Future Trends in Housing Tenure and the Adequacy of Retirement Income

FUTURE TRENDS IN HOUSING TENURE AND THE

ADEQUACY OF RETIREMENT INCOME

RACHEL SLAYMAKER, BARRA ROANTREE, ANNE NOLAN AND

CONOR O'TOOLE

RESEARCH

SERIES

NUMBER 143

JULY 2022

FUTURE TRENDS IN HOUSING TENURE AND THE

ADEQUACY OF RETIREMENT INCOME

Rachel Slaymaker

Barra Roantree

Anne Nolan

Conor O’Toole

JULY 2022

RESEARCH SERIES

NUMBER 143

Available to download from www.esri.ie

 The Economic and Social Research Institute

Whitaker Square, Sir John Rogerson’s Quay, Dublin 2

https://doi.org/10.26504/rs143

This Open Access work is licensed under a Creative Commons Attribution 4.0 International License

(https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and

reproduction in any medium, provided the original work is properly credited.

ABOUT THE ESRI

The mission of the Economic and Social Research Institute is to advance evidence-

based policymaking that supports economic sustainability and social progress in

Ireland. ESRI researchers apply the highest standards of academic excellence to

challenges facing policymakers, focusing on 12 areas of critical importance to 21st

Century Ireland.

The Institute was founded in 1960 by a group of senior civil servants led by

Dr T.K. Whitaker, who identified the need for independent and in-depth research

analysis to provide a robust evidence base for policymaking in Ireland.

Since then, the Institute has remained committed to independent research and its

work is free of any expressed ideology or political position. The Institute publishes

all research reaching the appropriate academic standard, irrespective of its

findings or who funds the research.

The quality of its research output is guaranteed by a rigorous peer review process.

ESRI researchers are experts in their fields and are committed to producing work

that meets the highest academic standards and practices.

The work of the Institute is disseminated widely in books, journal articles and

reports. ESRI publications are available to download, free of charge, from its

website. Additionally, ESRI staff communicate research findings at regular

conferences and seminars.

The ESRI is a company limited by guarantee, answerable to its members and

governed by a Council, comprising 14 members who represent a cross-section of

ESRI members from academia, civil services, state agencies, businesses and civil

society. The Institute receives an annual grant-in-aid from the Department of

Public Expenditure and Reform to support the scientific and public interest

elements of the Institute’s activities; the grant accounted for an average of 30 per

cent of the Institute’s income over the lifetime of the last Research Strategy. The

remaining funding comes from research programmes supported by government

departments and agencies, public bodies and competitive research programmes.

Further information is available at www.esri.ie

THE AUTHORS

Rachel Slaymaker and Barra Roantree are Research Officers at the Economic and

Social Research Institute (ESRI). Anne Nolan and Conor O’Toole are Associate

Research Professors at the ESRI. All hold adjunct positions at Trinity College Dublin

(TCD).

ACKNOWLEDGEMENTS

The research was funded by the Pensions Council under a Research Programme on

‘Housing and Pensions’. The authors thank the Central Statistics Office (CSO) for

access to Survey on Income and Living Conditions (SILC) microdata, and the Irish

Longitudinal Study on Ageing (TILDA) for access to the TILDA microdata. Thanks

also to the members of the Pensions Council, and to the members of the Research

Programme Steering Committee (Ciaran Diamond, Des Henry, Brendan Kennedy,

Munro O’Dwyer) for helpful comments and guidance throughout the period of this

research programme. We are grateful to participants at an ESRI lunchtime seminar,

and to the external and internal reviewers for helpful comments. All views, errors

and omissions remain the sole responsibility of the authors.

This report has been accepted for publication by the Institute, which does not itself take institutional

policy positions. All ESRI Research Series reports are peer reviewed prior to publication. The authors

are solely responsible for the content and the views expressed.

Table of contents | v

TABLE OF CONTENTS

ABBREVIATIONS .................................................................................................................................. VII

EXECUTIVE SUMMARY ......................................................................................................................... IX

CHAPTER 1 INTRODUCTION ................................................................................................................ 1

1.1 Context ................................................................................................................ 1

1.2 Existing evidence ................................................................................................. 3

1.3 Contribution and report structure....................................................................... 7

CHAPTER 2 POLICY CONTEXT .............................................................................................................. 9

2.1 The Irish pension system ..................................................................................... 9

2.2 The Irish housing market ................................................................................... 11

CHAPTER 3 HOUSING TENURE AND COST SCENARIOS ..................................................................... 15

3.1 Housing tenure .................................................................................................. 15

3.2 Current housing costs ........................................................................................ 18

3.3 Homeownership simulation scenarios .............................................................. 20

3.4 Summary ............................................................................................................ 31

CHAPTER 4 INCOME ADEQUACY IN RETIREMENT ............................................................................ 33

4.1 Introduction ....................................................................................................... 33

4.2 Data and methodology ...................................................................................... 33

4.3 Results and empirical findings ........................................................................... 35

4.4 Summary ............................................................................................................ 40

CHAPTER 5 CONCLUSIONS ................................................................................................................ 43

5.1 Summary of main findings ................................................................................. 43

5.2 Policy implications ............................................................................................. 45

REFERENCES ........................................................................................................................................ 49

APPENDIX I SUPPLEMENTARY RESEARCH FOR CHAPTER 3.................................................................

APPENDIX II SUPPLEMENTARY RESEARCH FOR CHAPTER 4 ............................................................... 55

vi | Future trends in housing tenure and the adequacy of retirement income

LIST OF TABLES

Table 3.1 What % of non-transitioners are constrained by different factors? ................................ 24

Table 3.2 Homeownership rate simulations by age cohort (%) – additional scenarios ................... 26

Table 3.3 Homeownership rate simulations for 25-34 cohort (%) .................................................. 28

Table 4.1 Poverty line (AROP) measures of income adequacy ........................................................ 36

Table 4.2 Poverty gap for various measures of income adequacy .................................................. 38

Table 4.3 Logistic regression of the factors associated with AROP status, baseline simulations,

broader income measure ................................................................................................. 40

Table A.1 Evidence of likelihood of receiving gift / amount by recipient characteristics ................ 53

Table A.2 25th percentile house price by NUTS 3 region – 2019 ..................................................... 53

Table A.3 Logistic regression of the factors associated with AROP, baseline simulations, narrow

income measure ............................................................................................................... 55

LIST OF FIGURES

Figure 1.1 Homeownership rates by decade of birth cohort .............................................................. 2

Figure 2.1 Trends in housing costs (year-on-year change, %) ........................................................... 12

Figure 2.2 Tenure shares Ireland ....................................................................................................... 13

Figure 2.3 Housing supply changes – ESB connections – overall and per 1,000 persons

aged 20-50 ........................................................................................................................ 14

Figure 3.1 % Homeowners by age cohort 1987-2015 ....................................................................... 16

Figure 3.2 % Households in housing tenures by age cohort 2018-2019 ........................................... 17

Figure 3.3 Mean housing cost-to-income ratio (HCTI) % by tenure and age cohort ........................ 18

Figure 3.4 % Households with high housing costs (HCTI>30%) by age cohort .................................. 19

Figure 3.5 Overview of static tenure assignment method ................................................................ 21

Figure 3.6 % Homeowners after static tenure transition, by age cohort .......................................... 23

Figure 3.7 % Homeowners 25-34 cohort – different house price and income growth scenarios .... 29

Figure A.1 Distribution of after housing costs income, by scenario (broad definition) .................... 56

Figure A.2 Distribution of after housing costs income, by scenario (narrow definition) .................. 56

Abbreviations | vii

ABBREVIATIONS

AHC After housing cost

AROP At risk of poverty

AT-HOME Affordability and Tenure Housing Microsimulation Engine

BHC Before housing cost

CSO Central Statistics Office

HAP Housing Assistance Payment

PRSI Pay Related Social Insurance

RAS Rental Assistance Scheme

RS Rent Supplement

SILC Survey on Income and Living Conditions

SPA State Pension Age

TILDA The Irish Longitudinal Study on Ageing

TRIAM TILDA Retirement Income Adequacy Model

Executive summary | ix

EXECUTIVE SUMMARY

CONTEXT AND RESEARCH RATIONALE

In recent years, Ireland has experienced a notable decline in homeownership and

a rise in the proportion of households in private rental accommodation.

Concurrently, the cost of both rents and house prices have risen markedly and

affordability challenges in terms of meeting housing costs have been well

documented. While these challenges are immediate in terms of the housing costs

burden, a longer-term issue arises as to how renting households will manage the

financial adjustment to retirement with ongoing rent payments. This transition

historically in Ireland has been smoothed by high homeownership rates and

mortgage amortisation which lowered the cost of housing payments and increased

financial assets. Indeed, existing international evidence suggests that being a

homeowner can provide a ‘double dividend’ of lowering housing costs and

increasing assets in retirement. This can help manage changes to income in

retirement and allow households to maintain their standard of living.

The continued requirement to cover housing costs for renters in retirement can be

seen as an additional vulnerability in terms of retirement income adequacy in

Ireland, over and above the general concern of sufficient pension income. While

considerable research to date has documented this well-known challenge of

providing a sufficient standard of living in retirement for many households in

Ireland, the drop in the homeownership rate layers an additional complication in

terms of the policy challenge.

To address these related issues, this report explores indicative future paths for

homeownership rates in Ireland and explores the impact in terms of income

poverty in retirement. Using data from the Survey on Income and Living Conditions

(SILC) and the Irish Longitudinal Study on Ageing (TILDA), we consider a number of

scenarios to assess the possibility of renting households becoming homeowners,

and then test the impact on income poverty rates if households were to continue

to have rental costs into retirement.

SUMMARY OF MAIN FINDINGS

A number of key findings emerge from this research:

Findings on homeownership rates

• Ireland has experienced a marked drop in homeownership rates in recent

years. This has been particularly acute for younger-aged households, with the

share of 25-34s who owned their own home halving between 2004 and 2015,

falling from 60 per cent to just 30 per cent. This has continued to fall for

x | Future trends in housing tenure and the adequacy of retirement income

households aged 25-34 and 35-44 since 2015. For these age groups, the

predominant tenure for non-homeowners is private rental with no policy

support (such as HAP, RAS etc).

• At present the share of households with high housing costs

clearly falls with

age. Over 1-in-5 households aged 25-34 face high housing costs, compared to

around 1-in-17 aged 55-64 and only 1-in-100 of those currently aged 65+. For

all age cohorts the incidence of high housing costs is notably higher for non-

homeowners.

• We undertake a range of simulations that vary income, credit, savings and

house prices to explore the possible transition of current renters to

homeownership. The results indicate a likely, and considerable, drop in the

achievable rate of homeownership relative to previous generations (and

current retirees) across all age groups;

o Homeownership rates are around 10 percentage points lower

(approximately 80 per cent) for those currently aged 55-64 and 45-54

compared with current retirees (approximately 90 per cent); this

differential is unlikely to close substantially for these groups given their

position in the lifecycle;

o The homeownership rate for those aged 35-44 is currently 58 per cent.

Our scenarios indicate that this age cohort could, at a maximum, reach

a homeownership rate of 71 per cent through home purchases in the

open market;

o There is likely greater uncertainty over the 25-34 age group given their

age and lifecycle earnings prospects. However, while alternative

scenarios (e.g. down-payment supports) would boost this cohorts’

homeownership rates, they remain significantly below those of

previous cohorts, and this is unlikely to change considerably unless their

income growth outpaces house price growth for a sustained period;

o As a result, the simulated rate of homeownership is lower again for the

youngest aged group, 25-34, with approximately one-in-two

households becoming homeowners in the majority of scenarios

explored.

• Multiple factors are likely causing the drop in homeownership rates including

labour market dynamics, the relative growth of house prices and incomes,

supply bottlenecks and credit availability. Using a static model, we do not

formally test for the drivers of homeownership in this report, nor do we take

High housing costs is defined as a housing cost-to-income ratio greater than 30 per cent.

Executive summary | xi

account of possible demographic changes (e.g. migration) that would change

the demand for housing.

Findings on income adequacy of future cohorts

• We explore the likely impact of lower rates of homeownership on measures

of income poverty for Irish individuals approaching retirement. Using data on

the group who are due to retire over the next five years (92 per cent of whom

own their home), we estimate that income poverty

in retirement rises from

9 per cent on a before housing costs basis to 14 per cent on an after housing

costs basis.

• Lower rates of homeownership are likely to lead to substantially higher after

housing costs income poverty rates. To demonstrate the impact of changes in

homeownership on income adequacy, and linking to our research above on

lower homeownership for younger households, we simulate income poverty

rates for those approaching retirement would reach 31 per cent under a low

homeownership scenario of 63 per cent, 28 per cent under a medium

homeownership scenario of 70 per cent, and 21 per cent under a high

homeownership scenario of 78 per cent, compared to 14 per cent under the

baseline homeownership rate of 92 per cent.

• Our analysis suggests that women, those with lower levels of education and

those living alone during working life are particularly vulnerable to income

poverty in retirement after housing costs are taken into account.

POLICY CONSIDERATIONS

A number of considerations for policy arise from our research:

• Ireland’s pension system (both public and private) has to date leveraged the

‘double dividend’ associated with homeownership – lower housing costs and

higher assets in retirement. It is unlikely that homeownership rates will be as

high in future and therefore consideration will have to be given to the most

appropriate policy response. This could include policies aimed at addressing

the housing costs element, the income adequacy element, or both.

• Policies could be targeted to improve incomes of older households or to

incentivise renting households to build up more resources earlier in their

lifecycle while in employment.

• Income-based support for older households could come in the form of

targeted supports to those groups of households most at risk of income

poverty. This would be more efficient in deadweight terms than blanket

measures to raise state pension rates which would also be fiscally expensive.

This is based on the official at risk of poverty threshold and a ‘broad’ definition of income. The ‘broad’ definition of

income includes income from state and supplementary pensions, as well as the annuitised value of net financial assets

(i.e. savings, stocks/shares, investment property, etc., minus any outstanding non-mortgage debt).

xii | Future trends in housing tenure and the adequacy of retirement income

Indeed, the latter approach may not be fiscally sustainable in the context of

the broader demographic pressures.

• The second area of intervention that shifts the housing costs risk in retirement

is to directly intervene for households with housing costs challenges. The

availability of a targeted instrument such as an income supplement or rent

subsidy to directly cover housing costs for those households who are proven

to be in housing costs distress could feature as part of the future policy

architecture.

• However, the availability of, and even the signalling of, such instruments could

alter the incentives for households to sufficiently accumulate resources earlier

in their working lives. Furthermore, such extensive subsidy supports for those

living in private sector accommodation risks adding to rental inflation and the

costs are likely to be prohibitively expensive for the State given the scale of

the projected fall in homeownership for future retirement cohorts.

• More generally in terms of broader housing policy, interventions at earlier

stages of the lifecycle are critical to help lower the housing costs future

cohorts will face in retirement. Instruments targeting increased supply which

impact the share of homeowners, increased direct provision of social housing

or which develop alternative, non-market renting cohorts such as cost rental

and expanded social housing can all help to lower the cost of housing at

retirement.

• It is however worth noting that while the scenarios in this report attempt to

provide an indicative range for future homeownership rates and the impact

on households’ income adequacy, any assessment that extrapolates across

such a long horizon is subject to considerable uncertainty. In particular, the

current 25-34 age cohort may be expected to experience further career

progression and income growth, which leads to considerable uncertainty in

projecting their likely homeownership rates in retirement.

• In addition, there is also uncertainty about the extent to which current (and

future) housing policy initiatives may increase the share of future

homeowners and/or ensure lower and more stable housing costs for future

cohorts of retirees. Therefore, there is certainly time for policies to influence

the outcomes of this research from both a housing tenure and income support

perspective.

• However, given future demographic projections and associated sustainability

challenges, it is unlikely the State will also be able to fully carry the cost burden

for the challenge of higher housing costs in retirement. Therefore, developing

a policy strategy which mixes lifetime incentives to accumulate retirement

assets during employment with an adequate safety net is most likely to be

required.

