Mapping Poverty - Combat Poverty Agency

RESEARCH 

MAPPING POVERTY: NATIONAL, REGIONAL 

AND COUNTY PATTERNS 

DOROTHY WATSON, CHRISTOPHER T. WHELAN, 

JAMES WILLIAMS, SYLVIA BLACKWELL

AA.Prelims 7/7/05 6:53 am Page iii 

Mapping Poverty: 

National, Regional and 

County Patterns 

Dorothy Watson 

Christopher T. Whelan 

James Williams 

Sylvia Blackwell

AA.Prelims 7/7/05 6:53 am Page iv 

First published 2005 

by the 

Institute of Public Administration 

57–61 Lansdowne Road 

Dublin 4 

and 

Combat Poverty Agency 

Bridgewater Centre 

Conyngham Road 

Islandbridge 

Dublin 8 

© 2005 Combat Poverty Agency 

British Library Cataloguing-in-Publication Data 

A catalogue record for this book is available from the British Library. 

ISBN 1 904541 25 9 

This study forms part of the Combat Poverty Agency Research Series, in which it 

is No. 37. 

The views expressed in this text are the authors’ own and not necessarily those of 

Combat Poverty Agency. 

Publications and printed matter will be made available, on request, in a range of 

formats, including audio tape, large print, Braille and computer disc. 

Cover design by Red Dog Design Consultants, Dublin 

Typeset by Computertype Limited, Dublin 

Printed in Ireland by Betaprint, Dublin

AA.Prelims 7/7/05 6:53 am Page v 

Table of Contents 

List of Tables 

List of Maps 

List of Figures 

About the Authors 

Acknowledgements 

Foreword 

Executive Summary 

ix 

xiv 

xvi 

xvii 

xx 

xxi 

xxxi 

Chapter 1: Introduction 1 

1.1 Background to the Study 1 

1.2 Rationales for Spatial Intervention 5 

1.3 Data Sources 9 

1.4 Outline of the Study 10 

Chapter 2: Spatial Trends in Deprivation Surrogates 11 

2.1 Introduction 11 

2.2 Age Structures 12 

2.3 Economic Status and Activity 16 

2.4 Variations in Levels of 

Educational Attainment 30 

2.5 Variations in Levels of Social Class 37 

2.6 Conclusions 45 

v

AA.Prelims 7/7/05 6:53 am Page vi 

Mapping Poverty 

Chapter 3: Measurement of Poverty at the 

Household Level 46 


3.2 Measuring Poverty and 

Deprivation in the LII Surveys 47 

3.3 Measuring Poverty and 

Deprivation in the NSHQ 49 

3.4 The National Context 55 


Chapter 4: Poverty and Deprivation by 

Region and Local Authority Area 59 


4.2 Poverty Risk and Incidence by Regional 

Authority in the LII Survey 60 

4.3 Disparities in Income Poverty Risk 

by Local Authority Area from the NSHQ 64 

4.4 Modified Consistent Poverty (MCP) 69 

4.5 MCP by Local Authority Area 70 

4.6 Direct and Indirect Measures of 

Poverty and Deprivation 73 


Chapter 5: Poverty by Area and Tenure Type 78 


5.2 Measuring Area Type and Tenure 79 

5.3 Poverty and Deprivation by Area Type 80 

5.4 Poverty and Deprivation by Tenure Type 85 

5.5 Disparities in Poverty Risk by Size 

of Place and Tenure 89 

5.6 Modified Consistent Poverty 

by Area Type and Tenure 93 


vi

AA.Prelims 7/7/05 6:53 am Page vii 

Contents 

Chapter 6: Disparities in Deprivation 97 


6.2 Methodology 97 

6.3 Disparities in Deprivation by 

Region and Local Authority Area 98 

6.4 Disparities in Deprivation by 

Size of Place and Tenure 113 


Chapter 7: Understanding Variation in Poverty 

by Area and Tenure Type 119 


7.2 Assessing the Net Effect of Location 

and Tenure 119 

7.3 Methodology 120 

7.4 Gross and Net Effects of Region 

on Poverty in the NSHQ 122 

7.5 Disparities in Poverty by 

Local Authority Areas (LAA) 125 

7.6 Gross and Net Effects of Size of 

Area and Tenure on Risk of 

Poverty in the NSHQ 127 


Chapter 8: Conclusions 139 


8.2 Surrogate Measures of Deprivation 140 

8.3 Overview of CSO Results 141 

8.4 Geographical Variation: Evidence from 

Analysis of Indirect Measures of 

Deprivation 141 

8.5 Variation by Size of Area and Tenure 143 

8.6 Policy Implications 143 

vii

AA.Prelims 7/7/05 6:53 am Page viii 


Appendix 1: Additional Tables 151 

Appendix 2: Methodology 167 

Glossary 193 

References 195 

viii

AA.Prelims 7/7/05 6:53 am Page ix 


Table 2.1: 

