Mike Fahy1, Kieran McQuinn2, Conor O’Toole2 and Rachel Slaymaker2

Department of Finance1 and Economic & Social Research Institute2

14th June 2018

The Financial Crisis and the Changing Profile of Mortgage Arrears in Ireland

The analysis and views set out in this paper are those of the authors

only and do not necessarily reflect the views of the Department of

Finance or the Minister for Finance, or the Economic and Social

Research Institute.

Disclaimer

2

This work on “the Changing Profile of Mortgage Arrears in Ireland” is part of theJoint Research Programme between the Department of Finance and the ESRI.

The Research is structured in two phases:i. Phase 1: To address a gap in the literature, and shed some light on the

composition of mortgage arrears in Ireland – we explore whetherthe composition of borrowers who went into arrears during thefinancial crisis period was different from those who were inarrears in the pre-crisis period.

ii. Phase 2: We are developing a new household mortgage stress testing modelto assess the triggers of mortgage default in Ireland. Channelsinclude: the interest rate, income (affordability) channels andhouse price (equity) channel.

Motivation:Understanding which households go into mortgage arrears during both boomand bust periods, is of critical importance to ensure suitable policies aredeployed to safeguard future financial stability.

Introduction

3

Irish research:1. Tsiropoulos (2018) - calculate a vulnerability index for currently performing

mortgages under a certain adverse scenario which involves movements in the unemployment rate, house prices and interest rates

2. Byrne et al. (2017) - highlight the interest rate channel on default3. Kelly and O'Malley (2016) - develop an arrears transitions model that is used in

mortgage stress testing by CBI and builds on dual trigger of mortgage default4. Kelly and McCann (2015) –using CBI Loan Level Data and Standard Financial

Statement 2012 & 2013, show that high LTV ratios, consumer credit growth, shocks to mortgage affordability and unemployment contribute to mortgage defaults

5. Ireland - McCarthy (2014) - shows that unemployment and income shocks determine arrears rate in Ireland using novel dataset collected by the Central Bank of Ireland. This is an important study as it includes the up to date information on income (current income)

6. Lydon and McCarthy (2011) - use CBI Loan-Level Data to look at equity and income shocks. Find a considerable role for both channels but unemployment was only proxy for the income shocks

7. McCarthy and McQuinn (2011) - focus on income in the period just before the severe crisis period 2009/2010.

To date - no research has documented the change in mortgage arrears over the crisis period in terms of household composition and other characteristics - mainly due to the lack of loan-level data for the pre-crisis period.

Previous Research

4

To examine mortgage arrears, we draw on the Survey on Income and Living Conditions (SILC) data set from 2004 to 2015 undertaken by the CSO

The value of SILC - is that it has rich demographic data which is not available (in for example the CBI’s loan level data), or any other any other data set.

It is also the only source of arrears data in Ireland during pre-crisis period

SILC contains household characteristics: age,

marital status,

education,

household composition,

health status,

income,

poverty indicators (ability to save, ability to make ends meet), and

regions in Ireland (at NUTS 3 level)

Enables us to explore the difference in the type of households that went into arrears during the systemic crisis period relative to the pre-crisis period

5

Analysis based on use of SILC Data

Data: How well does SILC measure PDH Mortgage Arrears?

6

0

2

4

6

8

10

12

14

16

18

20

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

%

SILC data (population weighted) - % of PDH loans accounts in arrears (in the last 12 months)

CBI data - % of total PDH loan accounts in arrears outstanding (average across 4 quarters)

CBI data - % of total PDH loan accounts in arrears outstanding for more than 90 days (average across 4 quarters)

Correlation coefficient 0.91

How well does SILC measure Negative Equity?

7

0

5

10

15

20

25

30

35

40

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

%

SILC data (t-1 and population weighted) - % of PDH loans accounts in negative equity

CBI data - % of total PDH loan accounts in negative equity

Correlation coefficient 0.99

8

Share of Households in Arrears Pre and Post Crisis

5-83-51-3

%

2004-2009 2010-2013

9

The relationship between Mortgage Arrears & Unemployment, by Region (2004-2009 vs 2010-2013)

Post crisis years – greater dispersion between regions and a high correlation between high level of arrears and high unemployment rate

Bord

Mid

West

Dub

ME

MWSE

SW

Bord

Mid

WestDub

ME

MW

SE

SW

0.0

5.1

.15

.2Sh

are

of H

ouse

hold

s in

Arr

ears

5 10 15 20% Unemployment

2004-2009 2010-2013

10

The relationship between Mortgage Arrears & Negative Equity, (2004-2015)

Post crisis years – a high correlation between high level of arrears and negative equity