Introduction | 1

CHAPTER 1

Introduction

1.1 CONTEXT

A key area of policy concern is the extent to which those who are retired will have

adequate resources to fund their consumption in retirement (European

Commission, 2021; Government of Ireland, 2018; Mulligan et al., 2019). Assessing

the adequacy of incomes in retirement is not a straightforward task, particularly

for those who have not yet retired. A number of assumptions are required to

simulate the level of retirement income, while there are multiple definitions of

what one may want to consider as retirement income. In this report, we consider

adequacy with reference to some fixed level of income deemed sufficient to buy

essential goods and services: the so-called ‘poverty-line’ approach.

Much of the existing work in this area in the Irish context has focused on the extent

to which those who are either already retired or very close to retirement are at risk

of inadequate incomes in retirement (Beirne et al., 2020; Bercholz et al., 2019;

Nivakoski, 2014; Nivakoski and Barrett, 2019). While assessments of retirement

income adequacy are sensitive to the measure of adequacy adopted, and the

definition of retirement income, the most recent Irish estimates suggest that

approximately 9 per cent of those approaching the state pension age (SPA) over

the period 2022-2027 would be at risk of poverty using the 60 per cent of median

household disposable income poverty threshold, based on a definition of

retirement income that includes state and supplementary pension income, as well

as non-primary residence asset income (Beirne et al., 2020).

However, demographic, social and economic changes over recent decades mean

that future cohorts of retirees are likely to look very different to the current cohort

of retirees. One of the most important of these dimensions is housing tenure and

costs. According to the 2016 Census of Population, over 85 per cent of those aged

65+ owned their own home, the vast majority outright (Central Statistics Office,

2017). In addition, most of those aged 65+ renting do so from a local authority or

voluntary body, and thus may expect to have low and relatively stable housing

costs in retirement (Doolan et al., 2022). However, the age at which two-thirds of

the cohort own their own home (with or without a mortgage) has been increasing

steadily over time, from 28 years of age in 1991 to 41 in 2016 (Central Statistics

Office, 2017). This implies substantial differences in the projected tenure and

housing costs of younger cohorts.

See (Beirne et al., 2020) and (Crawford and O’Dea, 2012) for more detailed discussions of the conceptual and

measurement issues involved in assessing retirement income adequacy.

2 | Future trends in housing tenure and the adequacy of retirement income

These cohort-level changes in housing tenure can be illustrated in greater detail in

Figure 1.1. While more than 60 per cent of those born in the 1960s lived in a home

they or their partner owned by age 30, this proportion had fallen to 39 per cent for

those born in the 1970s and 32 per cent for those born in the early 1980s. While

there may be some catch-up at later ages (in part due to changing patterns of

family formation), the figure shows that rates of homeownership for previous

cohorts tend to level off after age 45 (Roantree et al., 2021).

FIGURE 1.1 HOMEOWNERSHIP RATES BY DECADE OF BIRTH COHORT

Source: Reproduced from Roantree et al. (2021) using the Survey of Income Distribution, Poverty and Usage of State Services; the Living

in Ireland survey; and the Survey on Income and Living Conditions.

Notes: Homeownership defined as being the head – or the partner of the head – of a household that lives in an owner-occupied dwelling

(i.e. adult children of owner-occupiers not counted as owning their own home).

The drop in homeownership documented above is likely caused by a range of

factors. As discussed in McQuinn et al. (2021), the tightening of credit access rules

following the onset of the financial crisis, coupled with historically low housing

supply and rising demographic demand, have led to rapidly rising house prices and

rents in Ireland. These factors have led to an increasingly evident housing

affordability challenge, in particular for young, urban renting households who have

faced rapidly increasing rental prices (Corrigan et al., 2019). While these problems

are currently acute, the longer-term implications for these households of how to

cover housing costs in retirement (in lower expected homeownership contexts) has

received less attention in the policy or research literature.

20%

40%

60%

80%

100%

20 25 30 35 40 45 50 55

% who own their own home

Age

1940s 1950s 1960s 1970s 1980s 1990s

Introduction | 3

1.2 EXISTING EVIDENCE

1.2.1 Retirement income adequacy

Most assessments of retirement income adequacy in an Irish context have tended

to focus on those already retired, or close to retirement. For example, using data

from the Irish Longitudinal Study on Ageing (TILDA), Nivakoski (2014) and Nivakoski

and Barrett (2019) found that income replacement rates

were highest for women

and those at the bottom of the earnings distribution, more reflecting the lower

earnings of these groups immediately prior to retirement than high levels of

resources available in retirement. As replacement rates are sensitive to how

income (both pre- and post-retirement) is defined, and generally do not take into

account the wider resources of the household (e.g. from other family members

and/or from assets), or future risks to longevity, investments and healthcare needs

(Knoff et al., 2016), alternative benchmarks are required. An alternative approach

assesses the adequacy of retirement income by determining whether it is sufficient

to meet basic needs (i.e. having enough income to be able to purchase necessities

such as housing, food and clothing). This is assessed in terms of whether retirement

resources are below some benchmark of income poverty, typically the at-risk-of-

poverty threshold.

In this context, a recent study attempted to move beyond a single replacement

rate to consider multiple indicators of retirement income adequacy, varying not

only the definition of income, but also the benchmarks used to assess adequacy.

Beirne et al. (2020) used data from TILDA on the cohort of people born between

1955 and 1960 (and therefore expected to reach the state pension age over the

period 2022-2027) to assess the adequacy of their projected income in retirement,

using a variety of different benchmarks and definitions of retirement income.

The

analysis found that assessments of income adequacy in retirement were very

sensitive to the benchmark used. Even small changes in the benchmark resulted in

large changes in the proportion assessed as having inadequate resources in

retirement. For example, moving from a poverty-line benchmark of 60 per cent of

median household disposable income (the current at-risk-of-poverty threshold) to

a 50 per cent of median household disposable income benchmark (the OECD at-

risk-of-poverty threshold) resulted in a sharp fall in those with inadequate levels of

income in retirement (from 9.2 per cent to 3.0 per cent using a broad definition of

income that includes non-primary residence asset income in addition to state and

The retirement income replacement rate is defined as the ratio of post-retirement income (from pensions, annuitised

wealth holdings, etc.) to pre-retirement income. The current Pensions Roadmap 2018-2023 aims to ensure that the

state pension, in combination with supplementary pensions, replaces a sufficient proportion (between 50 and 60 per

cent) of an individual’s pre-retirement earnings so as to enable them to maintain a reasonable standard of living after

retirement (Government of Ireland, 2018).

In this report, we use this model, the TILDA Retirement Income Adequacy Model (TRIAM) developed in Beirne et al.

(2020), to assess the retirement income adequacy of current and future cohorts of retirees on an after housing costs

basis.

4 | Future trends in housing tenure and the adequacy of retirement income

supplementary pension income). In addition, across all benchmarks, broadening

the definition of income in retirement to include financial assets, and further still

to include half the value of owner-occupied housing wealth, substantially reduced

the proportions of those who were considered at risk of having inadequate

resources in retirement. While the issue of whether individuals should draw on

wealth in the form of primary residences to support living standards in retirement

is contentious, less so is that individuals should draw on other forms of wealth.

These results suggested that private non-pension, non-primary residence assets

play an important role in individuals’ preparedness for retirement.

1.2.2 Housing costs and retirement income adequacy

While the general literature on retirement income adequacy is considerable, few

assessments of retirement income adequacy consider explicitly the role of housing.

Housing can refer to a service (the accommodation that housing provides) and a

capital asset (the dwelling that produces the service) (Fahey and Norris, 2010). One

strand of the literature considers housing from the latter perspective, i.e. as an

asset that can be considered as a source of income in retirement. Fahey (2003)

notes that housing is a valuable asset in retirement which, in principle at least, can

be drawn upon to provide cash income later in life should the need arise (e.g. by

downsizing, equity release, or through family arrangements whereby younger

family members provide support to older family members in return for an explicit

or implicit promise of an eventual bequest). Banks et al. (2005), Crawford and

O’Dea (2012) and Beirne et al. (2020) all considered the role of owner-occupied

housing wealth in their assessments of retirement income adequacy in England

and Ireland, by examining the extent to which these assessments were affected by

the inclusion of half the annuitised value of owner-occupied housing wealth in the

definition of retirement income.

Dewilde and Raeymaeckers (2008) used data from the 2001 European Community

Household Panel (ECHP) to test the idea that housing policies can reduce poverty

in later life by promoting outright ownership, which in turn provides a ‘hidden’

source of income (also addressed using earlier data by Fahey et al., 2004a). Across

the ten Member States included in the ECHP (including Ireland), they found that

neither generous pensions nor high ownership rates had the strongest poverty-

reducing potential; lower poverty rates in older age were most strongly associated

with the provision of social housing for older people. Furthermore, they identified

a group of older people who were faced with a double disadvantage, in the sense

that in high homeownership countries, those who did not possess their own homes

also tended to receive low pension benefits.

In a comparative analysis of a selection of countries with differing housing and

pension systems in the early 2000s, Delfani et al. (2014) found little support for the

hypothesis that housing and pensions in older age were to some extent

Introduction | 5

interchangeable; the exception was in countries such as Ireland with high rates of

homeownership, and no mandatory earnings-related pension scheme. A detailed

analysis of the homeownership-pension trade-off in four European countries

(Belgium, Germany, Ireland and the Netherlands) using Survey on Income and

Living Condition (SILC) data for 2009 showed that outright homeownership among

the income-poor older population played a significant role in reducing poverty

(both income poverty and deprivation)

in old age (Delfani et al., 2015). In Ireland,

outright homeownership had a particularly strong effect on reducing deprivation

in older age.

Focusing on housing costs, a recent paper used data from the German Socio-

Economic Panel (GSOEP) to examine how rising housing costs over the period

1996-2017 have affected income poverty among the older population (aged 65+)

(Lozano Alcántara and Vogel, 2021). The German case is of particular interest as it

has one of the highest percentages of tenants (48.9 per cent in 2019) in Europe.

They found that the share of tenants aged 65+ at income poverty risk after housing

costs rose from 23.9 per cent in 1996 to 33.9 per cent in 2017, while for

homeowners the share remained broadly stable (and much lower) throughout the

period (at approximately 6-8 per cent). A Blinder-Oaxaca decomposition showed

that changes in housing costs between 1996 and 2017 explained most of the

increase in the old-age income poverty rate over the period.

Using data from Australia for 2017-2018, Bradbury and Saunders (2021) compared

before and after housing costs poverty rates across different population groups.

They hypothesised that the before housing costs measure would over-estimate

poverty among older people who often have low (mainly pension) incomes but also

face low housing costs because many of them own their homes outright. In

contrast, they found that the after housing costs poverty rate for single older

people was substantially higher than the before housing costs rate (23.4 per cent

vs. 15.9 per cent), and largely the same on a before and after housing costs basis

for older couples (at approximately 9 per cent). Examining trends over time since

2007, they found that the sharp increase in house prices that has left many

Australians with unpaid mortgages when they reach pension age, and the growing

number of both young and old Australians who are unable to gain access to

homeownership and forced to rely on the private rental market, explained these

patterns.

In the Irish context, Fahey (2003) calculated before and after housing costs at risk

of poverty rates by age cohort using data from 1998. He found that while the after

housing costs poverty rate among older households was 10 percentage points

I.e. inability to afford social activities, healthy food, clothing etc.

6 | Future trends in housing tenure and the adequacy of retirement income

lower than the before housing costs poverty rate, it was still high in absolute terms

(at 30 per cent) and was the highest of any age group. This reflected the relatively

low mortgage costs across the age cohorts observed at this time. Updating the

analysis using data for 2000, Fahey et al. (2004b) found that the contrast between

before and after housing costs income poverty measures was largest for those

aged over 65: the percentage of persons aged 65+ below the before housing costs

income poverty threshold was 43 per cent, in comparison to 34 per cent when

using the after housing costs measure. These figures have to be interpreted in the

light of the patterns of income poverty across age cohorts at that time; in contrast

to now, older adults (aged 65+) were at the highest risk of (before housing cost)

poverty in the early 2000s, with 43.3 per cent in income poverty in 2000, compared

with 16.9 per cent of the population aged 18-64, and 24.9 per cent of children aged

under 18 (Fahey et al., 2004b).

To our knowledge, there has been no research of

after housing costs income poverty in Ireland since the early 2000s, though the CSO

has recently begun to publish such estimates as a supplement to the official

estimates of poverty and deprivation (Central Statistics Office, 2021b).

A particularly important aspect in the homeownership debate, and one which has

focused research where younger adults are considered, is the extent to which

saving for homeownership and/or rental payments affects how younger adults

save for retirement. For example, Crawford and Simpson (2020) found that higher

house prices reduced the probability that young adults (aged 21-35) in England,

particularly those working in the public sector and in the middle of the earnings

distribution, enrolled in a private pension, although the magnitude of the effect

was small. Yates and Bradbury (2010) examined the potential for asset-based

welfare

to protect households from poverty after retirement by focusing on the

role of homeownership in preventing poverty among older Australians and on

likely future trends in asset accumulation. They found that although asset-based

welfare has the potential to ease the fiscal constraints faced by the state, it may

well lead to poorer social insurance outcomes for households with limited saving

capacity over their lifetime. Using data for Australia (which has high rates of

homeownership), they showed that older households who ‘missed out’ on

homeownership were multiply disadvantaged: they had lower non-housing

wealth, lower disposable incomes and higher housing costs in retirement than

homeowners and they had significantly higher after housing poverty rates.

In 2020, children (aged 0-17) had the highest income poverty rate (16.4 per cent) across the age groups, while those

aged 65+ had the lowest (9.8 per cent) (CSO, 2021b).

The indicator of after housing costs income is calculated based on an alternative measure of equivalised income,

excluding the total rent paid and mortgage interest. The total rent paid includes housing supports such as the Housing

Assistance Payment (HAP), Rent Supplement, Rental Assistance Scheme (RAS) which were included in the household

income. Any person with an equivalised income after rent and mortgage interest of less than 60 per cent of the median

before rent and mortgage interest is considered at risk of poverty after rent and mortgage interest (i.e. the same

threshold is used for calculating the rate before and after rent and mortgage interest is deducted) (CSO, 2021b).

Asset-based welfare refers to the use of policies to incentivise accumulation of asset wealth (including homeownership)

in working life, in order to fund consumption in retirement.

Introduction | 7

Projections suggested that this group will grow in size in the coming decades (see

also Yates et al., 2008).

1.3 CONTRIBUTION AND REPORT STRUCTURE

Given the existing literature, both internationally and for Ireland, there is a clear

gap in terms of research covering the intersection of the changing nature of

housing costs and tenure and the adequacy of income in retirement. The purpose

of this report is to address this gap and to explore the following research questions:

• What is the expected proportion of homeowners and renters in Ireland going

forward?

• What is the typical housing costs burden for these households?

• What implications do projected tenure patterns have for the adequacy of

retirement income, in particular for renters?

More specifically, our contribution is as follows. We build on an existing model of

retirement income adequacy developed at the ESRI to assess the impact of housing

costs on retirement income adequacy. This model, TRIAM, uses data from the Irish

Longitudinal Study on Ageing (TILDA) to simulate the incomes in retirement of a

cohort of individuals born between 1955 and 1960 (who can expect to reach the

state pension age (SPA) of 67 between 2022 and 2027) (Beirne et al., 2020).

The

model uses similar data and methods to those used in the UK to assess retirement

income adequacy (Banks et al., 2005; Crawford and O’Dea, 2012). In this model,

the adequacy of retirement income is assessed against a number of poverty line

benchmarks of adequacy. Different definitions of income (i.e. state and

supplementary pension income, adding financial assets, etc.) are also examined.

Cohort-level scenarios of tenure and housing costs are developed using the ESRI’s

micro-simulation tool for assessing housing affordability and tenure (AT-HOME).

This tool considers tenure transitions by households and provides a number of

different scenarios for savings, house prices and incomes to provide housing

tenure proportions. This provides an age-cohort based projection of the share of

renters and homeowners as well as an exploration of the factors that act as the

major barrier to homeownership (for example insufficient savings or income).

These tenure-share projections are incorporated into the TRIAM, coupled with

housing costs estimates for future renters, in a number of stages. In the first stage,

the sensitivity of estimates of retirement income adequacy for the 1955-1960

cohort to the inclusion of housing costs is assessed (i.e. before and after housing

When the model was developed (2019), the SPA was expected to increase to 67 from 2021.