Table 2.2: 

Table 2.3: 

Table 2.4: 

Table 2.5: 

Table 2.6: 

Table 2.7: 

Percentage of persons in each region 

according to age cohort 13 

Regional labour force participation and 

unemployment rates 19 

Distribution of the unemployed 

across regions 21 

Percentage of all persons engaged in 

farming by number of acres farmed 22 

Index of economic dependency 

in each region 28 

Distribution of persons 15 years and over 

whose education has ceased, classified by 

highest level of educational attainment 31 

Distribution of persons whose education 

has ceased in highest and lowest levels of 

educational attainment across regions 36 

Table 2.8: Percentage of persons in each social class 38 

Table 2.9: 

Table 3.1: 

Regional distribution of Professional and 

Unskilled Manual class categories 44 

Items used in the deprivation scales in 

the NSHQ 54 

ix

AA.Prelims 7/7/05 6:53 am Page x 


Table 3.2: 

Table 4.1: 

Table 4.2: 

Table 4.3: 

Table 4.4: 

Table 4.5: 

Trends in percentage of households below 

poverty lines 57 

Poverty risk in 2000 by regional authority 

and regional assembly 62 

Poverty incidence in 2000: Per cent of 

poor in each regional authority and 

regional assembly area (relative income 

poverty and consistent poverty) 64 

Disparities in income poverty risk by local 

authority area 65 

Disparities in risk of modified consistent 

poverty by local authority area 71 

Correlation of socio-demographic 

characteristics of county with risk of poverty 75 

Table 5.1: Risk of poverty by type of area for 1987, 

1994 and 2000 82 

Table 5.2: 

Incidence of poverty by type of area for 

1987, 1994 and 2000 83 

Table 5.3: Risk of poverty by tenure type for 1987, 

1994 and 2000 88 

Table 5.4: 

Table 5.5: 

Incidence of poverty by tenure type for 

1987, 1994 and 2000 90 

Disparities in income poverty risk by 

urban/rural location 91 

Table 5.6: Disparities in income poverty risk by tenure 92 

Table 5.7: 

Disparities in risk of MCP by urban/rural 

location 94 

x

AA.Prelims 7/7/05 6:53 am Page xi 


Table 5.8: Disparities in risk of MCP by tenure 95 

Table 6.1: 

Table 6.2: 

Table 6.3: 

Table 6.4: 

Table 6.5: 

Table 7.1: 

Table 7.2: 

Table A1.1: 

Table A1.2: 

Table A1.3: 

Table A1.4: 

Dimensions of deprivation (cannot afford 

items) 98 

Disparities in risk of basic, secondary, 

housing and environmental deprivation 

by regional authority area 100 

Disparities in risk of basic deprivation, 

secondary deprivation, housing 

deprivation and environmental deprivation 

by local authority 107 



by urban/rural area 114 



by tenure 115 

Gross and net odds ratios of being poor 

versus non-poor by regional authority 125 

Poverty variation between and within 

local authority areas 126 

Percentage of persons in each county 

according to age cohort 152 

County-level labour force participation 

and unemployment rates 154 

Distribution of the unemployed across 

regions 155 

Percentage of all persons engaged in 

agriculture classified by number of acres 

farmed 156 

xi

AA.Prelims 7/7/05 6:53 am Page xii 


Table A1.5: 

Table A1.6: 

Table A1.7: 

Index of economic dependency in 

each county 157 

Distributions of persons 15 years and over 

classified by highest level of educational 

attainment 158 

Distribution of persons whose education 

has ceased in highest and lowest levels of 

educational attainment across counties 159 

Table A1.8: Percentage of persons in each social class 160 

Table A1.9: 

Distributions of persons in Professional 

and Unskilled Manual class categories 

across region 161 

Table A1.10: Risk of poverty by planning region, 1987, 

1994 and 2000 163 

Table A1.11: Incidence of poverty by planning region, 

1987, 1994 and 2000 164 

Table A2.1: 