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Shar

e o

f H

ou

seh

old

s in

Ne

gati

ve E

qu

ity

No Arrears Arrears

11

Bord

Mid

West

Dub

ME

MW

SE

SW

.1.1

2.1

4.1

6.1

8.2

Shar

e of

Hou

seho

lds

in A

rrea

rs

.2 .3 .4 .5Share of Households in Negative Equity

The relationship between Mortgage Arrears & Negative Equity, by Region (2010-2013)

Post crisis years – a high correlation between high level of arrears and negative equity, and greater dispersion between regions

12

Table 1: Mortgage Arrears by Household Socioeconomic Characteristics, 2004-2013

CharacteristicPredicted Probability

of Arrears 2004-2009

Predicted Probability of

Arrears 2010-2013

Percentage Point

Difference

Percentage

Change (%)

Age

18-35 0.015 0.136 0.121 806.7

36-50 0.017 0.083 0.066 388.2

51+ 0.02 0.104 0.084 420.0

Marital Status

Married 0.014 0.09 0.076 542.9

Single 0.021 0.11 0.089 423.8

Divorced/Separated/Widowed 0.04 0.169 0.129 322.5

Employment Status

Employed 0.013 0.086 0.073 561.5

Unemployed 0.127 0.216 0.089 70.1

Other 0.031 0.13 0.099 319.4

Education

Primary 0.025 0.188 0.163 652.0

Secondary 0.024 0.108 0.084 350.0

Tertiary 0.01 0.089 0.079 790.0

Household Composition

1 adult, no children 0.011 0.048 0.037 336.4

1 adult, with children 0.086 0.105 0.019 22.1

2 adults, no children 0.01 0.069 0.059 590.0

2 adults, with children 0.022 0.141 0.119 540.9

Health Status

Very good/good 0.015 0.091 0.076 506.7

Fair 0.045 0.19 0.145 322.2

Bad/very bad 0.145 0.308 0.163 112.4

Income Distribution

<p(40) 0.041 0.28 0.239 582.9

p(40)-p(60) 0.029 0.153 0.124 427.6

>p(60) 0.013 0.069 0.056 430.8

LTV > 80 0.020 0.123 0.103 515.0

Unconditional Proportion of Households in Arrears (%) 3.2 13.8 10.6 331.3

No. Obs. 6,633 4,334

13

Table 2: Mortgage Arrears by Household Socioeconomic Characteristics, 2004-2013

Characteristic

Predicted

Probability of

Arrears

2004-2009

Predicted Probability

of Arrears

2010-2013

Percentage Point

Difference

Percentage

Change (%)

Utilities Arrears in last 12 months

Yes 0.116 0.348 0.232 200.0

No 0.006 0.096 0.09 1500.0

Consumer Loan Arrears in last 12 months

Yes 0.061 0.448 0.387 634.4

No 0.007 0.104 0.097 1385.7

Can Regularly Save Some Income

Yes 0.006 0.115 0.109 1816.7

No 0.01 0.141 0.131 1310.0

Ability to Make Ends Meet

With great difficulty/difficulty 0.025 0.19 0.165 660.0

With some difficulty 0.008 0.101 0.093 1162.5

Fairly easily/easily 0.004 0.101 0.097 2425.0

Very easily - - - -

Ability to take Annual Family Holiday

Yes 0.006 0.113 0.107 1783.3

No 0.015 0.157 0.142 946.7

No. Obs. 6,632 4,332

Results show that from a regional perspective;

In pre-crisis there wasn’t significant geographical variation in the proportion of

households in arrears, however, in the post-crisis years mortgage arrears rates were

significantly higher in the Border, Midlands and South-East regions.

taking a closer look at the equity and unemployment channels which act as a catalyst

for a household falling into arrears, results show that the relative importance of these

channels differs across regions.

In addition results show that

prior to the crisis (2004-2009), households in mortgage arrears were typically those

who had suffered unemployment, marital and health shocks, as well as particular risk

groups such as lone parents.

However, during the systemic crisis (2010-2013), the risk of falling into arrears spread

throughout society, and in fact healthier, younger, more highly educated, employed

households saw the largest increases in incidence of arrears.

Intention to publish - “The Financial Crisis and the Changing Profile of Mortgage Arrears in

Ireland” as a research note in the ESRI’s Quarterly Economic Commentary

Summary

14

Phase 2: We are developing a household mortgage stress testing model that can look at shocks to interest rates, incomes (net after tax) and house prices.

Motivation for phase 2 research:o Mortgage loans outstanding are the largest asset class held by Irish banks and

thus credit risk vulnerabilities that arise through economic shocks propagate quickly through into loan delinquencies problems for the banking sector.

o Within this context, it is important to understand the main drivers of mortgage default and to be able to disentangle the factors which lead to households missing loan payments (i.e. the factors which trigger mortgage loan defaults)

Policy scenarios - Model can be used to undertake a number of scenarios, for example impact of interest rates, income, and house price channel shocks

Next Steps

15

Thanks

-Any Questions ?

16