8 | Future trends in housing tenure and the adequacy of retirement income

costs measures of poverty are calculated). In the second stage, we assess the

broader implications of projected cohort-level differences in housing tenure and

costs (based on simulations from the AT-HOME model) for before and after

housing costs estimates of poverty. Three estimates of projected homeownership

rates (high, medium and low – based on the AT-HOME modelling results) are used

to illustrate the impact on retirement income adequacy estimates.

The report is structured as follows. Chapter 2 provides a brief overview of the

institutional background and policy context, summarising the current system of

pension payments in Ireland, as well as the key features of the Irish housing

market. Chapter 3 outlines the results of the modelling to derive cohort-specific

tenure and housing costs (from the AT-HOME model), while Chapter 4 presents the

results incorporating these projections into assessments of retirement income

adequacy using TRIAM. The final chapter, Chapter 5, of the report discusses the

findings, drawing out implications for policy.

Policy context | 9

CHAPTER 2

Policy context

2.1 THE IRISH PENSION SYSTEM

In common with many other OECD countries, the Irish pension system comprises

three main pillars: state, occupational and private pensions (Government of

Ireland, 2018; OECD, 2014; 2021). Data from the 2015 and 2016 Survey on Income

and Living Conditions (SILC) show that nearly 75 per cent of the gross income of

those aged 66+ comes from these sources (Nolan et al., 2019). The latest official

figures show that just under two-thirds of those aged 20 to 69 in employment in

2020 were covered by an occupational or private pension (Central Statistics Office,

2021a).

The first pillar comprises the state (contributory or non-contributory) pension,

which is a largely flat-rate payment (i.e. not earnings-related), payable from age

66. Entitlement to the state pension (contributory) is based on a complex set of

rules, and depends on when the individual first entered employment, the number

of pay-related social insurance (PRSI) contributions, and the average number of

contributions per year.

To qualify for the non-contributory state welfare pension,

a person must pass a means test based on both income and wealth, and must be

habitually resident in Ireland (Nivakoski and Barrett, 2019).

The State Pension Age (SPA) in Ireland is currently 66, with no provision to defer

any entitlement or to claim it earlier. Prior to 2014, the state pension (transition)

in Ireland was payable at age 65 to individuals who retired from insurable

employment and satisfied certain social insurance contribution conditions. The

state pension (transition) was abolished in 2014. In February 2021, a benefit

payment for those aged 65 who have ceased regular employment or self-

employment and who have the required PRSI contributions was introduced.

The benefit is paid at the same rate as the standard rate of Jobseeker’s Benefit

(€208 per week) and stops once the person reaches the SPA of 66.

It is proposed to move to a new ‘total contributions approach’ (TCA) for determining the level of contributory pension

received by an individual. Under this system, the amount of pensions will be directly proportionate to the number of

social insurance contributions made by a person over his or her working life, rather than the average (Government of

Ireland, 2018). While this was initially intended to be fully rolled out for those retiring in 2020 onwards, those reaching

the SPA are currently being assessed under the existing and TCA systems and given an entitlement corresponding to

the higher of the two.

The maximum contributory state pension payable is currently €253.30 per week. The non-contributory state pension

is means-tested and is payable at a maximum rate of €242 for those aged less than 80, and €252 for those aged 80+.

See https://www.citizensinformation.ie/en/social_welfare/social_welfare_payments/older_and_retired_people/ for

further details.

See

https://www.citizensinformation.ie/en/social_welfare/social_welfare_payments/older_and_retired_people/paymen

t_for_people_retired_at_65.html for further details.

10 | Future trends in housing tenure and the adequacy of retirement income

For both the contributory and non-contributory state pensions, it was announced

in 2010 that the SPA would further increase to 67 in 2021 and to 68 in 2028

(Government of Ireland, 2010; 2018). However, the planned increase to 67 was

suspended in 2020. The Government established a Commission on Pensions to

examine various issues around the sustainability of the state pension, including the

SPA. The report of the Commission, published in October 2021, recommended that

the SPA increase by three months per year starting from 2028 to reach 67 years in

2031, and at a slower pace (by three months every two years from 2033) to reach

68 in 2039 (Government of Ireland, 2021).

The second pillar comprises occupational pensions, which cover a broad section of

the population.

Tax relief at the marginal rate is available for contributions to

approved occupational pension schemes. It is generally not possible to contribute

to an occupational pension scheme after the normal age of retirement. There is no

legal obligation for employers to provide occupational pension schemes, but those

who do not must provide access to at least one Personal Retirement Savings

Account (PRSA). However, public service occupational pension schemes are in

place for staff across the civil service, local authorities, Garda Síochána (Irish

police), the Defence Forces, the health and education sectors and non-commercial

state bodies. In order to increase supplementary pension coverage, particularly in

the private sector, in 2019 the Government announced its intention to introduce a

system of automatic enrolment (OECD, 2021). The details of this Automatic

Enrolment Retirement Savings System were announced in March 2022. From

January 2024, all employees not already in an occupational pension scheme, aged

between 23 and 60 and earning over €20,000 across all of their employments, will

be automatically enrolled. Employer and employee contributions will initially be

1.5 per cent of gross income each (with an additional 0.5 per cent provided by the

State), increasing every three years by 1.5 per cent until they eventually reach 6 per

cent for both employee and employer (and 2 per cent by the State) in year ten

(2034) (Department of Social Protection, 2022).

Finally, the third pillar is a voluntary savings pillar, typically comprised of long-term

personal savings and financial investments used to fund retirement over and above

first and second-pillar arrangements. Private pension schemes are voluntary and

include Retirement Annuity Contracts (RACs), commonly used by the self-

employed, and PRSAs, which were introduced in 2002 as an attempt to increase

the pension coverage among low-coverage employee groups (Nolan et al., 2019).

In 2020, two-thirds of workers had an occupational and/or private pension. Of those workers who had an occupational

and/or private pension in 2020, 69.7 per cent had occupational pension cover only (from current or previous

employment), 12.3 per cent had personal pension coverage only and 18 per cent had both (Central Statistics Office,

2021a).

Policy context | 11

The Pensions Roadmap 2018-2023 sets out Government policy in relation to six key

strands of pensions policy, including state pension reform and automatic

enrolment (Government of Ireland, 2018). It recognises that while the state

pension provides basic protection against income poverty in old age, it is not

designed to ensure a high level of retirement income adequacy. Therefore, the

state pension should, in most cases, be combined with supplementary pension

arrangements in the form of occupational and personal pension products. While

international assessments of Ireland’s pension system highlight future

sustainability challenges, particularly if increases to the SPA are deferred

(European Commission, 2021; OECD, 2021), there has been little or no discussion

of the potential impact of cohort-specific changes in housing costs on the

retirement income adequacy of future cohorts.

However, in recognition of the

growing importance of housing costs in household budgets, the CSO started to

publish both before and after housing costs measures of income poverty in 2020

(Central Statistics Office, 2021b).

2.2 THE IRISH HOUSING MARKET

Affordability of, and access to, housing has become one of the critical policy issues

in Ireland in the past number of years. This is due to the culmination of worsening

affordability in both rental and house purchase markets as chronic supply

shortages and rising prices have persisted.

Figure 2.1 documents the trend in housing and rental price changes for the period

2006 to 2022. In the immediate aftermath of the financial crisis, both house prices

and rents dropped considerably reflecting the general economic turmoil of the

period. As the economy began to rebound in 2013/2014, the inflation rate for

housing and rental prices began to recover. In the home purchase market, prices

began to rapidly increase but cooled somewhat in 2015 as the Central Bank of

Ireland introduced a suite of macroprudential mortgage measures which capped

the loan-to-income and loan-to-value ratios for borrowers. The aim of these

measures was to build bank and borrower resilience as well as break the

procyclicality in credit cycles between housing and credit markets (Grace et al.,

2015). House prices began to increase in 2016 and 2017 but tapered off before the

COVID-19 pandemic. Inflationary pressures rebounded in 2021.

The European Commission notes that the housing costs overburden rate (1.5 per cent) for the current cohort of over

65s in Ireland is considerably below the EU average (9.8 per cent). The housing costs overburden rate is the share of

older people living in households where the total housing costs represent more than 40 per cent of the total disposable

household income (European Commission, 2021).

12 | Future trends in housing tenure and the adequacy of retirement income

FIGURE 2.1 TRENDS IN HOUSING COSTS (YEAR-ON-YEAR CHANGE, %)

Source: CSO House Price Index and CPI data (rental indicators).

Notes: Percentage change in index over 12 months reported.

While the macroprudential mortgage measures act to limit feedback loops

between credit and house prices,

these policies can lead to other externalities

such as tighter access to credit for first time homebuyers. Recent research suggests

that Irish households do face credit access constraints from a lack of sufficient

equity accumulation as well as tight income leverage limits (McQuinn et al., 2021;

Kelly and Mazza, 2019). To attempt to deal with these externalities, a number of

policy supports were introduced into the market such as the Help-to-Buy scheme

which provides for a tax rebate towards house purchase of new homes by first time

buyers and the Rebuilding Ireland Home Loan, a direct local authority loan product

(O’Toole and Slaymaker, 2020). Further measures to deal with credit and leverage

access (in the form of general and local authority specific shared-equity schemes)

are forthcoming as part of the ‘Housing for All’ plan announced in 2021.

In the rental sector, prices have also risen rapidly in recent years. The sector has

been accommodating a larger and larger share of the population relative to

historical norms (see Figure 2.2) and this has led to considerable rises in the

average level of rents in the period since 2014. To counteract the rapid inflationary

impulses, rent control policies have been introduced to put a cap on inflation.

These ‘Rent Pressure Zones’ placed limits on the allowable inflation rates per

property for rental tenancies in specific geographic zones (see Coffey et al., 2022).

Research has shown that the macroprudential rules have limited house price inflation in Ireland (McQuinn, 2021).

See De Burca et al. (2021) for a review of the First Home shared equity scheme.

-25

-20

-15

-10

-5

2006M01

2006M08

2007M03

2007M10

2008M05

2008M12

2009M07

2010M02

2010M09

2011M04

2011M11

2012M06

2013M01

2013M08

2014M03

2014M10

2015M05

2015M12

2016M07

2017M02

2017M09

2018M04

2018M11

2019M06

2020M01

2020M08

2021M03

2021M10

House Prices Private Rents Local Authority Rents

Policy context | 13

FIGURE 2.2 TENURE SHARES IRELAND

Source: CSO Census of Population, various years.

The inflationary conditions in both the house purchase and rental sectors have

given rise to serious concerns about housing affordability, with a number of

research studies pointing to affordability challenges for subgroups of the

population (low income, urban, private renters) (Corrigan et al., 2019; O’Toole et

al., 2020).

One of the oft-cited reasons behind the rapid price escalation is chronic shortage

of available properties and that new housing completions are well below estimates

of demographic demand (Bergin and Garcia-Rodriguez, 2020). Following the

financial crisis, new housing production in Ireland fell dramatically and has been

slow to recover as the economy has rebounded. Figure 2.3 displays the total

number of electricity connections in Ireland per year, a proxy for the number of

new housing units (left hand bar). The figure also displays (in red) the number of

completions per 1,000 persons aged 20-50 in the population. Given that this age

bracket is the most likely to provide new forming households, it is a useful proxy

for the demographic pressures coming from this group. Current new supply is well

below the historical levels shown in the data.

A range of policy supports have been introduced in recent years to attempt to

alleviate supply bottlenecks. These policies typically fall into three broad policy

groups: a) those that attempt to deal with the cost of construction such as the site

services fund and the local housing infrastructure activation fund (LIHAF);

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1991 2002 2006 2011 2016

Own Outright Other Renter Local Authority Renter

14 | Future trends in housing tenure and the adequacy of retirement income

b) policies to provide financing for building projects such as the establishment of

the Land Development Agency (LDA) and Home Building Finance Ireland (HBFI);

and c) policies to lessen time frictions in the planning process. While the number

of housing commencements has recovered in 2021, actual completions continue

to remain below any reasonable estimates of demographic demand.

FIGURE 2.3 HOUSING SUPPLY CHANGES – ESB CONNECTIONS – OVERALL AND PER 1,000

PERSONS AGED 20-50

Source: CSO NDA04 – ESB Connections dataset.

Given these housing challenges, there are a number of more recent and long-

standing policies to deal with affordability for lower-income renters outside the

suite of supports that are aimed at the homeownership transition. The bulk of the

policy support is aimed towards those low-income households whose

circumstances are such that they qualify for social housing provision. For this

population cohort a number of supports are available. First, local authorities

provide accommodation directly to social housing tenants who pay an income-

based differential rent. Subsidies are also provided through the Rental

Accommodation Scheme (RAS) and the Housing Assistance Payment (HAP)

whereby the tenant occupies a property in the private rental sector, but the local

authority pays a top-up to the landlord above the differential rent level (paid by

the tenant). Short-term income support is also available for households through

the Rent Supplement (RS) scheme. A more thorough overview of these schemes

and the mechanisms by which they operate can be found in Doolan et al. (2022)

while a documentation of their evolution overtime is contained in Corrigan and

Watson (2019).

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

100,000

Connections per 1000 persons Connections

Housing tenure and cost scenarios | 15

CHAPTER 3

Housing tenure and cost scenarios

Given the housing market context, and the observed lowering of the

homeownership rate in Ireland, consideration must be given as how persistent this

pattern may remain and what the likely implications are for income adequacy in

retirement. Within this context, the purpose of this chapter is threefold. First, we

examine current housing tenure patterns for different age cohorts and how these

have changed over time. Second, we analyse how the proportion of income spent

on housing costs varies across tenures and age groups. Third, we use the ESRI’s

Affordability and Tenure Housing Microsimulation Engine (AT-HOME) tool to

simulate a series of scenarios for homeownership rates for different age cohorts

based on various assumptions around credit access, house prices, incomes and

rental prices. While homeownership may not be the appropriate tenure for certain

households and high levels of homeownership are not necessarily desirable,

homeownership in retirement provides security in the form of low and stable

housing costs, as well as an asset that can be drawn upon as a potential source of

income (Fahey, 2003; Fahey and Norris, 2010). It must be noted that our

assessment of the rate of homeownership focuses on the ability of households to

become homeowners through house purchase and not channels such as whole (or

part) home inheritance. The findings from this chapter will then be used to

examine how housing costs are likely to impact retirement income adequacy for

different age cohorts in Chapter 4.

3.1 HOUSING TENURE

As noted in Chapter 2, the share of homeowners in Ireland has been declining since

2006. To explore this issue in more detail, using data from the Household Budget

Surveys, Figure 3.1 presents the share of homeowners in different age cohorts and

how this has varied over time. Between 1987 and 2004 ownership rates were fairly

stable across the different age groups; rising slightly over the period for the over

35s but falling around 5 percentage points for 25-34s. However, between 2004 and

2015 there was a notable decline in homeownership across those aged 25-64. The

share of 55-64 year olds owning their own home fell nearly 10 percentage points

over this period, while for those aged 35-44, by 2015 it had fallen to 65 per cent,

having fluctuated between 80-85 per cent between 1987 and 2004. The starkest

decline was for the youngest households in our analysis with the share of 25-34s

who owned their own home halving between 2004 and 2015, falling from 60 per

cent to just 30 per cent.

16 | Future trends in housing tenure and the adequacy of retirement income

FIGURE 3.1 % HOMEOWNERS BY AGE COHORT 1987-2015

Source: Authors’ calculations using Household Budget Surveys 1987-2015.

Notes: Homeowners includes both outright owners and owners with a mortgage.

The Household Budget Survey figures presented in Figure 3.1 are only available for

the period up to 2015. The period from 2015 onwards has seen tighter access to

credit with the introduction of macroprudential mortgage regulations, as well as

significant growth in house prices, as shown in Chapter 2. From Figure 3.2, which

uses Survey on Income and Living Conditions (SILC) data for 2018-2019, we see that

the shares of households aged 25-34 and 35-44 who are homeowners have

continued to fall further since 2015, standing at 27 per cent and 58 per cent

respectively for 2018-2019. Ownership rates for those aged 45 and over have

remained similar since 2015.

