03/2017 95 85 95 10

04/2017 90 80 90 15 06/2016

Speed limit LTV range

24

• On aggregate level, institutions comply with the LTV limits since 2017 H2, although differences between individual institutions persist and some loans with LTV above 90 % are still extended.

Banks‘ complience with LTV limits

• Evidence of reduction in loan activity above limits with rising concentration close to the limits.

• (Overly?) optimistic valuations of collateral.

• Likely co-funding of some mortgages by unsecured loans from other providers outside the groups (so far evidence of small quantities only).

Fulfilment of the recommended LTV limits

(share of loans in volume provided in %)

Source: CNB

0

5

10

15

20

25

30

35

4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12

17 2Q 17 3Q 17 4Q 18 1Q 18 2Q 18 3Q 18 4Q

LTV > 90 LTV 80–90

Aggregate limit Individual limit

25

• The CNB does not currently consider it necessary to change the applicable LTV limits (the upper LTV limit of 90% and the aggregate limit of 15% for loans with an LTV of 80%–90%).

• The CNB’s current measures in the LTV area have a positive effect.

• Growth in genuinely new mortgages has slowed after the tightening of LTV limits.

• In year-on-year comparison, banks have started to take greater account of clients’ risk characteristics when setting interest rates.

• Given the current overvaluation of housing prices of roughly 15% as indicated by the CNB model, the current LTV caps represent the limit as regards sufficient coverage of potential risks by banks.

LTV limits in last two FSRs

26

LTV – one tool is not enough

• We feel very comfortable that credit quality won't be an issue at Golden West going forward. Even if housing prices drop fairly dramatically, there is plenty of room in their loan-to-value ratio. Yes.

• Former Wachovia CEO Ken Thompson, describing in 2006 the acquisition of Golden West Financial, mortgages of which ruined Wachovia.

• During the boom period, appraisers had incentives to inflate transaction prices in order to accommodate the financial needs of their clients (banks or financial brokers) and obtain future appraisal assignments.

• Montalvo, J. G., Raya, J. (2017): Constraints on LTV as a macroprudential tool: a precautionary tale. UniversitatPompeu Fabra Barcelona Working Paper no. 1592 (survey of empirical evidence).

27

• In 2016, the CNB warned banks that it regarded new loans with a DTI (debt-to-income) ratio of higher than 8 and a DSTI (debt service-to-income) ratio of higher than 40% as highly risky.

• The indicators of borrowers’ ability to service loans from their current income are implying growth in potential systemic risks.

• When limiting loans with an LTV of 80%–90%, banks did not significantly reduce loans with a DSTI > 40% and a DTI > 8.

• Borrowers are becoming more vulnerable due to faster growth in housing prices relative to income.

• The probability that riskier clients with a greater tendency to take risks will increasingly apply for loans is simultaneously rising.

Borrowers‘ stretch – lending standards:DTI/DSTI

28

• Internal rules set by banks were allowing for high LTI and LSTI.

• Volume and share of loans with high LTIs and LSTIs substantial and virtually unchanged in 2017.

LTI distribution of new loans

(x-axis: LTI in years; y-axis: share of loans in volume in %)

Source: CNB

0

10

20

30

40

50

60

≤ 3 3–6 6–8 8–9 > 9

4Q 2017 1Q 2018 2Q 2018 3Q 2018 4Q 2018

LSTI distribution of new loans

(x-axis: LSTI in %; y-axis: share of loans in volume in %)

Source: CNB

0

5

10

15

20

25

30

35

40

≤ 10 10–20 20–30 30–40 40–50 50–60 > 60

4Q 2017 1Q 2018 2Q 2018 3Q 2018 4Q 2018

Borrowers‘ stretch – lending standards:DTI/DSTI

29

Housing loans in recent years

• Banks provided a record amount of new housing loans in several consecutive years.