External population characteristics used 

in the construction of household weights 

for the Living in Ireland Surveys 174 

Table A2.2: Population checks for sample weighting 179 

Table A2.3: 

Level of missing information on key 

variables and imputation procedure 181 

Table A2.4: Mean weekly household income (in £) 

using full measure and single-item 

measure by number of adults and number 

of persons at work 183 

xii

AA.Prelims 7/7/05 6:53 am Page xiii 


Table A2.5: 

Table A2.6: 

Table A2.7: 

Table A2.8: 

Table A2.9: 

Model based on LII to correct for 

understatement of income when a 

single-item measure is used 184 

Income category midpoints and coefficients 

applied to the survey of house quality 185 

Mean ‘corrected’ income for each original 

income category in the NSHQ 187 

Average income before and after correction 

by number of adults and number at work 

in the NSHQ 188 

Disparities in risk of income poverty by 

region, tenure, household size, economic 

status and social class from the LII Survey 

and the NSHQ 189 

Table A2.10: Disparities in risk of (modified) consistent 

poverty (60% mean and basic) by region, 

tenure, household size, economic status 

and social class from the LII Survey and 

the NSHQ 191 

xiii

AA.Prelims 7/7/05 6:53 am Page xiv 

List of Maps 

Map 2.1: 

Map 2.2: 

Map 2.3: 

Per cent age-dependent by local authority 

area 15 

Per cent of persons age 65 and over and 

living alone 16 

Labour force participation rate by local 


Map 2.4: Unemployment rate by local authority area 18 

Map 2.5: Per cent of persons in farming 23 

Map 2.6: 

Per cent of farmers farming less than 

30 acres 24 

Map 2.7: Per cent of farmers farming 30–49 acres 25 

Map 2.8: Per cent of farmers farming 50+ acres 26 

Map 2.9: 

Economic dependency rates by local 


Map 2.10: Per cent of persons age 15+ who have left 

school with no education or primary 

education 32 

Map 2.11: Per cent of persons age 15+ who have 

left school with lower secondary education 33 

xiv

AA.Prelims 7/7/05 6:53 am Page xv 

List of Maps 


school with upper secondary education 34 


school with third-level education 35 

Map 2.14: Per cent of persons classified as 

Professional/ Managerial/Technical 39 

Map 2.15: Per cent of persons classified as ‘Other 

Non-Manual’ 40 

Map 2.16: Per cent of persons classified as Skilled 

Manual 41 


Semi-skilled Manual 42 


Unskilled Manual 43 

Map 4.1: Consistent poverty risk by region 63 

Map 4.2: 

Map 4.3: 

Map 4.4: 


authority area (50 per cent poverty line) 67 


authority area (60 per cent poverty line) 68 

Disparities in risk of modified consistent 

poverty by local authority area 72 

Maps 6.1a to 6.1e: Disparities in risk of deprivation by 

regional authority area 101 

Maps 6.2a to 6.2e: Disparities in risk of deprivation by 

local authority area 108 

xv

AA.Prelims 7/7/05 6:53 am Page xvi 

List of Figures 

Figure 7.1: Gross and net ratios of being poor versus nonpoor 

at the 50 per cent income line by tenure 

and urban-rural location 129 

Figure 7.2: Gross and net ratios of being poor versus 

non-poor at the 50 per cent income line by 

tenure and Dublin versus elsewhere 131 


non-poor at the 60 per cent income line by 

tenure and urban-rural location 132 


non-poor at the 60 per cent income line 

by tenure and Dublin versus elsewhere 133 


non-poor at the MCP line by tenure and 

urban-rural location 134 


non-poor at the 50 per cent MCP line by 

tenure and Dublin versus elsewhere 136 

xvi

AA.Prelims 7/7/05 6:53 am Page xvii 


Dorothy Watson is a Senior Research Officer at the 

Economic and Social Research Institute. She managed the 

Living in Ireland Survey and, with James Williams, the 

2001/2002 Survey of Housing Quality and the 2003 National 

Survey of Domestic Abuse. Together with Bernadette Ryan 

and Bertrand Maitre of the ESRI, she is presently working on 

an international consortium to establish the Euro-Panel User 

Network to provide support and avenues of communication 

to users of the European Community Household Panel 

(ECHP). She has been a member of the EU-SILC Task Force 

and a consultant to Eurostat on the EU-SILC Pilot. Her 

current research includes a study of the changes in income 

associated with retirement (with Gerry Hughes), the 

development of a questionnaire for a national study of 

disability in Ireland (with Brian Nolan and James Williams) and 

an analysis of the spatial distribution of poverty and 

disadvantage in Ireland (with James Williams and Chris 

Whelan). 