Regarding the levels of homeownership, Figure 3.2 shows that around 90 per cent

of those in or approaching retirement age (65+) own their own homes. The vast

majority own outright, with fewer than 3 per cent of this age group still making

mortgage payments. Among those who are renting, the majority are in receipt of

some form of housing supports (either living in local authority provided

accommodation or in private sector accommodation with state housing supports).

Fewer than 2 per cent of this cohort live in the private rental sector without

supports. For the subsequent cohort, those currently aged 55-64, the

homeownership rate is around 80 per cent, 10 percentage points lower than the

65+ group. Again, the majority of renters are in receipt of supports, but 5 per cent

of this cohort do live unaided in the private rental sector. The ownership rate for

those aged 45-54 is just under 80 per cent, slightly lower but similar to the 55-64

group. By contrast, far fewer of the subsequent 35-44 cohort are currently

homeowners, with the rate standing at a much lower 58 per cent. For the youngest

cohort (25-34), only 27 per cent currently own their own home, with just over half

of this age group renting unsupported in the private sector.

100

1987 1994 1999 2004 2009 2015

% Homeowners

25-34 35-44 45-54 55-64 65+

Housing tenure and cost scenarios | 17

FIGURE 3.2 % HOUSEHOLDS IN HOUSING TENURES BY AGE COHORT 2018-2019

Source: Authors’ calculations from Survey on Income and Living Conditions (SILC) 2018 and 2019.

Notes: The group ‘rental – with supports’ includes households in in the private rental sector who also report paying rent to a local

authority i.e. HAP or Rental Accommodation Scheme recipients, as well as those living in local authority owned properties.

Households who responded that they are living in their property ‘rent free’ are omitted.

While there may be some degree of homeownership catch-up at later ages,

particularly regarding the youngest age group (in part reflecting increased

education, delayed entry into the workforce and delayed patterns of household

formation), ownership rates for previous birth cohorts have tended to level off by

40-45 (see Figure 1.1). Corlett and Judge (2017) find a similar tapering off of

homeownership rates after 40 in the UK. These findings are consistent with the

fact that banks are typically only willing to offer a mortgage term up to retirement

age and therefore the maximum permissible mortgage term begins to decline after

this point, rendering ownership more challenging.

Corlett and Judge (2017) also find that in the UK both levels and the growth rates

in homeownership between the ages of 30-40 are lower for younger age cohorts.

They conclude the fall in homeownership rates among younger families observed

in the wake of the financial crisis is therefore not likely to fully reverse. Similarly,

Stebbing and Spies-Butcher (2016) find that the rate of catch-up is declining in

Australia and also and conclude that falls in homeownership for younger cohorts

are unlikely to be fully reversed in time. These findings are consistent with Arundel

and Doling (2017) who argue that declines in homeownership are not solely due to

the financial crisis, but that this merely accelerated pre-existing labour market

dynamics such as increased labour market insecurity and are therefore not likely

to be reversed. For Ireland, such a large gap between the ownership rates of the

44-54 and 35-44 groups therefore raises the question of how much this gap is likely

to narrow. We return to this in Section 3.3.1.

100

25-34 35-44 45-54 55-64 65+

Rental - Supports

Rental - No Supports

Mortgaged

Own outright

18 | Future trends in housing tenure and the adequacy of retirement income

3.2 CURRENT HOUSING COSTS

A critical issue in the policy considerations around the Irish housing market is the

issue of high housing costs for particular cohorts of the population. Figure 3.3

presents the mean share of net income spent on housing payment costs by tenure

and age cohort. We define housing payment costs as either rental or mortgage

payments. It does not therefore take into account maintenance, insurance and

other costs associated with housing (such as utilities). The high levels of outright

homeownership for those aged 65+, documented above, mean that on average

these homeowners are currently only paying 0.5 per cent of their income on

mortgage costs, as the vast majority have no housing payment costs. The roughly

3 per cent still with a mortgage are paying an average of 22 per cent of net income

on payments, similar to the 21.6 per cent paid by private renters not in receipt of

housing supports. For all other age groups, non-supported private renters are

clearly paying a higher share of income on their housing costs than other tenures.

While mortgaged households typically pay no more than 20-21 per cent on

average, private renters not in receipt of housing supports in the youngest cohort

(25-34), as well as those aged 45-54, are on average paying 27 per cent of their

incomes on rent. In contrast, the housing payment-to-income ratios of local

authority renters are much lower and fairly consistent regardless of age, ranging

from 11-13.5 per cent on average.

FIGURE 3.3 MEAN HOUSING COST-TO-INCOME RATIO (HCTI) % BY TENURE AND AGE COHORT

Source: Authors’ calculations from Survey on Income and Living Conditions (SILC) 2018 and 2019.

Notes: Owner (all) includes both those who own outright and those who are mortgaged homeowners. We omit housing cost-to-income

ratios for supported renters living in private rental sector accommodation due to the difficulties associated with capturing the

rents actually paid by these tenants in SILC (see Doolan et al., 2022, for more information).

25-34 35-44 45-54 55-64 65+

Owner (all) Mortgaged PR - no supports LA rent

Housing tenure and cost scenarios | 19

Figure 3.4 presents the share of households in each age cohort faced with high

housing costs. We define high housing costs as having a housing cost-to-income

ratio (HCTI) greater than 30 per cent, a commonly used benchmark for high housing

costs (Corrigan et al., 2019). At present the share of households with high housing

costs clearly falls with age. Just over 1-in-5 households aged 25-34 face high

housing costs under this definition, compared to around 1-in-17 aged 55-64 and

only 1-in-100 of those currently aged 65+. For all age cohorts the incidence of high

housing costs is notably higher for non-homeowners.

FIGURE 3.4 % HOUSEHOLDS WITH HIGH HOUSING COSTS (HCTI>30%) BY AGE COHORT

Source: Survey on Income and Living Conditions (SILC) 2018 and 2019.

Notes: Insufficient observations to report share of non-homeowner households aged 65+ with HCTI>30 per cent. Includes all households

with housing payment costs i.e. all but outright owners.

Figures 3.3 and 3.4 highlight the challenges associated with declining

homeownership rates. As noted, while homeownership may not be the

appropriate tenure for certain households and high levels of homeownership are

not necessarily desirable, homeownership in retirement provides security in the

form of low and stable housing costs, as well as an asset that can be drawn upon

as a potential source of income in retirement. Indeed, as highlighted in Section 1.2,

in the literature homeownership and public pensions are often framed as

substitutes in countries with high rates of homeownership (Delfani et al., 2014).

Additionally, Norris (2016) notes how the Irish welfare system has traditionally

focused on property redistribution, with income redistribution and provision of

state services relegated to a secondary objective. As such, high homeownership in

Ireland may have allowed these households to counteract low public or private

income sources in retirement. If the trends of reduced homeownership outlined

above persist into retirement, a substantial portion of future retirement cohorts

will be faced with housing payment costs in their retirement. We now turn to

modelling the likelihood that current renters in different age groups will be able to

transition into homeownership.

25-34 35-44 45-54 55-64 65+

% Households with HCTI>30%

All Non-Homeowners

20 | Future trends in housing tenure and the adequacy of retirement income

3.3 HOMEOWNERSHIP SIMULATION SCENARIOS

3.3.1 Overview: Affordability and Tenure Housing Microsimulation

Engine (AT-HOME)

Having observed the current tenure structure of Irish households, in this section

we use the Affordability and Tenure Housing Microsimulation Engine (AT-HOME)

outlined in McQuinn et al. (2021) to simulate the likelihood of current rental

households in different age cohorts being able to transition to homeownership

under a prudent credit risk assessment policy and current market conditions and

regulations. This assessment has two stages: (i) determining whether a household

is inherently high credit risk and (ii) an assessment of whether a household could

pass the current credit requirements (such as the macroprudential mortgage

regulations) regarding wealth, income and affordability. For this analysis we

combine the 2018 and 2019 waves of the nationally representative Survey on

Income and Living Conditions (SILC). A diagrammatic overview of our AT-HOME

simulation model is presented in Figure 3.5.

Broadly this can be considered a static style modelling approach in that it uses data

on the incomes, employment characteristics, expenditures and housing costs of

existing households and therefore does not focus on household formation or more

dynamic demographic changes (e.g. migration and birth rates) that will

undoubtedly impact the housing market over time. While these aspects are

important, they are not the focus of this work. Our analysis can therefore be

thought of as assessing the proportion of current households in each age cohort

that are: a) currently homeowners and b) likely to be able to become homeowners.

We focus on five age bands: 25-34; 35-44; 45-54; 55-64; and 65+. We only include

cohorts aged 25 and over in our analysis as we do not model the household

formation process specifically. When interpreting our findings, particularly for this

youngest group aged 25-34, it must however be kept in mind that we are only able

to consider those in each age cohort living independently and not those who

remain in their family home. Eurostat figures suggest around one-in-four 25-34

year olds in Ireland live with their parents.

It is important to note that the homeownership simulations presented here are

based on the current market and regulatory conditions. In practice future

homeownership rates could be affected by various policy levers, a greater

provision of social and affordable housing, changes in the availability of policy

supports for homeownership, regulatory changes to private rental sector and

macroprudential regulations, as well as future economic conditions. It must also

be noted that the homeownership transitions that we model in this paper are

Available at:

https://ec.europa.eu/eurostat/databrowser/view/ILC_LVPS08__custom_2353321/default/table?lang=en.

Housing tenure and cost scenarios | 21

concerned solely with those relating to home purchase in the open market. We do

not consider any changes in tenure status that come about due to the inheritance

of whole (or substantial part) properties.

FIGURE 3.5 O

VERVIEW OF STATIC TENURE ASSIGNMENT METHOD

Source: Authors’ analysis.

Notes: We set LTV=90 per cent and LTi=3.5 in line with current macroprudential regulations for first-time buyers. d is down-payment;

is gross household income; P

is the maximum potential monthly repayment (net income minus necessary expenditure);

is the prevailing interest rate for new mortgage loans plus 2 per cent; τ is the mortgage term set at the lower of either

30 years or the number of years until the household head reaches retirement age.

In the first stage of our assessment, we determine whether a household is high

credit-risk, and therefore unlikely to be able to pass a credit check and obtain a

mortgage at present. First, if the household head has been unemployed or inactive

in the labour market during the past 12 months. Second, we utilise a series of

indicators as a proxy for whether a household has a poor credit record. We

determine a household to be high credit risk if they had any rent, utilities or

consumer loan arrears in the past 12 months due to financial difficulties; state that

their existing debt

is a burden or that they are unable to meet an unexpected

expense; and also either face at least one deprivation indicator in SILC. Combining

these subjective measures of burden with more objective measures such as

deprivation indicators aims to capture those facing material difficulties because of

these burdens.

These households are in the rental sector and have no mortgage debt. Existing debt therefore refers to items such as

car loans, credit cards and other consumer loans.

22 | Future trends in housing tenure and the adequacy of retirement income

In the second stage of our assessment, we use data on household income,

estimated savings that could be used as a down-payment, and expenditure to

assess whether they could pass the current regulatory hurdles regarding income,

wealth and affordability. Our methodology draws on the framework outlined in

Kelly et al. (2018) and McQuinn et al. (2021) by calculating three potential house

prices that a household could obtain given their gross income, estimated savings

and an affordability stress test.

Wealth

First-time buyers face a maximum Loan-to-Value (LTV) ratio of 90 per cent under

the current macroprudential mortgage regulations. The maximum house price

permissible under this LTV constraint is therefore d/1-LTV where d is the deposit.

We utilise detailed information on current income and expenditures to determine

what a household could reasonably be expected to accrue as a deposit. Specifically,

we allow a household to save three years’ worth

of any remainder from their net

monthly income after current rent and other minimum necessary expenditure. We

also model the likelihood a household receives an intergenerational wealth

transfer or gift. We allow 25 per cent of households to receive a €10,000 gift. This

choice of parameters is based on evidence on the prevalence and value of gift

giving from TILDA participants to their non-homeowning children (see Appendix

Table A.1). The findings from TILDA are consistent with evidence from the

Department of Housing, Local Government and Heritage’s survey of mature

homeowners (Corrigan, 2019) which found for mature homeowners providing gifts

to their children, the modal gift value was €10,000.

Income

First-time buyers face a maximum Loan-to-Income (LTI) ratio of 3.5

under the

current macroprudential mortgage regulations. Under this LTI constraint the

maximum house price is therefore 3.5 times gross income (i.e. maximum loan size)

plus deposit.

Affordability

Mortgage lenders perform a so-called affordability stress test on prospective

borrowers to ensure they would be able to continue to afford their mortgage

payments in the face of an interest rate rise. We therefore calculate the maximum

house price a household could afford in order to have sufficient residual income

after making mortgage payments to attain their minimum required expenditure if

they were faced with a 2 percentage point increase in the interest rate. We do so

using the amortisation formula outlined in Figure 3.5.

We choose three years based on the findings of Kelly and McCann (2016). The sensitivity of this choice is examined

further in Corrigan et al. (2020) Table 12.

The issue of more substantial inheritances is discussed in Section 3.3.3.

Up to 20 per cent of the value of FTB lending is permitted to be above this 3.5 LTI (Kinghan, 2018). In our sensitivity

analysis presented later in Table 3.1 we therefore increase the maximum permissible LTI to 4.5.

Housing tenure and cost scenarios | 23

If the lowest of these three house prices (i. wealth; ii. Income and iii. affordability)

is above the 25th percentile of the actual transacted house price distribution in the

household’s broad NUTS 3 region, then we determine they are likely to be able to

become homeowners. We choose the 25th percentile because Gaffney (2018)

finds that below this point the majority of transactions are by investors and cash

buyers rather than first-time buyers (see Table A.2 in Appendix II for further

details).

3.3.2 Tenure simulation results

Figure 3.6 presents the baseline results from our tenure simulations. The blue bars

correspond to the current percentage of each age band who are homeowners

(either outright or with a mortgage), while the red bars show the additional

proportion of each age group likely to be able to transition into homeownership

under the simulation methods outlined in Section 3.3.1. Homeownership rates for

those currently aged 65+ sit at 90 per cent. By definition we see no increase in

ownership rates for this group as the model does not permit mortgage borrowing

beyond retirement age. The findings from our simulations indicate that ownership

rates for the subsequent cohort, those aged 55-64, are 10 percentage points lower

and unlikely to rise any further. Ownership rates for the 45-54 cohort are expected

to be similar but marginally lower than for the 55-64 group, with simulations

suggesting an additional 2 per cent of this group could become owners, resulting

in an overall ownership rate of 79 per cent.

FIGURE 3.6 % HOMEOWNERS AFTER STATIC TENURE TRANSITION, BY AGE COHORT

Source: Authors’ calculations using SILC 2018-2019.

100

25-34 35-44 45-54 55-64 65+

Owner - Actual Owner - Baseline Simulation

24 | Future trends in housing tenure and the adequacy of retirement income

There is a marked fall in homeownership rates for the 35-44 cohort relative to the

45-54 group, with simulations suggesting a 65 per cent ownership rate for this

cohort, 14 percentage points below that of the previous age cohort. It is likely many

of this cohort will have been badly affected by the global financial crisis and

resulting changed credit access, as well as by the rapid house price and rental

growth observed since the mid-2010s. Focusing on our final group, those aged

25-34, we see a further substantial reduction, with the simulated homeownership

rate reaching just under 40 per cent. However, given the ages of this group, this

near-term assessment of their likelihood of being able to transition by home

purchase into homeownership is likely to significantly underestimate ownership

rates for this cohort across their lifetimes. We will return to this in Section 3.3.3.

Table 3.1 explores the constraints faced by current renters that do not transition

into homeownership in our simulations. At least two-thirds of these households in

each age cohort are determined to be high credit risk, suggesting they are unlikely

to be able to pass a credit check and obtain mortgage credit. The share of high

credit risk households rises with age, reaching nearly 84 per cent of non-

homeowners in the 55-64 cohort. More than half of these households in each age

cohort are unable to save any of their monthly income after paying their rent and

other essential expenditures. This inability to save is further highlighted by the fact

that more than 80 per cent of those unable to transition in the 25-34 and 35-44

cohorts have insufficient savings for the required down-payment (LTV

constrained).