New housing loans and mortgage loans

(annual totals in CZK billions)

Source: CNB

0

50

100

150

200

250

300

350

400

2014 2015 2016 2017 2018

New loans for house purchase

New mortgage loans

New mortgage loans excluding refixations and refinancing

30

• The CNB therefore issued a recommendation in June 2018 that banks take account of loan applicants’ income as of 1 October 2018 and comply with: • an upper DTI limit of 9 times the applicant’s net

annual income,

• an upper DSTI limit of 45% of the applicant’s net income.

• The warning to banks regarding risky DTI and DSTI levels remains in effect.

• The CNB respects the fact that a small proportion of loans have specific characteristics and that strict insistence on the application of the caps could lead to excessive regulatory hardship.

• It is therefore allowing lenders to cover specific cases under a 5% exemption.

Last mortgage lending measures: DTI and DSTI

31

Potential impact of DTI and DSTI limits

• DTI at 9 multiple does not represent a major constraint.

Fulfilment of the recommended DTI limits

(share of loans in volume provided in %)

Source: CNB

0

2

4

6

8

10

12

14

16

18

7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12

17 Q3 17 Q4 18 Q1 18 Q2 18 Q3 18 Q4

LTI > 9 DTI > 9

DTI distribution of new loans

(x-axis: DTI in years; y-axis: share of loans in volume in %)

Source: CNB

Note: Relative to the volume of loans provided a quarter earlier.

0

10

20

30

40

≤ 3 3–6 6–8 8–9 > 9

2018 Q3 2018 Q4

32

Potential impact of DTI and DSTI limits

• The loans with DSTI limit above 45% were used on excessive scale in 2H 2018.

DSTI distribution of new loans

(x-axis: DSTI in %; y-axis: share of loans in volume in %)

Source: CNB

Note: Relative to the volume of loans provided a quarter earlier.

0

5

10

15

20

25

30

≤ 10 10–20 20–30 30–40 40–45 > 45

2018 Q3 2018 Q4

Fulfilment of the recommended DSTI limits

(share of loans in volume provided in %)

Source: ČNB

0

5

10

15

20

25

30

35

7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12

17 Q3 17 Q4 18 Q1 18 Q2 18 Q3 18 Q4

DSTI > 45 LSTI > 45

Venn diagram for selected group of loans in Q4/18

(in % of volume)

Source CNB

33

Riskiness of mortgages extended in 2018

• Number of loans had simultaneously high ratios in Q4/2018.

Distribution of mortgages with LTV > 80%

(in % of volume)

Source: CNB

0

2

4

6

8

10

12

14

2018-3 2018-4

LTV > 80 %LTV > 80 % and DSTI > 40 %LTV > 80 %, DSTI > 40 % and DTI > 8

34

Impact of DTI/DSTI limits on mortgage provision

• Sharp rise in 2H 2019 followed by sharp drop in 1Q 2019.

• CNB expects the impact on volumes from DTI/DSTI (ceteris paribus) of roughly 10%.

Monthly volumes of genuinely new mortgages(CZK bn.)

Source: CNB

0

5

10

15

20

25

1-14 4-14 7-14 10-14 1-15 4-15 7-15 10-15 1-16 4-16 7-16 10-16 1-17 4-17 7-17 10-17 1-18 4-18 7-18 10-18 1-19 4-19 7-19

Monthly volumes of genuinely new mortgages Monthly average since 2015

35

Concluding remarks

• Forward-looking action key to policy credibility.

• Necessary to respond early once potential risks move beyond certain point (before unwelcome practices become “new normal” or generally accepted standard).

• Effectiveness of current tools will show up in next recession only.

• CNB activated most of its tools in good times when banks had plenty of room for expansion in terms of capital and caps provided limited constraint.

• Mandatory tools should improve effectiveness and create a level playing field.

• The CNB and the Ministry of Finance await a joint proposal for binding LTV, DTI and DSTI limits to be discussed by the Parliament soon.

THANK YOU FOR YOUR ATTENTION

www.cnb.cz

Jan Frait

Financial Stability Reports and other documentsavailable at https://www.cnb.cz/en/financial-stability/

Financial Stability DepartmentCzech National Bank

Na Prikope 28CZ-11503 Prague