Christopher T. Whelan is a Research Professor at the 

Economic and Social Research Institute. His current research 

interests include the causes and consequences of poverty 

and inequality, the measurement and monitoring of poverty 

and social exclusion, social mobility and inequality of 

opportunity and quality of life. He is involved in a range of 

xvii

AA.Prelims 7/7/05 6:53 am Page xviii 


Irish and European projects involving analysis of the Living in 

Ireland Panel Survey and the European Community 

Household Panel Study. These activities include acting as a 

co-ordinator of an EU-funded network of European 

institutions focusing on the inter-disciplinary study of 

Economic Change, Unequal Life-Chances and Quality of Life 

(CHANGEQUAL). He has contributed on these topics to a 

range of academic journals. He is currently an associate 

editor of the European Sociological Review and is 

Chairperson of the Standing Committee for the Social 

Sciences of the European Science Foundation. 

James Williams is a Research Professor at the Economic 

and Social Research Institute and is also Head of the 

Institute’s Survey Unit, where he is responsible for all aspects 

of survey-based data collection. From 1993 to 1996 he 

managed the Irish component of the European Community’s 

Household Panel Survey (ECHP). In that role he had a major 

involvement with Eurostat in the planning and implementation 

of the survey. He has undertaken an extensive range of 

commissioned research and consultancy projects and has 

both contributed numerous chapters to books and coauthored 

others. He has also been involved in a wide range 

of commissioned reports and consultancy projects - both 

national and international - spanning a diverse range of 

topics. He also carries out both ad hoc and ongoing surveys 

for a range of organisations and bodies, mostly Irish public 

sector, European Commission and Eurostat. 

Sylvia Blackwell has been the Surveys Executive in the 

ESRI Survey Unit since 2000, where she has worked on 

business projects such as the Annual Business Surveys coordinated 

by Forfás among client bases of the IDA; Enterprise 

xviii

AA.Prelims 7/7/05 6:53 am Page xix 


Ireland; Údarás na Gaeltachta and Shannon Development 

and the Survey of Business Expenditure on R & D within 

Ireland. More recently she has worked on such projects as 

the Dublin Docklands Authority on their Educational 

Database, a Survey of Employers’ and Employees’ Views and 

Experiences and also on the National Survey of Vacancies in 

the Private Non-Agriculture Sector and Public Sector. 

xix

AA.Prelims 7/7/05 6:53 am Page xx 

Acknowledgements 

This study forms part of on going research programme being 

carried out by the Economic and Social Research Institute 

(ESRI) that is sponsored by the Department of Social and 

Family Affairs. The study draws on a number of major data 

sources, including the National Survey of Housing Quality. The 

Combat Poverty Agency has provided specific funding for this 

project and we would like to express our appreciation of their 

support. 

The Department of the Environment, Heritage and Local 

Government provided funding for the National Survey of 

Housing Quality 2001/2002, which is an important source of 

data for some of the detailed regional analysis presented here. 

In carrying out this study we have benefited from the 

comments of Jim Walsh and Vanessa Coffey of the Combat 

Poverty Agency and an anonymous referee. 

We also gratefully acknowledge the assistance of Mary 

Dowling, Daphne McNamara Lancha and Pat Hopkins. 

xx

AA.Prelims 7/7/05 6:53 am Page xxi 

Foreword 

Introduction 

Combat Poverty is a state advisory body which develops and 

promotes evidence-based proposals and measures to 

combat poverty. The spatial aspects of poverty are a longstanding 

concern of Combat Poverty, in terms of both 

research and area-based policies. 1 This study updates and 

extends previous research on the spatial distribution of 

poverty using recent national data sources: the Census of 

Population (2002), the Living in Ireland Survey (2000) and, for 

the first time, the National Survey of Housing Quality 

(2001/2002). It applies direct and indirect measures of poverty 

(household income, deprivation, socio-demographic 

1 Brian Nolan, et al. (1998), Where are Poor Households The Spatial 

Distribution of Poverty and Deprivation in Ireland, Dublin: Oak Tree 

Press and Combat Poverty Agency; Brian Nolan and Christopher 

Whelan (1999), Loading the Dice A Study of Cumulative 

Disadvantage, Dublin: Oak Tree Press and Combat Poverty Agency; 