TABLE 3.1 WHAT % OF NON-TRANSITIONERS ARE CONSTRAINED BY DIFFERENT FACTORS?

Constraining Factor

25-34

35-44

45-54

55-64

High credit risk 64.9 76.0 82.0 83.6

Zero or negative monthly savings

54.3

62.5

57.6

59.0

LTV constrained

81.2

81.6

71.8

71.3

LTI constrained

41.5

37.3

48.5

57.5

Affordability constrained

56.1

62.7

60.4

68.7

Source: Authors’ calculations using SILC 2018-2019.

Note: Negative monthly savings refers to a situation where minimum necessary expenditure on essential goods exceeds the

household’s monthly disposable income.

Scenarios for potential credit access supports

The simulation modelling framework outlined in Section 3.3.1 applies strict credit

conditions based on current macroprudential regulations (see baseline scenario 1

in Table 3.2). As Table 3.1 shows, the majority of those unable to transition into

homeownership face insufficient down-payment constraints, while around two-in-

five are also constrained by the loan size relative to their income. Given the

importance of credit access conditions for household transitions, there is potential

Housing tenure and cost scenarios | 25

scope for policy support (both on the income and down-payment side) to

potentially aide households to become homeowners.

To explore the potential impact of easier credit access conditions, in addition to

our baseline scenario (scenario 1), we also deploy a number of illustrative scenarios

in Table 3.2 which attempt to capture the possible range of options that could be

used by Irish policymakers (such as instruments that would increase the income-

leverage ratio like local authority loans or shared-equity or deposit support

arrangements). We do not attempt to model any specific Irish policy instruments,

rather we attempt to deploy illustrative scenarios that capture the mechanisms by

which these measures would work. We deploy the following scenarios:

1. On the income side, we loosen the maximum LTI from 3.5 to 4.5 (scenario 2);

2. On the wealth side, we allow an equity, or down-payment support of €20,000

for all households

(scenario 3); and

3. Combine both (1) and (2) above (scenario 4).

Implicitly these scenarios assume an elastic housing supply i.e. that this additional

credit in the market would have no impact on house prices because housing supply

would adjust. In reality, in such a supply constrained housing market as the Irish

market, this is an unrealistic assumption. In scenarios 5-7 we therefore account for

the feedback loops between credit and house prices i.e. in each of the scenarios

deployed we take the estimated increase in credit extended and increase house

prices proportionately given an illustrative elasticity. We apply an elasticity of 0.15,

meaning that a one per cent increase in credit is associated with a 0.15 per cent

increase in house prices. This parameter is consistent with the international

findings of Favara and Imbs (2015), as well as McQuinn et al. (2021) who find

elasticities in the range 0.143-0.179 for Ireland (depending on the model

specification).

Note this is a simplified, illustrative exercise and, while its impact targets a similar channel, is not intended to mirror

current equity supports such as the Help-to-Buy scheme which is restricted to new build properties only and is only

available to those with sufficient prior tax contributions.

26 | Future trends in housing tenure and the adequacy of retirement income

TABLE 3.2 HOMEOWNERSHIP RATE SIMULATIONS BY AGE COHORT (%) – ADDITIONAL

SCENARIOS

Scenario

25-34

35-44

45-54

55-64

>65

Change from Baseline Scenario (1)

39.7

64.7

79.2

81.4

90.3

39.9

64.7

79.3

81.4

90.3

LTI

3 52.2 72.5 84.2 82.9 90.3 Equity support

54.8

73.7

85.0

83.3

90.3

LTI & equity support

39.9

64.7

79.3

81.4

90.3

LTI & feedback to house prices

48.8

70.5

83.0

82.7

90.3

Equity support & feedback to house prices

51.4

71.0

83.3

82.8

90.3

LTI & equity support & feedback to house prices

Source: Authors’ calculations using SILC 2018 and 2019.

Note: Sensitivity scenario parameters: LTI – increased from 3.5 to 4.5 for all; Equity support – €20,000 towards savings for down-payment

(note this is a flat rate payment for all, not dependent on household’s tax contribution as current Help-to-Buy scheme); scenarios

5-7 take account of the impact increased credit has on house prices. To do so we use an elasticity of 0.15. This is consistent with

the findings for McQuinn et al. (2021) for Ireland who find the elasticity for Ireland ranges between 0.143-0.179 depending on the

model specification.

Homeownership rates in these additional scenarios remain similar to those

presented in Figure 3.6 for households aged 45 and over. This is consistent with

previous findings that homeownership rates level off after 40-45 (Corlett and

Judge, 2017; Roantree et al., 2021). Unsurprisingly there is more variation for the

35-44 and 25-34 cohorts. Focusing on scenarios 5-7, we see that providing

households with this significant down-payment support could raise the ownership

rates by 9 and 6 percentage points respectively for the 25-34 and 35-44 cohorts

(scenario 6). Solely increasing the multiple of income households can borrow has

minimal impact on the homeownership rate (scenario 5). This is not to say that LTI

constraints are not binding, but rather that LTV or wealth constraints are more

prevalent. Indeed, this is evident from scenario 7 where combining income and

down-payment supports further raises the homeownership rates, particularly for

the youngest 25-34 cohort. Nevertheless, even in the face of significant supports,

the homeownership rate for the 35-44 cohort would be expected to reach a

maximum of 71 per cent, remaining around 12 percentage points below that of the

previous 44-54 cohort under the same supports.

Even with these significant supports, ownership rates for the youngest cohort

remain substantially lower still at just over 50 per cent. However, given the ages of

this group, this near-term assessment of their likelihood of being able to transition

into homeownership is likely to underestimate ownership rates for this cohort

across their lifetimes. We will now examine this further.

3.3.3 25-34 cohort

The median age of first-time buyers in Ireland in 2019 was 34 (Gaffney and Kinghan,

2021). Many in their mid to late 20s and early 30s may therefore expect to see

significant career progression and therefore income growth before they attempt

to purchase a home. The near-term modelling approach outlined in Section 3.3.1

Housing tenure and cost scenarios | 27

is therefore arguably not suitable for the youngest cohort in our analysis.

To address this, in this section we transition households in the 25-34 age cohort

forwards ten years based on a series of growth rates outlined below. As above,

when interpreting our findings, it must be kept in mind that we are only able to

consider those in this age cohort living independently and not those who remain

in their family home.

In our baseline dynamic scenario we grow each household’s income forward based

on a sector-age-education growth rate matrix. Essentially this grows forward a

25-34 aged household’s income to the level of those with similar employment

sector and education profiles in the current 35-44 cohort.

In addition, we allow

house prices to grow at 5 per cent per year over this ten-year period (broadly in

line with long-run average house price growth), rental growth is capped at 2 per

cent annually as per current regulations in Rent Pressure Zones, and households’

minimum necessary expenditures on other goods and services are assumed to

grow at 2 per cent per year, as are incomes (over and above the transformation

described above). We will examine the sensitivity of these parameter assumptions

below. In addition, we remove the current credit risk assessment outlined in

Section 3.3.1 as it is based on the current situation of the household and may

therefore not be relevant ten years into the future. Assumptions around the time

to accumulate a down-payment, the parameterisation of intergenerational wealth

transfers (gifts) and LTV/LTI requirements under the macroprudential measures for

first-time buyers remain unchanged from the framework outlined in Section 3.3.1.

We present the results of these simulations in Table 3.3. We do so both under

current macroprudential conditions (Baseline) and in the case of income and

down-payment supports, accounting for the likely impact of this on house prices

(LTI and equity support and feedback to house prices). Under these baseline

dynamic assumptions, our simulations suggest that 52 per cent of the 25-34 cohort

would be able to transition into homeownership in ten years’ time (Table 3.3,

scenario 2). This is significantly higher than the 39.7 per cent suggested by the near-

term static simulations for this cohort presented in Section 3.3.2 (also shown in

Table 3.3, scenario 1), but still notably lower than for previous age cohorts.

Combining income and down-payment supports whilst accounting for their likely

impact on house prices would further raise the homeownership rate to just over

55 per cent, 4 percentage points higher than for the equivalent near term static

scenario for this cohort (Table 3.2, scenario 7).

Specifically, we take the difference between the median income level for workers aged 25-34 and those aged 35-44 in

each NACE sector-education grouping. We then uprate household incomes for the 25-34 group by this difference to

bring their income levels in line with similar households in the 35-44 age group. On top of this, we allow their incomes

to grow 2 per cent per year over this ten-year period.

28 | Future trends in housing tenure and the adequacy of retirement income

TABLE 3.3 HOMEOWNERSHIP RATE SIMULATIONS FOR 25-34 COHORT (%)

Scenario

Baseline

LTI & equity support &

feedback to house prices

Scenario Notes

1 39.7 51.4

Static (taken from Table 3.2 scenarios 1 (baseline) and

7 (LTI & equity support & feedback to house prices)

52.0 55.3

Dynamic

Sensitivity:

55.4

59.3

3% annual house price growth

4 48.6 52.5

House prices growing faster than incomes

(7% annual house price growth)

5 60.0 64.3

Uniform house price and income growth

(5% annual income growth for all)

51.7

55.3

4% annual rental price growth

48.5

51.8

8% annual rental price growth

52.0

55.4

Increase retirement age to 70

Source: Authors’ calculations using SILC 2018-2019.

Note: Baseline: LTI 3.5; three years to save deposit; 25 per cent get €10,000 gift towards deposit; no current high credit risk filter; implicit

3.2 per cent average annual income growth; 2 per cent annual rental price growth; 5 per cent annual house price growth – using

the 25th percentile (p(25)) house price sold in NUTS 3 area as threshold; minimum expenditures 2 per cent annual growth. In each

scenario we change one of these parameters from its baseline level. Scenarios 3-8 should be compared against scenario 2.

It must be noted that these figures should not be seen as maximum possible

achievable rates of homeownership for a number of reasons. This includes the

considerable uncertainty around projections concerning long-term income and

housing market dynamics, the impact of whole home inheritances and other policy

measures. Rather our scenarios should be seen as illustrative possible transition

likelihoods given the current household structure and different scenarios for

incomes, house prices and credit access. In the medium- to long-term, all of these

aspects can be impacted by policy changes.

Given the considerable uncertainty around the future paths of incomes, house

prices and rents used for projecting these households’ situation ten years into the

future, it is useful to understand how changing key parameters (and therefore the

relativities between them) would affect the likelihood of homeownership for

households currently aged 25-34.

Housing tenure and cost scenarios | 29

FIGURE 3.7 % HOMEOWNERS 25-34 COHORT – DIFFERENT HOUSE PRICE AND INCOME GROWTH

SCENARIOS

(a) (b)

Source: Authors’ calculations using SILC 2018-2019.

Note: Red bars represent baseline scenario (Table 3.3 Baseline column, scenario 2); blue bar represents Table 3.3 Baseline column,

scenario 3; green bar represents Table 3.3 Baseline column, scenario 4; purple bar represents Table 3.3 Baseline column,

scenario 5.

Figure 3.7a explores how the likelihood of homeownership changes when we alter

the rate of annual house price growth over the ten-year period, holding all other

parameters constant. In the blue bar we lower the annual house price growth to

3 per cent (down from 5 per cent in the baseline scenario). This results in the

homeownership rate rising by 3 percentage points to 55 per cent relative to the

baseline scenario. In the green bar we instead increase annual house price growth

to 7 per cent. This highlights the case where house price growth is significantly

more rapid than income growth. In this instance the ownership rate falls by

4 percentage points from the baseline scenario to 48 per cent.

In Figure 3.7b the purple bar illustrates a scenario where we instead allow all

incomes to grow at 5 per cent per year, in line with the 5 per cent annual house

price growth assumption. This contrasts to the baseline scenario where

households’ income growth paths are determined individually depending on their

employment sector and education levels. Instead, here we see that if incomes and

house prices were to grow in line with each other for all households, the ownership

rate would increase by almost 8 percentage points to around 60 per cent. The

simulations presented in Figure 3.7 demonstrate the importance of the relative

growth rates of incomes and house prices (as well as how income growth is

distributed across households) in determining the homeownership rate for this

cohort.

Annual House Price Growth

% 25-34 Homeowners

3% Baseline (5%) 7%

Annual Income Growth

% 25-34 Homeowners

Baseline (Average 3.2%) 5% for all

30 | Future trends in housing tenure and the adequacy of retirement income

The remainder of our sensitivity checks are presented in scenarios 6-8 in Table 3.3.

Increasing the annual rental price growth from 2 per cent to 4 per cent only has a

small negative impact on the ownership rate (scenario 6). This finding is likely

explained by the high level of rents relative to incomes for many households that

may prevent them from saving sufficiently for a deposit. Indeed, in the SILC data

we find just over one-in-three of the age 25-34 households are not currently able

to save anything on a monthly basis once they have paid their housing costs and

other essential expenditures. Significantly higher rental growth of 8 per cent per

year (scenario 7) would however have a more notable negative impact on

ownership rates. Increasing the retirement age for this cohort to 70 also has little

effect on the overall ownership rate (scenario 8). Increasing the retirement age

affects the affordability stress test by lengthening the available mortgage term.

However, as the majority of households are constrained by insufficient wealth or

income to overcome the LTV and LTI conditions, rather than solely by the

affordability stress test, this has a negligible effect on the overall homeownership

rate.

Regardless of the specific parameters chosen, the results in Table 3.3 clearly show

that the 25-34 cohort are likely to see significantly lower homeownership rates

relative to previous age cohorts. It is however important to keep in mind a number

of caveats when modelling homeownership rates for this youngest cohort. We are

only able to model households living independently with the required income,

housing costs and expenditure data, meaning that those living with parents or

family members are not included in the analysis.

Furthermore, given the dynamic

nature of this analysis and uncertainty around both how household incomes and

house prices are likely to develop, there is more uncertainty around the estimates

for this age group.

One major barrier to homeownership for younger age cohorts is the lack of long-

term job and earnings security (Lersch and Dewilde, 2015; Akdogan et al., 2019).

While we do not explicitly model this, analysis of the type of employment contract

for those we find could transition into homeownership confirms that this is largely

being captured implicitly in our simulation modelling framework.

One often mentioned potential mitigating factor for the fall in homeownership

rates observed for younger age cohorts is that of future inheritance (see above). In

It is unclear whether we would expect those remaining in the family home to have a higher or lower likelihood of

becoming homeowners relative to those living independently that we are able to model. On the one hand those living

with family may have more opportunities to save due to not making rent payments (or making smaller contributions

to the family) which could increase their likelihood of transition into homeownership. On the other hand, it could be

the case that many of those remaining in the family home do so because they have lower incomes and cannot afford

to live independently, which could make them less likely to transition.

Housing tenure and cost scenarios | 31

our model we parameterise parental gifts towards a deposit. However, we do not

model the inheritance of whole homes.

Rather our framework only estimates the

transitions related to market home purchases as noted above. The role future

inheritance will play in ownership rates for younger age groups is unclear. While

some households will benefit from inheriting an entire home, in practice there are

a number of reasons why we may not expect this to significantly counteract the

observed fall in homeownership in younger age cohorts. First, international

evidence suggests that inheritance tends to amplify inequality within generations,

i.e. between those with richer and those with poorer parents (Bourquin et al.,

2021). This suggests that many of those in a position to inherit a family home will

have already received parental gifts towards a deposit. In contrast, those who

remain non-homeowners at a later age are those less likely to receive a substantial

inheritance. Second, given current life expectancy, it is quite likely that many

recipients may inherit in their 50s and 60s, past the age at which mortgaged

homeownership is typically feasible. If any inheritance was therefore insufficient

to purchase a home outright, such as in the case of several siblings receiving an

equal share, these households would likely struggle to obtain mortgage credit and

transition into homeownership at that point. Despite these caveats, it is possible

that the homeownership rate for renters would rise if explicit modelling of this

potential inheritance channel were to be conducted.

3.4 SUMMARY

The aims of this chapter were to examine current housing tenure and cost patterns

for different age cohorts and to simulate the likelihood that households will

become homeowners. To do so we used the ESRI’s Affordability and Tenure

Housing Microsimulation Engine (AT-HOME) tool to simulate the likelihood of

current rental households in different age cohorts being able to transition to

homeownership under a prudent credit risk assessment policy and current market

conditions and regulations.

Our findings show the current retirement cohort have very high rates of

homeownership (90 per cent) and therefore very low levels of housing costs. In

addition, the majority of non-homeowners in this cohort are in the supported

rental sector, either living in local authority provided accommodation or housed in

private sector accommodation and in receipt of housing supports. On average

housing cost-to-income ratios are significantly higher in the private rental sector

than other tenures. For those below the age of 55, while mortgaged households

typically pay no more than 20-21 per cent on average, private renters not in receipt

of housing supports in the youngest cohort (25-34) as well as those aged 45-54 are

on average paying 27 per cent of their incomes on rent.

Future research which can incorporate this into the framework would be a natural extension to this research.

32 | Future trends in housing tenure and the adequacy of retirement income

Relative to current pensioners, homeownership rates are around 10 percentage

points lower for the subsequent 55-64 and 45-54 age cohorts and are unlikely to

rise further for these groups. Moreover, our simulations based on current market

and regulatory conditions would suggest a further significant decline in

homeownership rates for younger cohorts. Our baseline estimates suggest around

two-thirds of those aged 35-44 are likely to become homeowners and only 52 per

cent of those aged 25-34. While combining an equity support towards a down-

payment and loosening the amount households can borrow relative to their

incomes would boost these cohorts’ homeownership rates, they remain

significantly below those of previous cohorts.

A number of caveats must be kept in mind when interpreting these findings. Our

simulations are based on current market and regulatory conditions. Ownership

rates for younger cohorts are sensitive to future economic and regulatory

conditions. In particular, we show that estimates are most sensitive to future house

price and income growth and the relativities between them. Furthermore, there

are levers which could help to alleviate such stark declines in homeownership rates

for younger cohorts, particularly a large-scale increase in the supply of affordable

housing both in the owner occupation sector and the rental sector which could

facilitate the accrual of deposits.

One limitation of our model is that households are assessed on their ability to

purchase a home within their current NUTS 3 region. For those in Dublin facing the

highest house prices, this could be a restrictive assumption as in practice they may

be able to purchase in a neighbouring area. However, while this may on the one

hand work to increase homeownership, it could also likely drive up prices in those

areas, which could then crowd out those currently living in those areas.

Income adequacy in retirement | 33

CHAPTER 4

Income adequacy in retirement

4.1 INTRODUCTION

The previous chapter clearly outlined the expected fall in homeownership in

younger age cohorts for Ireland over the coming years. The lower rate of

homeownership will naturally lead to a greater share of the population paying

housing costs into retirement (relative to the natural mortgage amortisation

process that occurs for homeowners). The critical question remains as to the extent

to which these households will face income inadequacy given that housing costs

remain part of their monthly expenditure.

Previous research by Beirne et al. (2020) assessed the income adequacy of those

who were approaching retirement (born 1955-1960) by calculating income poverty

rates on a before housing costs basis. The analysis identified a number of groups

who were already vulnerable to income poverty on this basis, and it is likely that

these and additional groups may face additional challenges if potential housing

costs in retirement are taken into account.

The aim of this chapter is to attempt to assess the scale of this challenge. We use

an augmented version of TRIAM (the ESRI model of income adequacy in retirement

developed by Beirne et al. (2020) based on TILDA, the Irish Longitudinal Study on

Ageing) to calculate after housing costs measures of income poverty for a) the

current cohort of those approaching retirement (i.e. those born 1955-1960) and b)

future cohorts of retirees (by adjusting the homeownership rates of the current

cohort of retirees to reflect projected homeownership rates of future cohorts with

reference to the range of homeownership scenarios outlined in Chapter 3).

4.2 DATA AND METHODOLOGY

For this analysis, we follow Beirne et al. (2020) in using the main TILDA dataset; a

nationally representative survey of community-dwelling individuals aged 50 years

and over, and their spouses or partners of any age (i.e. individuals living in nursing

homes or other institutions were excluded at baseline). The survey is harmonised

with other international longitudinal studies of ageing – such as the US Health and

Retirement Study (HRS), the Survey of Health, Ageing and Retirement in Europe

(SHARE) and the English Longitudinal Study on Ageing (ELSA) – and contains rich

information on the demographic, health and socio-economic circumstances of

older people in Ireland. Data are collected primarily via computer-aided personal

interviewing (CAPI). Data collection for the first wave took place over the period

34 | Future trends in housing tenure and the adequacy of retirement income

October 2009 to February 2011, when 8,504 individuals were sampled, of whom

8,175 were aged 50+ years (Barrett et al., 2011).

TRIAM simulates future disposable – after tax and welfare payments, but before

housing costs (BHC) – income in retirement for the sample who were born between

1955 and 1960 (i.e. those aged between 50 and 55 at TILDA Wave 1, who are

expected to reach the state pension age between 2022 and 2027).

The model is

described in detail by Beirne et al. (2020). We use detailed information on

employment history, income, wealth, pension cover and family status, collected in

TILDA Wave 1 in 2010, along with assumptions about likely income and asset

growth rates,

to simulate future disposable income at the point of retirement. In

this report, we augment these simulations by incorporating an after housing costs

(AHC) measure of disposable income in retirement. The housing costs we deduct

from disposable income are zero for those who own their house outright, which is

the vast majority of the sample, and equal to rents – private or social – for renters.

The TILDA data do not allow us to distinguish mortgage capital and interest

repayments, so for the very small share of homeowners who do not expect to own

their dwelling outright at the point of retirement we include both payments.

Following the approach of Crawford and O’Dea (2012), we consider alternate

definitions of income which differ in their treatment of assets accumulated before

retirement. Our first definition, which we refer to as ‘narrow income’, is that

derived from the public and private pensions of household members and market

income of spouses who are still in work, less applicable direct taxes. The second

adds to this income from financial assets, or more specifically the annuitised value

of net financial assets (excluding owner-occupied but including holiday or

investment property). Where the individual is a member of a couple, incomes of

both partners are considered and total household income is adjusted (equivalised)

for household size.

To assess the adequacy of income in retirement we adopt a poverty-line approach

that conceives of income adequacy as related to the surpassing the minimum

Further information on the sample design is available elsewhere (e.g. Whelan and Savva, 2013).

A number of these individuals reported being retired at the time of the 2010 TILDA survey (n=58). We drop these from

our sample leaving a final sample size of 1,904.

Tables A1.1 and A1.2 in Beirne et al. (2020) detail the various growth rate assumptions used in the model.

This differs from the definition of housing costs used by both Eurostat and the CSO, which also includes insurance, light

and heat (information on which are not reliably available in TILDA), but excludes capital repayments. Given these

represent the bulk of repayments towards the end of a mortgage, their inclusion means that we are likely overstating

housing costs in retirement for a very small share of our sample.

We use the national (CSO) equivalence scale, which gives the first adult in a household a weight of 1, subsequent adults

aged 14+ a weight of 0.66 and any children a weight of 0.33.

Income adequacy in retirement | 35

resources needed to buy essential goods and services.

What should be

considered an essential good or service is ultimately a subjective question and will

naturally evolve over time. We therefore examine three related measures that

form part of policy discussions both in Ireland and abroad, namely living in a

household with less than 60, 50 or 40 per cent of median equivalised disposable

household income. These thresholds correspond to those used for the national at-

risk-of-poverty (AROP) line (60 per cent), the OECD AROP line (50 per cent), and

what is sometimes called the ‘severe’ AROP line (40 per cent).

4.3 RESULTS AND EMPIRICAL FINDINGS

Table 4.1 presents these AROP measures on a before (BHC) and after (AHC) housing

costs basis for our baseline simulations, i.e. using respondents’ recorded housing

tenure, which corresponds to a 92 per cent homeownership rate for our current

cohort of TILDA retirees. We also present AROP measures on an AHC basis for the

sample given three sets of homeownership scenarios (high, medium and low) of:

• 78 per cent (high scenario);

• 70 per cent (medium scenario); and

• 63 per cent (low scenario).

The calibration of these scenarios is informed by the analysis in Chapter 3. As our

approach uses data on the cohort currently approaching retirement, we explore a

set of plausible scenarios that attempt to cover a range of likely homeownership

rates presented in Chapter 3 (Table 3.2). As there was considerable variation across

the age specific cohorts in Chapter 3, we chose a set of parameter values that

attempt to provide an informed range of possible homeownership levels that

capture the range in the baseline scenario for the groups aged 35-44 and 45-54 of

64.7 and 79 per cent).

We do not model a scenario lower than 63 per cent (such

as those for the youngest age category in Chapter 3) given the considerable

uncertainty around this cohort.

The scenarios are modelled by first estimating the relationship between a set of

household characteristics and the likelihood of transitioning to homeowner status

in AT-HOME. These estimates are then used to predict the probability of

homeownership for our TILDA sample given the same set of characteristics. We

then simulate rental costs for the 14 per cent (high scenario), 22 per cent (medium

There is a large literature in Ireland on the problems with relying on income poverty alone to assess living standards

(Fahey et al., 2004b). However, we do not have the required information in TILDA to be able to calculate measures of

deprivation, and by extension, consistent poverty.

See, for example, Brewer et al. (2010). These measures are set with reference to population-wide AHC or BHC median

income as recorded in the 2017 Survey on Income and Living Conditions, with nominal figures uprated to expected

retirement date by the same factor as earnings (forecast wage growth).

The scenarios modelled here do not map precisely onto those presented in Table 3.2 due to small discontinuities in the

distribution of homeownership probabilities described below. As a result, 78 per cent is the closest calibration to 79 per

cent that we can model, and likewise 63 per cent the closest calibration to 64.7 per cent.

36 | Future trends in housing tenure and the adequacy of retirement income

scenario) and 29 (low scenario) per cent of respondents with the lowest predicted

probability of homeownership, assigning them the median regional rent observed

for similarly aged renters in TILDA. We finally deduct these housing costs from

disposable BHC income to give our simulated AHC income in retirement.

Our

assessment of after housing costs poverty is based on a definition of housing costs

that includes just rent payments (we assume all homeowners in TILDA have repaid

their mortgage by the SPA). It therefore does not take into account other

maintenance, insurance, etc. costs associated with housing.

TABLE 4.1 POVERTY LINE (AROP) MEASURES OF INCOME ADEQUACY

(%)

Share with income in retirement that is:

<60% median

<50% median

<40% median

A. Baseline simulations (BHC):

Narrow

24.4

7.1

3.1

Plus annuitised value of net financial assets

9.2

3.0

1.4

B. Baseline simulations (AHC):

Narrow

28.1

12.2

6.3

Plus annuitised value of net financial assets

14.1

7.9

3.9

C. High homeownership scenario (AHC):

Narrow

38.7

20.9

8.2

Plus annuitised value of net financial assets

23.6

14.1

4.6

D. Medium homeownership scenario (AHC):

Narrow

42.8

24.7

8.9

Plus annuitised value of net financial assets

27.8

16.7

4.6

E. Low homeownership scenario (AHC):

Narrow

47.2

28.3

9.6

Plus annuitised value of net financial assets

31.4

19.1

4.8

Source: Authors’ calculations using data from the Irish Longitudinal Study of Ageing (TILDA).

Note: Median equivalised income for poverty lines taken from 2017 Survey on Income and Living Conditions, uprated to expected

retirement date by forecast wage growth.

Panel A in Table 4.1 shows that for our baseline simulation using the narrow

definition of income, 24.4 per cent of those approaching the SPA are likely to have

income below the 60 per cent AROP line on a BHC basis. This rate falls to 9.2 per

cent when one uses the broader measure of income that includes the annuitised

value of net financial assets, which Beirne et al. (2020) show play an important role

in individuals’ preparedness for retirement. The share likely to fall below the 50

and 40 per cent AROP lines is – by definition – lower but substantially so, at 3 per

cent and 1.4 per cent respectively using our broader measure of income. This

We also examined results using a different approach where rents were set at 20 per cent of (BHC) incomes before

retirement and remained at that (real) level into retirement. These provided similar results to those presented in

Table 4.1.

See Fahey et al. (2004b) for a discussion of the limitations of using before and after housing costs measures to assess

income poverty. The after housing costs income poverty measure ignores the fact that people on a similar income may

simply make different choices about how much to spend on housing versus other goods and services: higher housing

expenditure may be associated with higher-quality housing.

Income adequacy in retirement | 37

illustrates the sizeable share of individuals in our baseline simulation with incomes

of between 50 and 60 per cent of median equivalised income.

Panel B shows the income adequacy measures for this group on an AHC, rather

than BHC, basis. The AROP rates are higher in all instances, rising from 9.2 per cent

to 14.1 per cent using the broader definition of income under the 60 per cent AROP

threshold, and from 1.4 to 3.9 per cent under the 40 per cent threshold. This

relatively muted rise in AROP rates reflects the 92 per cent homeownership rate in

our baseline simulation, with the vast majority of these facing housing costs of

zero.

However, AROP rates are substantially higher on an AHC basis when we

incorporate lower expected homeownership rates. Panel C shows that using our

broader measure of income, 23.6 per cent are likely to fall below 60 per cent AROP

line in retirement under the high-homeownership scenario of 78 per cent

compared to 14.1 per cent in our baseline.

Panels D and E show that the medium- and low-homeownership scenarios result

in even higher AROP rates – of 27.8 and 31.4 per cent respectively – using the

broader definition of income and the 60 per cent AROP line. Both are around

double the AROP rate in the baseline (14.1 per cent AROP assuming a 92 per cent

homeownership rate) homeownership scenario, and illustrate the potential

impacts being exposed to even modest housing costs can have on the adequacy of

income in retirement. Similarly, our simulations for AROP rates using the broader

definition of income and the 50 per cent threshold are at least double the baseline

rate of 7.9 per cent in the medium- and low-homeownership scenarios, and almost

double even in the high-homeownership scenario.

While the ‘severe’ AROP rate is also higher in the simulated homeownership

scenarios than the baseline, the increase is much more muted than for the national

(60 per cent) and OECD (50 per cent) AROP lines. This is because the distribution

of AHC income is particularly dense around 50-60 per cent of the median (i.e.

around the national and OECD AROP line), with most households brought beneath

the national or OECD AROP line by housing costs not brought significantly below it.

38 | Future trends in housing tenure and the adequacy of retirement income

TABLE 4.2 POVERTY GAP FOR VARIOUS MEASURES OF INCOME ADEQUACY

(%)

Poverty gap for:

<60% median

<50% median

<40% median

Baseline Before Housing Costs (BHC) results:

Narrow

16.5

21.9

25.2

Plus annuitised value of net financial assets

18.7

35.4

59.6

Baseline After Housing Costs (AHC) results:

Narrow 22.0 27.1 31.2

Plus annuitised value of net financial assets

27.0

30.7

42.6

High homeownership scenario:

Narrow

22.4

20.9

29.3

Plus annuitised value of net financial assets

24.2

21.5

40.3

Medium homeownership scenario:

Narrow

22.6

19.0

28.1

Plus annuitised value of net financial assets

22.8 18.3 39.1

Low homeownership scenario:

Narrow

22.5

17.8

26.5

Plus annuitised value of net financial assets

22.3

16.9

37.6

Source: Authors’ calculations using data from the Irish Longitudinal Study of Ageing (TILDA).

Note: Median equivalised income for poverty lines taken from 2017 Survey on Income and Living Conditions, uprated to expected

retirement date by forecast wage growth.

This is confirmed by Figure A.1 in the Appendix (which plots the distribution of AHC

income in our baseline and simulated homeownership scenarios), and by looking

at estimates of the poverty gap (the average distance from the poverty line for

those beneath the poverty line, expressed as a percentage of the poverty line).

Estimates of the poverty gap are shown in Table 4.2 for the various poverty line

measures we consider. The key result to note is that estimates of the poverty gap

are almost all smaller in our lower homeownership scenarios than in our baseline

AHC simulations. In other words, while the share of the population beneath the

poverty line is higher under each of our simulated homeownership rate scenarios,

those underneath those lines are beneath by less. Although our focus in this paper

is on poverty line measures of income adequacy, Figure A.1 also shows that lower

rates of homeownership have a substantial impact on the distribution of AHC

income across the distribution. This reflects the fact that lower future rates of

homeownership are likely to reduce the level of AHC resources available to

households, by exposing those renting to higher levels of housing costs in

retirement.