Sarah Craig and Kieran McKeown (1994), Progress through 

Partnership: Final Evaluation Report on the PESP Pilot Initiative on 

Long-Term Unemployment, Dublin: Combat Poverty Agency; Brian 

Harvey (1994), Combating Exclusion: Lessons from the Third EU 

Poverty Programme in Ireland 1989–1994, Dublin: Combat Poverty 

Agency; Barry Cullen (1994), A Programme in the Making: A Review of 

the Community Development Programme, Dublin: Combat Poverty 

Agency. 

xxi

AA.Prelims 7/7/05 6:53 am Page xxii 


indicators) to a range of spatial categories, including 

regional and local administrative units, area type and 

housing tenure. 2 

The study addresses three key aspects of the spatial 

distribution of poverty: 

• identifies patterns with regard to the concentration of 

poverty and how these have evolved over time 

• assesses if these patters are significant in terms of the 

overall incidence of poverty 

• considers the processes underlying poverty clustering, 

distinguishing between factors which influence the 

location and the causes of poverty. 

These questions are also relevant to the government policy 

priority attached to tackling ‘poverty blackspots’ in urban 

and rural areas. The National Anti-Poverty Strategy 

highlights the ‘concentrated and cumulative nature’ of 

poverty in urban and rural areas, which has prompted 

numerous area programmes and related initiatives. 3 The 

uneven distribution of poverty is also linked to a wider 

policy concern with balanced regional development, as is 

reflected in the regional and Peace II programmes under the 

National Development Plan. From this policy perspective, 

2 A fivefold area classification is used of open countryside, village, town, 

provincial city and Dublin. While tenure is not a spatial category per se, 

it does have a spatial expression through the segmented nature of the 

housing market. The categories include owned outright, owned with 

mortgage, tenant purchase, public rented and private rented. 

3 These include the Local Development Social Inclusion Programme, 

Community Development Programme, RAPID Programme and Clár 

Programme. Spatial initiatives also exist to address educational 

disadvantage, drug misuse, unemployment, child care and other forms 

of disadvantage. 

xxii

AA.Prelims 7/7/05 6:53 am Page xxiii 

Foreword 

research on the spatial dimension of poverty is needed to 

accurately identify poor areas in the first instance, to 

understand why these areas exist and to assess the impact 

of area policies. 

Key Research Findings 

The study finds a pattern of regional and county variation in 

poverty-related indicators such as age dependency, 

economic activity, educational attainment and social class. 

This provides an interpretative context for subsequent 

analyses of direct measures of poverty. 

The study identifies an uneven spatial distribution of poor 

households at regional and local levels, with a more 

differentiated pattern of poverty apparent for each 

descending unit of analysis. Thus, the slightly higher poverty 

rate in the NUTS 2 Border, Midlands and West region is 

accentuated in NUTS 3 planning regions such as the Border, 

Mid-West and West. Among local authorities, Donegal, 

Cavan, Leitrim, Longford, Mayo and the provincial cities 

record above-average rates of poverty. 

Turning to area differences, this study finds that the 

countryside and villages have a slighter higher poverty risk, 

especially compared to Dublin and the larger cities. The 

differential for income poverty has worsened in recent years, 

though the variation in the consistent poverty measure has 

narrowed. 

Tenure differences are very marked, with local authority 

tenants at particular risk of poverty and deprivation (33 per 

cent compared with an average of 6 per cent). Furthermore, 

their relative position has worsened over time though the 

xxiii

AA.Prelims 7/7/05 6:53 am Page xxiv 


overall level of poverty has fallen and their share of the 

housing market has declined. Another high-risk category is 

tenants in private housing, who have twice the average 

poverty rate. 

The ranking of regional and local authorities varies a lot 

depending on the specific deprivation measures. Thus, 

Dublin, which otherwise has a low poverty risk, scores high in 

housing and environmental items. Deprivation disparities by 

tenure are striking, with public tenants experiencing the 

highest levels. Furthermore, these disparities are consistent, 

indicating a pattern of multiple deprivation among this 

housing tenure. 

In summary, poverty remains a spatially diffuse phenomenon, 

with variations in the risk of poverty being counterbalanced 

by a more even incidence of poverty as higher-risk areas tend 

to have a lower share of the total population. The greatest 

poverty concentration is by housing tenure, with local 

authority and private rented tenant households accounting for 

60 per cent of all households experiencing poverty and 

deprivation, though only representing 17 per cent of the total 

population. 

The study is constrained by data limitations in exploring the 

micro distribution of poverty within counties. However, it is 

likely that the more acute poverty clusters at this spatial level 

represent a relatively small proportion of the total numbers in 

poverty. 

In reviewing the factors which underlie the observed spatial 

patterns in poverty, the study finds little evidence of a causal 

relationship with particular locations or tenures. Where 

variation exists, the determining factors relate to the socioeconomic 

composition of poor households rather than 

xxiv

AA.Prelims 7/7/05 6:53 am Page xxv 

Foreword 

location per se. The significance of tenure over other spatial 

units is its selection effect: the housing market, aided by 

public policy, reserves public and social rented housing for 

low-income households in the main. However, the study does 

indicate an additional tenure effect in urban areas which 

warrants further investigation. 

Policy Implications 

The main policy implications arising from this study identified 

by Combat Poverty are as follows. 

• Policies to tackle poverty must continue to prioritise the 

structural causes of the problem over a focus on the 

spatial outcomes. 

• Area programmes cannot be justified on a targeting basis 

alone given the diffuse nature of poverty, though other 

rationales have merit. 

• Policies should distinguish between ‘people poverty’, 

which is linked to structural factors, and ‘place poverty’, 

which reflects more local issues. 

• Greater policy emphasis must be placed on 

neighbourhood issues, such as access to services and 

estate management in area regeneration policies. 

• Action must be taken to address the segregated nature of 

the housing market, which goes beyond a concern with 

planning issues to address core inequalities. 

• Regional and local poverty trends should be monitored 

and targets set for poverty reduction in line with national 

targets. 

xxv

AA.Prelims 7/7/05 6:53 am Page xxvi 


This study reaffirms that poverty is primarily a structural, 

rather than a spatial, phenomenon. It arises from sociodemographic 

variables associated with an increased risk of 

poverty, e.g. unemployment and low-paid work, low 

educational attainment, old age, child dependency and lone 

parenthood. The spatial distribution of poverty largely reflects 

variations in the existence of these variables among the 

population. Policies to tackle poverty must prioritise these 

structural issues. At the same time, there is a positive 

message in the report’s rejection of the idea of a spatially 

concentrated ‘underclass’ that is cut off socially and 

economically from the rest of society. 

From a targeting perspective, spatial categories are of limited 

benefit, especially as compared with a traditional focus on 

socio-demographic factors. While it may be possible to 

distinguish micro clusters of poor households, often 

associated with public housing estates, these are not 

significant in terms of the total number of households in 

poverty. Spatial programmes have other policy rationales 

besides targeting. They can improve the delivery of existing 

services through promoting greater integration, community 

access and user involvement and develop new outreach 

mechanisms to meet the needs of vulnerable groups. They 

also achieve a multiplier effect by combining investment in 

physical regeneration with economic and social initiatives. 

More attention should be focused on the added value of 

area-based interventions in relation to these objectives as 

distinct from a crude targeting mechanism. 

Targeting is relevant when applied to tenants in local authority 

and private accommodation. Even then, this must be 

contextualised in wider policy issues affecting this group, 

especially in urban areas. The main challenge is the selection 

xxvi

AA.Prelims 7/7/05 6:53 am Page xxvii 

Foreword 

effect associated with social housing. While there has been 

a policy shift in regard to the scale of social housing in order 

to avoid large concentrations, there has been no change in 

the restrictive allocation of social housing. Furthermore, the 

dynamic nature of social housing ensures a constant 

recruitment of poor tenants to replace those who, having 

improved their situation, move on. In addition, there is an 

internal process within social housing whereby tenants 

move from different estates, usually from low-demand to 

high-demand areas. These often reflect differences in how 

estates are managed as much as the quality of housing. 

Previous research indicates that within local authority 

estates, there can be subtle differences between individual 

streets and blocks. 4 Urgent measures are required to 

counter this internal process of ghettoising vulnerable 

households. 