Finally, we turn to look at which groups of individuals are at greater risk of being

below the various poverty lines. Table 4.3 displays the odds ratios from logistic

regressions of individual characteristics on the probability of having disposable

income in retirement below the 60 per cent AROP line on both a BHC basis

(Column 1) and an AHC basis (Column 2). These are for our baseline simulations

(i.e. correspond to the estimates in Panels A and B of Tables 4.1 and 4.2) using the

Income adequacy in retirement | 39

more broadly defined measure of income.

Odds ratios are best explained with an

example. An odds ratio of 2 says that an individual in group A faces twice the

chance of someone in the excluded group B of having an inadequate income in

retirement.

The key information to take from Table 4.3 is whether the coefficient

is above or below 1, indicating that an individual with a certain characteristic (e.g.

being a woman) has respectively a greater or lower chance of having inadequate

income in retirement than a similar individual in the excluded category (in this

example, a man).

The odds ratios show that men, individuals with higher levels of education and

those living in a farming household are less likely to have incomes below the 60 per

cent AROP line on both a BHC and an AHC basis, though the last of these estimates

is not statistically significant on a BHC basis at the 95 per cent level of confidence.

Nor is there any statistically significant relationship between the region of

residence and the likelihood of being below the AROP line. However, perhaps the

most striking result is that while there is no statistically significant association

between being single in working life and being below the 60 per cent AROP line in

retirement on a BHC basis (conditional on other factors), there is an extremely

strong and statistically significant association on an AHC basis. This aligns with the

findings of previous research that those living alone are particularly vulnerable to

having inadequate resources in retirement, with our results suggesting that such

concerns might be even more pronounced after one accounts for housing costs

(Nolan et al., 2019; Beirne et al., 2020).

Table A.3 in Appendix II presents similar models for the ‘narrow’ definition of income.

These allow us to easily compare the relative chance that someone in each group has of having inadequate income in

retirement (which is constant across other covariates), but do not allow the absolute or relative risk of income

adequacy to be directly inferred (as these depend on the values taken by other covariates).

40 | Future trends in housing tenure and the adequacy of retirement income

TABLE 4.3 LOGISTIC REGRESSION OF THE FACTORS ASSOCIATED WITH AROP STATUS, BASELINE

SIMULATIONS, BROADER INCOME MEASURE

(1)

(2)

Share with income in retirement that is:

< 60% median BHC

< 60% median AHC

Living in

Dublin (reference)

Urban area outside Dublin

0.954

0.918

Rural area

1.005

0.760

Household type

Non-farming household (reference)

Farming household

0.502

0.447**

Cohabitation status in working life

Married/cohabitating (reference)

Single

0.891

5.714***

Gender

Man (reference)

Woman

1.531

1.429**

Education

Primary or lower (reference)

Lower secondary

0.779

0.515***

Upper secondary

0.595

0.321***

Post-secondary < degree

0.438

0.265***

Post-secondary degree or higher

0.287

***

0.132***

Observations

1,904

Source: Authors’ calculations using data from the Irish Longitudinal Study of Ageing (TILDA).

Note: Coefficients show odds ratios from logistic regressions. Stars show statistical significance, indicating a p value of < 0.05 (*), < 0.01

(**) and < 0.001 (***) respectively.

4.4 SUMMARY

The results of this modelling exercise highlight the potential scale of the challenge

in ensuring that future cohorts of retirees will have adequate resources for their

retirement. The current cohort of those approaching retirement (born 1955-1960)

has high homeownership rates (above 90 per cent), which means that the

projected increases in the AROP rate for this cohort when housing costs are taken

into account are relatively muted. For example, using the broader definition of

income and the 60 per cent of median income AROP threshold, the AROP rate rises

from 9.2 per cent to 14.1 per cent. However, as detailed in Chapter 3, younger

cohorts are likely to have substantially lower homeownership rates. Based on the

modelling conducted in Chapter 3, assessments of AROP rates assuming alternative

homeownership rates of 78, 70 and 63 per cent results in higher AROP rates. Under

the high-homeownership scenario of 78 per cent, 23.6 per cent are likely to fall

below the 60 per cent AROP line in retirement. The medium- and low-

homeownership scenarios result in even higher AROP rates of 27.8 and 31.4 per

cent respectively; around double the AHC rate for those currently approaching

retirement.

Income adequacy in retirement | 41

The results of this research highlight the additional vulnerabilities in terms of

poverty risks that are likely to occur in Ireland with lower homeownership rates.

As more households have to carry housing costs in the form of rental payments

into retirement, this will increase the proportion of households at risk of poverty

and pose challenges to the pensions architecture. It is possible that these

assessments represent a lower bound on possible effects. The assessments of

income poverty using the TRIAM model in this chapter assume that future cohorts

of retirees will have similar characteristics to those who are currently approaching

retirement (i.e. born 1955-1960). This was done in order to isolate the impact of

projected changes in homeownership rates and housing costs on the retirement

income adequacy of future cohorts of retirees. However, it is likely that future

cohorts of retirees are likely to have different employment histories, savings

profiles, living arrangements, etc. than the current cohort, which in turn may affect

their risk of income poverty in retirement. For example, higher rental costs in early

to mid-adulthood may affect the capacity of future generations to build up the

same level of financial assets as the current cohort of retirees. The next chapter

draws out the implications of these and related issues for policy.

Conclusions | 43

CHAPTER 5

Conclusions

5.1 SUMMARY OF MAIN FINDINGS

As house prices and rents have risen sharply in recent years, Ireland’s housing

market has been characterised as having acute affordability challenges for cohorts

of the population, particularly low-to-moderate income urban renters. A defining

feature of the housing market since 2011 is a declining share of homeowners (in

particular mortgaged homeowners). While this of course contributes to

affordability challenges at present, the fall in homeownership also has important

implications for the provision of pensions and income adequacy in retirement. The

existing Irish pension system (and welfare state) has been geared towards high

homeownership which both lowers housing costs payments and raises asset values

for retirement consumption (Norris, 2016). Lower homeownership naturally

means a higher proportion of households in the rental sector and the continuation

of rental payments past the income retirement cliff without the security of the

family home asset.

This research attempts to consider a range of possible scenarios to quantify the

likely share of homeowners in the population going forward and then to explore

the likely impact of these homeownership rates on income adequacy for retired

households. Under a range of scenarios which vary savings levels (down-

payments), incomes, house prices, and credit access conditions, we find that the

rate of homeownership is likely to be considerably lower for future age cohorts

than current retirees. Our results suggest that a considerably higher proportion of

households aged under 45 will be renting come retirement than in previous

generations. While over 90 per cent of the current cohort of retirees are

homeowners, our baseline estimates suggest that around two-thirds of those aged

35-44 are likely to become homeowners through home purchase in the open

market and, with much greater uncertainty, we estimate only 52 per cent of those

aged 25-34. While combining an equity support towards a down-payment and

loosening the amount households can borrow relative to their incomes may boost

these cohorts’ homeownership rates, they are projected to remain significantly

below those of previous cohorts. However, projecting the likely homeownership

rates in retirement of the current 25-34 age cohort is particularly difficult as this

age cohort may be expected to experience further career progression and income

growth in the period during which homeownership typically occurs (up to age 45).

Using the current group of individuals who are nearing retirement as an

illustration, our results suggest that these projected falls in homeownership will be

associated with a rise in poverty rates on an after housing costs basis given current

44 | Future trends in housing tenure and the adequacy of retirement income

income levels. For example, while 14 per cent of current retirees are at risk of

poverty on an after housing costs basis (using the official 60 per cent of median

income threshold and a broad definition of income), this proportion could be as

high as 31 per cent under the lowest homeownership rate scenario examined.

There are a number of limitations to our analysis that are important to point out.

First, our housing simulations are based on currently-formed households. Broadly

this can be considered a static style modelling approach in that it uses data on the

incomes, employment characteristics, expenditures and housing costs of existing

households and therefore does not focus on household formation or more

dynamic demographic changes that will undoubtedly impact the housing market

over time.

Second, while we do allow for intergenerational wealth transfers, we do not model

explicitly the inheritance of full homes. If whole homes were to be inherited by

renters, this could raise the homeownership rate above the level in our scenarios

of home purchase only. However, previous research (e.g. Lawless and Lynch, 2017)

shows that the receipt of such inheritances is positively associated with

beneficiaries’ own level of income. As transitioning to a homeowner in our

approach is more likely for higher income households (McQuinn et al., 2021), this

evidence suggests the inclusion of whole home inheritances would be unlikely to

substantially improve the homeownership prospects of those we model as

remaining renters into retirement.

Third, it is important to note that the homeownership simulations presented here

are based on the current market and regulatory conditions. In practice future

homeownership rates could be affected by various policy levers, a greater

provision of social and affordable housing, changes in the availability of policy

supports for homeownership, regulatory changes to private rental sector and

macroprudential regulations, as well as future economic conditions. The most

important economic determinant of our scenarios is the relative growth rate of

house prices versus incomes; if real income growth was faster than real house price

growth for a period this would facilitate a greater likelihood of homeownership.

However, for such a dynamic to occur it would mean demand-side factors (such as

income growth and demographics) would have to be outweighed by supply-side

influences in the housing market and would require a reversal of supply-demand

imbalances that are observed in the market at present.

Finally, the assessments of income poverty using the TRIAM model assume that

future cohorts of retirees will have similar characteristics to those who are

currently approaching retirement (i.e. born 1955-1960). This was done in order to

Conclusions | 45

isolate the impact of projected changes in homeownership rates and housing costs

on the retirement income adequacy of future cohorts of retirees. However, it is

likely that future cohorts of retirees are likely to have different employment

histories, savings profiles, living arrangements, etc. than the current cohort, which

in turn may affect their risk of income poverty in retirement. For example, higher

rental costs in early to mid-adulthood may affect the capacity of future generations

to build up the same level of financial assets as the current cohort of retirees.

5.2 POLICY IMPLICATIONS

Given the multifaceted nature of the problem of managing housing costs through

the income retirement shock, it is important to consider how policy can best adapt

to the likely lower structural rate of homeownership for future cohorts of retirees.

Policymakers attempting to consider the optimal policy mix to ensure adequate

income in retirement might consider the policy options as targeting the housing

costs element and the income adequacy element.

On the one hand, policies could be targeted to either improve incomes in older

households or to support renting households to build up more retirement

resources earlier in their lifecycle while in employment where possible. In the

former case, the approach would be to support the incomes of future retirees to a

greater extent so that even after accounting for higher expected housing costs,

incomes are sufficient to keep most above the poverty line. One way this could be

done is through increases in the maximum rates of the state pension above and

beyond that needed to keep track with earnings. However, the trade-off between

income support and fiscal cost efficiency would need to be considered here as such

an approach could be very expensive and would likely have considerable

deadweight given its untargeted nature. Furthermore, this approach may not be

fiscally sustainable in the context of the broader demographic pressures (Irish

Fiscal Advisory Council, 2020).

Another more targeted option could be to direct the additional income support to

those households for whom the income adequacy challenge is most acute. Our

results suggest that those living alone (during working life) are at even greater risk

of having inadequate income in retirement on an AHC basis than those who are

married or cohabiting. This is a group who already experience a much greater risk

of poverty, even before housing costs are considered. A bespoke instrument for

such a group, for example through an enhanced living alone allowance, offers one

more targeted option than broader-based increases in the state pension for all.

The population aged 65-79 is projected to grow by 88 per cent and the 80+ population by 240 per cent over the period

2020-2050 (with a consequent increase in the old-age dependency ratio from 22 to 47 per cent). Assuming current

policies, public spending on pensions is projected to increase from 7.7 per cent of GNI* to 11.9 per cent of GNI* over

the same period (Irish Fiscal Advisory Council, 2020).

46 | Future trends in housing tenure and the adequacy of retirement income

However, this approach still may be fiscally costly and shifts the entire risk onto

future taxpayers with implications for intergenerational equity.

The second area of intervention that shifts the housing costs risk in retirement is

to directly intervene for households with housing costs challenges. At the point of

retirement this could be achieved by a targeted instrument such as an income

supplement or rent subsidy to directly cover housing costs for those households

who are proven to be in housing costs distress. However, the availability of, and

even the signalling of, such instruments could alter the incentives for households

to accumulate sufficient resources earlier in their working lives and potentially

expose the Exchequer to significant cost.

As incomes fall on retirement, many households would likely qualify for existing

supports such as HAP. However, Doolan et al. (2022) highlight the challenges

associated with this increased reliance on the private rental sector to house those

in receipt of housing supports. They show low levels of available properties under

scheme rent limits, particularly for high-risk groups such as single adult

households. Where rent limits do not keep pace with market rents, tenants face

the prospect of making additional top-up payments to landlords thus exposing

them to market forces, which is particularly problematic for retirees on fixed

incomes. Furthermore, such extensive supports for those living in private sector

accommodation risks fuelling rental inflation and the costs are likely to be

prohibitively expensive for the State given the scale of the projected fall in

homeownership for future retirement cohorts. Over and above the cost

implications of renting in retirement, another crucial aspect is the potential for

instability of tenure at a vulnerable stage in the lifecycle. This would be a

particularly prominent concern for those in private sector rather than local

authority accommodation.

Finally, considerations on the rate of homeownership, and implicitly for house

prices and rents, going forward are affected by many intertwined macroeconomic

and housing market factors. Policy interventions at earlier stages of the lifecycle

are critical to help lower the costs of housing that future cohorts will face in

retirement. Increased supply and other housing policy interventions can impact the

share of homeowners, while increased direct provision of social housing or policies

which develop alternative, non-market renting cohorts such as cost rental which

can all help to lower the cost of housing at retirement. In particular, if rents in cost

rental properties grow at a slower pace than incomes over time, this lowers the

affordability burden.

Conclusions | 47

More generally, the uncertainties attached to the timeframe in any longer-term

assessment of housing tenure and income adequacy in retirement would suggest

that rather than attempting to calibrate individual measures in the short term, a

longer-term policy planning exercise should be undertaken. This should consider a

suite of options and a range of policy levers that could be initiated depending on

the path which homeownership and poverty rates take. Indeed, given the general

challenges around ageing and pension provision (Irish Fiscal Advisory Council,

2020), it is unlikely the State will be able to fully carry the cost burden for the

challenge of higher housing costs in retirement. Therefore developing a policy

strategy which mixes lifetime incentives to accumulate retirement assets during

employment with an adequate safety net is most likely to be required. Further

research, on the sensitivity of projections of income adequacy to alternative

assumptions about work, pension and family structure throughout working life; on

the relative costs and benefits of alternative policy responses; and on comparable

policy responses in other countries, would help to tease out the appropriate policy

mix in the Irish context.

References | 49

REFERENCES

Akdogan, K., E. Karacimen and A.A. Yavuz (2019). ‘Cross-country evidence on the link

between job security and housing credit’, Journal of Housing and the Built

Environment, 34(4), pp. 947-963.

Arundel, R. and J. Doling (2017). ‘The end of mass homeownership? Changes in labour

markets and housing tenure opportunities across Europe’, Journal of Housing and

the Built Environment, 32(4), pp. 649-672.

Banks, J., C. Emmerson, Z. Oldfield and G. Tetlow (2005). Prepared for Retirement? The

Adequacy and Distribution of Retirement Resources in England. Institute for Fiscal

Studies.

Barrett, A., G. Savva, V. Timonen and R. Kenny (2011). Fifty Plus in Ireland 2011. The Irish

Longitudinal Study on Ageing, Dublin.

Beirne, K., A. Nolan and B. Roantree (2020). Income Adequacy in Retirement: Evidence

from the Irish Longitudinal Study on Ageing (TILDA). ESRI Research Series Report

No. 107, Economic and Social Research Institute: Dublin.

https://doi.org/10.26504/rs107.

Bercholz, M., A. Bergin, T. Callan, A. Garcia Rodriguez and C. Keane (2019). ‘A micro-

macro economic analysis of pension auto-enrolment options’. Working Paper No.

640. Economic and Social Research Institute: Dublin.

Bergin, A. and A. Garcia-Rodriguez (2020). Regional demographics and structural housing

demand at a county level. ESRI Research Series Report No. 111. Economic and

Social Research Institute: Dublin. https://doi.org/10.26504/rs111.

Bourquin, P., R. Joyce and D. Sturrock (2021). Inheritances and inequality over the life

cycle: what will they mean for younger generations? (No. R188). IFS Report.

Bradbury, B. and P. Saunders (2021). ‘Housing costs and poverty: Analysing recent

Australian trends’, Journal of Housing and the Built Environment.

https://doi.org/10.1007/s10901-021-09899-w.

Brewer, M., D. Philips and L. Siebieta (2010). What has happened to ‘severe poverty’

under Labour? Institute for Fiscal Studies. https://ifs.org.uk/bns/bn90.pdf.

Central Statistics Office (2017). Census of Population 2016—Profile 1 Housing in Ireland.

CSO. https://www.cso.ie/en/releasesandpublications/ep/p-cp1hii/cp1hii/.

Central Statistics Office (2021a). Pension Coverage 2020. Central Statistics Office.

Central Statistics Office (2021b). Survey on Income and Living Conditions (SILC) 2020.

Central Statistics Office. https://www.cso.ie/en/releasesandpublications/ep/p-

silc/surveyonincomeandlivingconditionssilc2020/.

Coffey, C., P.J. Hogan, C. O’Toole, K. McQuinn and R. Slaymaker (2022). Rental Inflation

and Stabilisation Policies: International Evidence and the Irish Experience. ESRI

Research Series Report No.136, Economic and Social Research Institute: Dublin.

https://doi.org/10.26504/rs136.

Corlett, A. and L. Judge (2017). Home Affront: Housing Across the Generations. Resolution

Foundation.

50 | Future trends in housing tenure and the adequacy of retirement income

Corrigan, E. (2019). ‘Attitudinal survey of mature homeowners – preliminary analysis’

mimeo.

Corrigan, E., D. Foley, K. McQuinn, C. O’Toole and R. Slaymaker (2019). ‘Exploring

affordability in the Irish housing market’, The Economic and Social Review,

50(1,Spring), 119-157.

Corrigan, E. and D. Watson (2019). ‘Social housing in the Irish housing market’, The

Economic and Social Review, 50 (1, Spring), pp. 213-248.

Crawford, R. and C. O’Dea (2012). The Adequacy of Wealth among those Approaching

Retirement. Institute for Fiscal Studies.

Crawford, R. and P. Simpson (2020). ‘The impact of house prices on pension saving in

early adulthood’, IFS Working Paper 2020/38.

De Burca, O., R. Kelly and E. O’Brien (2021). ‘A financial stability perspective on the First

Home shared equity scheme’, Financial Stability Notes 12/FS/21, Central Bank of

Ireland.

Delfani, N., J. De Deken and C. Dewilde (2014). ‘Home-Ownership and Pensions: Negative

Correlation, but No Trade-off’, Housing Studies, 29(5), 657-676.

https://doi.org/10.1080/02673037.2014.882495.

Delfani, N., J. De Deken and C. Dewilde (2015). ‘Poor because of low pensions or

expensive housing? The combined impact of pension and housing systems on

poverty among the elderly’, International Journal of Housing Policy, 15(3),

260-284. https://doi.org/10.1080/14616718.2015.1004880.

Department of Social Protection (2022). The Design Principles for Ireland’s Automatic

Enrolment Retirement Savings System. Government of Ireland.

Dewilde, C. and P. Raeymaeckers (2008). ‘The trade-off between home-ownership and

pensions: Individual and institutional determinants of old-age poverty’, Ageing

and Society, 28(6), 805-830. Cambridge Core.

https://doi.org/10.1017/S0144686X08007277.

Doolan, M., B. Roantree and R. Slaymaker (2022). Low Income Renters and Housing

Supports, ESRI Research Series Report No. 141. Economic and Social Research

Institute: Dublin. https://doi.org/10.26504/rs141.

European Commission (2021). The 2021 Pension Adequacy Report: Current and Future

Income Adequacy in Old Age in the EU. European Commission.

Fahey, T. (2003). ‘Is there a Trade-Off between Pensions and Home Ownership? An

exploration of the Irish case’, Journal of European Social Policy, 13(2), 159-173.

https://doi.org/10.1177/0958928703013002004.

Fahey, T., B. Nolan and B. Maître (2004a). ‘Housing Expenditures and Income Poverty in

EU Countries’, Journal of Social Policy, 33(3), 437-454. Cambridge Core.

https://doi.org/10.1017/S0047279404007767.

Fahey, T., B. Nolan and B. Maître (2004b). Housing, Poverty and Wealth in Ireland.

Institute of Public Administration and Combat Poverty Agency.

Fahey, T. and M. Norris (2010). ‘Housing’, in The Oxford Handbook of the Welfare State.

Oxford University Press.

Favara, G. and J. Imbs (2015). ‘Credit supply and the price of housing’, American Economic

Review, 105(3), pp. 958-92.

References | 51

Gaffney, E. (2018). ‘Residential property price segments and mortgage finance’, Financial

Stability Notes 11/FS/18, Central Bank of Ireland.

Gaffney, E. and C. Kinghan (2021). ‘Mortgage lending in Ireland during the 2010s’,

Financial Stability Notes 09/FS/21, Central Bank of Ireland.

Government of Ireland (2010). National Pensions Framework. Stationery Office.

Government of Ireland (2018). A Roadmap for Pensions Reform 2018-2023. Stationery

Office.

Government of Ireland (2021). Report of the Commission on Pensions. Government of

Ireland. https://www.gov.ie/en/publication/6cb6d-report-of-the-commission-

on-pensions/.

Grace, T., N. Hallissey and M. Woods (2015). ‘The Instruments of Macro-Prudential

Policy’, Quarterly Bulletin Articles, issue 01, pp. 90-105. Central Bank of Ireland.

Irish Fiscal Advisory Council (2020). Long Term Sustainability Report: Fiscal Challenges and

Risks 2025-2050.

Kelly, C. and F. McCann (2016). ‘Rental markets, savings and the accumulation of

mortgage deposits’, Quarterly Bulletin Articles, issue 04, pp. 56-70. Central Bank

of Ireland.

Kelly, R. and E. Mazza (2019). ‘A measure of bindingness in the Irish mortgage market’,

Financial Stability Notes 12/FS/19, Central Bank of Ireland.

https://EconPapers.repec.org/RePEc:cbi:fsnote:12/fs/19.

Kelly, R., F. McCann and C. O’Toole (2018). ‘Credit conditions, macroprudential policy and

house prices’, Journal of Housing Economics, 41, 153-167.

Kinghan, C. (2018). ‘Macroprudential measures and Irish mortgage lending: Insights from

H1 2018’, Financial Stability Notes, 08/FS/18, Central Bank of Ireland.

Knoff, M., J. Been, R. Alessie, K. Caminada, K. Goudswaard and A. Kalwu (2016).

‘Measuring retirement savings adequacy: Developing a multi-pillar approach in

the Netherlands’, Journal of Pension Economics and Finance, 15(1), 55-89.

Cambridge Core. https://doi.org/10.1017/S1474747214000341.

Lawless, M. and D. Lynch (2017). ‘Gifts and inheritances in Ireland’. Working Paper No.

579. Economic and Social Research Institute: Dublin.

Lersch, P.M. and C. Dewilde (2015). ‘Employment insecurity and first-time

homeownership: Evidence from twenty-two European countries’. Environment

and Planning A, 47(3), pp.607-624.

Lozano Alcántara, A. and C. Vogel (2021). ‘Rising housing costs and income poverty

among the elderly in Germany’, Housing Studies, 1-19.

https://doi.org/10.1080/02673037.2021.1935759.

McQuinn, K. (2021). 'House prices and mortgage credit: Empirical evidence for Ireland —

An update', ESRI Quarterly Economic Commentary Research Note No. 20210101.

Economic and Social Research Institute: Dublin.

https://doi.org/10.26504/rn20210101.

McQuinn, K., C. O’Toole and R. Slaymaker (2021). ‘Credit access, macroprudential rules

and policy interventions: Lessons for potential first-time buyers’, Journal of Policy

Modeling, 43(5), 944-963. https://doi.org/10.1016/j.jpolmod.2021.03.013.

52 | Future trends in housing tenure and the adequacy of retirement income

Mulligan, E., D. Wijeratne and M. Maher (2019). Pensions in Ireland: The Perspectives of

Irish Citizens. Whitaker Institute.

Nivakoski, S. (2014). ‘Determinants of Pension Coverage and Retirement Income

Replacement Rates – Evidence from TILDA’, The Economic and Social Review,

45(3), 299-328.

Nivakoski, S. and A. Barrett (2019). ‘Estimating, and Interpreting, Retirement Income

Replacement Rates’, The Economic and Social Review, 50(3), 587-609.

Nolan, A., A. Whelan, S. McGuinness and B. Maître (2019). Gender, Pensions and Income

in Retirement. ESRI Research Series Report No. 87. Economic and Social Research

Institute: Dublin. https://doi.org/10.26504/rs87.

Norris, M. (2016). Property, family and the Irish welfare state. Palgrave Macmillan, Cham,

Switzerland.

OECD (2014). OECD Reviews of Pensions Systems: Ireland. OECD.

OECD (2021). Pensions at a Glance 2021 OECD and G20 Indicators. OECD.

https://www.oecd-ilibrary.org/docserver/ca401ebd-

en.pdf?expires=1646922288&id=id&accname=guest&checksum=E84FC4F0BDA

40C00D8F0D2EE7E597438.

O’Toole, C. and R. Slaymaker (2020). A Review of the Rebuilding Ireland Home Loan, ESRI

Research Series Report No. 104. Economic and Social Research Institute: Dublin.

https://doi.org/10.26504/rs104.

O’Toole, C., R. Slaymaker, K. McQuinn, C. Coffey and E. Corrigan (2020). Exploring the

short-run implications of the COVID-19 pandemic on affordability in the Irish

private rental market, ESRI Research Series Report No. 108. Economic and Social

Research Institute: Dublin. https://doi.org/10.26504/rs108.

Roantree, B., B. Maître, A. McTague and I. Privalko (2021). Poverty, income inequality and

living standards in Ireland. Economic and Social Research Institute and The

Community Foundation for Ireland: Dublin.

https://doi.org/10.26504/bkmnext412.

Stebbing, A. and B. Spies-Butcher (2016). ‘The decline of a homeowning society? Asset-

based welfare, retirement and intergenerational equity in Australia’, Housing

Studies, 31(2), pp.190-207.

Whelan, B.J. and G.M. Savva (2013). ‘Design and Methodology of The Irish Longitudinal

Study on Ageing’, J. Am. Geriatr. Soc., 61, S265-S268.

https://doi.org/10.1111/jgs.12199.

Yates, J. and B. Bradbury (2010). ‘Home ownership as a (crumbling) fourth pillar of social

insurance in Australia’, Journal of Housing and the Built Environment, 25(2),

193-211. JSTOR.

Yates, J., H. Kendig, B. Phillips, V. Mulligan and R. Tanton (2008). Sustaining fair shares:

The Australian housing system and intergenerational sustainability (AHURI Final

Report No. 111). Australian Housing and Urban Research Institute.

https://www.ahuri.edu.au/sites/default/files/migration/documents/AHURI_Fin

al_Report_No111_Sustaining_fair_shares_the_Australian_housing_system_and

_intergenerational_sustainability.pdf.

Appendix I | 53

APPENDIX I SUPPLEMENTARY RESEARCH FOR CHAPTER 3

TABLE A.1 EVIDENCE OF LIKELIHOOD OF RECEIVING GIFT / AMOUNT BY RECIPIENT

CHARACTERISTICS

Overall Age 16-24 Age 25-49 Age 50+ Homeowner

Non-

Homeowner

Gift (%)

No gift >€5,000

73.3

81.3

70.9

86.6

71.8

75.0

Gift >€5,000

26.7

18.7

29.1

13.4

28.2

25.0

Value of Gift (€)

Mean

38,681

16,368

40,187

35,954

41,128

34,381

Median 10,000 6,667 10,000 10,000 10,000 10,000

Source: Authors’ calculations from TILDA.

Note: Gift amounts conditional on having received a gift. Assumes that all non-resident children of the TILDA participant(s) receive an equal

amount.

TABLE A.2 25TH PERCENTILE HOUSE PRICE BY NUTS 3 REGION – 2019

NUTS 3 region

25th percentile house price 2019 (€)

Border

80,000

Midlands

111,000

West

105,000

Dublin

278,000

Mid-East

210,000

Mid-West

113,000

South-East

127,500

South-West

160,000

Source: Authors’ calculations from Residential Property Price Register.

Appendix II | 55

APPENDIX II SUPPLEMENTARY RESEARCH FOR CHAPTER 4

TABLE A.3 LOGISTIC REGRESSION OF THE FACTORS ASSOCIATED WITH AROP, BASELINE

SIMULATIONS, NARROW INCOME MEASURE

(1)

(2)

(3)

BHC

AHC

Share with income in retirement that is:

<60% median <50% median <60% median <50% median

Living in

Dublin (reference)

Urban area outside Dublin

1.091

0.833

0.883

0.608**

Rural area

1.480**

1.214

0.938

0.671**

Household type

Non-farming household (reference)

Farming household

1.365

0.941

1.065

1.228

Cohabitation status in working life

Married/cohabitating (reference)

Single

1.387

0.755

4.824***

3.681***

Gender

Man (reference)

Woman

1.252

1.331 1.514*** 1.265*

Education

Primary or lower (reference)

Lower secondary

1.146

1.324

0.622**

1.017

Upper secondary

0.928

1.071

0.396***

0.845

Post-secondary < degree

0.736

0.804

0.318***

0.635*

Post-secondary degree or higher

0.686

0.982

0.170***

0.518**

Observations

1,904 1,904 1,904 1,904

Source: Authors’ calculations using data from the Irish Longitudinal Study of Ageing (TILDA).

Note: Coefficients show odds ratios from logistic regressions. Stars show statistical significance, indicating a p value of < 0.05

(*),< 0.01(**) and < 0.001 (***) respectively.

56 | Future trends in housing tenure and the adequacy of retirement income

FIGURE A.1 DISTRIBUTION OF AFTER HOUSING COSTS INCOME, BY SCENARIO (BROAD

DEFINITION)

Source: Authors’ calculations using data from The Irish Longitudinal Study of Ageing (TILDA).

Notes: After housing costs disposable income constructed as described in Chapter 4, equivalised using the modified OECD equivalence

scale. Estimate from kernel density plot with data censored at €50,000 per year for legibility.

FIGURE A.2 DISTRIBUTION OF AFTER HOUSING COSTS INCOME, BY SCENARIO (NARROW

DEFINITION)

Source: Authors’ calculations using data from The Irish Longitudinal Study of Ageing (TILDA).

Notes: After housing costs disposable income constructed as described in Chapter 4, equivalised using the modified OECD equivalence

scale. Estimate from kernel density plot with data censored at €50,000 per year for legibility.

0.00001

0.00002

0.00003

0.00004

0.00005

0.00006

0.00007

0.00008

0.00009

€0 €10,000 €20,000 €30,000 €40,000 €50,000

Density

Annual equivalised after housing costs disposable income

Baseline High Medium Low

0.00002

0.00004

0.00006

0.00008

0.0001

0.00012

0.00014

0.00016

0.00018

0.0002

€0 €10,000 €20,000 €30,000 €40,000 €50,000

Density

Annual equivalised after housing costs disposable income

Baseline High Medium Low

Whitaker Patrons and Corporate Members

The ESRI plays a leading role in producing independent research which allows policymakers in

Ireland to better understand the economic and social landscape which shapes Ireland. One goal

of the ESRI is to ensure that the Institute has a strong revenue stream to fund valuable,

independent social and economic research initiatives that will have a long-term impact on Irish

society. In 2021, we launched a compelling new membership programme with two categories

of membership:

1) Whitaker Patron and

2) Corporate Membership

We would like to acknowledge the following organisations and thank them for their valuable

support:

WHITAKER PATRONS

EirGrid

ICON

Davy

ESB

Intel

Microsoft

Kingspan