A stronger focus on ‘place’ or neighbourhood poverty is 

warranted. The residualisation of public housing has resulted 

in tenants experiencing inferior housing quality and a range 

of other neighbourhood drawbacks. The discontinuity 

between regeneration initiatives and social programmes is 

disconcerting, with the exception of some localised 

initiatives, e.g. Ballymun. One example of how the physical 

and social dimensions can be combined is in regard to fuel 

poverty. If combined with social programmes, energy 

efficiency measures can have a multiplier effect for poor 

tenants. For this synergy to develop, local authorities, along 

with other supporters and providers of social housing (health 

4 Tony Fahey (ed.) (1999), Social Housing in Ireland: A Study of Success, 

Failure and Lessons Learned, Dublin: Oak Tree Press in association 

with Combat Poverty Agency and the Katharine Howard Foundation. 

xxvii

AA.Prelims 7/7/05 6:53 am Page xxviii 


boards and voluntary housing associations) should have a 

role in designing and delivering area programmes which 

address this high-risk poverty sector. 

A neighbourhood effect is also likely to arise from the 

limited access to services and opportunities for residents 

of poor areas. For example, employer recruitment 

practices may discriminate against people from certain 

localities, either through crude selection policies or 

because of limited access to informal recruitment 

networks. Also, the quantity and quality of services 

provided can be less than are available in other, more 

affluent areas, e.g. health care, schools, financial and retail 

services, sports and recreation facilities, transport. Antidiscrimination 

monitoring of public and private services is 

required to address these area effects. Another factor is 

the level of social capital, as reflected in voluntary activity 

and social networking, in poor neighbourhoods. 

Finally, there is a need to monitor poverty trends at 

regional and local levels on an ongoing basis and to relate 

these to national trends. The provision of sub-national data 

on poverty provides a baseline to measure progress and to 

inform regional and local poverty strategies. Regional and 

local poverty reduction targets should also be set in line 

with the national target to reduce the combined poverty 

and deprivation measure to 2 per cent by 2007. It is 

possible to adapt this target based on reducing the 

poverty differentials that exist between regional and local 

authorities. These targets should be set in the context of 

administrative authorities drawing up plans to reduce 

poverty in their areas as part of an overall national 

strategy. 

xxviii

AA.Prelims 7/7/05 6:53 am Page xxix 

Foreword 

Data Issues 

Geographical analysis of poverty remains underdeveloped. 

A key constraint is the lack of poverty-specific data on a 

spatially disaggregated basis. The limitations of survey data 

as used in this study have already been highlighted. The 

Census of Population, though inclusive of all households, has 

acute drawbacks in regard to the measurement of poverty 

and the spatial units of analysis. The Department of the 

Environment, Heritage and Local Government recently issued 

a consultation paper on spatial data infrastructure. Various 

avenues should be explored to gather specific data on 

poverty at the local level, including the Census of Population. 

A related task is to develop a system for the gathering and 

presentation of existing poverty-related data at the local level, 

e.g. welfare payments, medical cards, unemployment 

records. In Northern Ireland and the United Kingdom, quite 

advanced methodologies have been developed to analyse 

poverty-related indicators. In Ireland, Hasse has developed 

earlier work by Combat Poverty to produce a composite 

index of deprivation based on the Small Areas Statistics 

Population from the Census of Population. 5 Further work is 

required to refine this methodology and to incorporate other 

data sources so as to create a comprehensive spatial 

database on poverty. 

Separately, there is a requirement for better research and 

monitoring of poor neighbourhoods. A dedicated study is 

required which would address two key issues: what is the 

5 Trutz Haase (1999), ‘Affluence and Deprivation: A Spatial Analysis 

Based on the 1991 Census of Population’, in Dennis Pringle et al. 

(eds.), Poor People, Poor Places, Dublin: Oak Tree Press and the 

Geographical Society of Ireland. 

xxix

AA.Prelims 7/7/05 6:53 am Page xxx 


neighbourhood effect on clusters of poor households, and 

what is the impact of government policies on the living 

conditions in poverty clusters An earlier study by Combat 

Poverty and the Katharine Howard Foundation demonstrates 

the potential of micro studies in understanding the factors 

shaping local conditions. 6 Ideally, this type of study should be 

established on a longitudinal and a comparative basis so that 

change over time can be tracked and comparisons made 

between the evolution of different localities. Such a study will 

be considered in Combat Poverty’s strategic plan 2005–2007. 

6 Tony Fahey (ed.), op cit. 

xxx

AA.Prelims 7/7/05 6:53 am Page xxxi 


Is poverty concentrated in certain ‘black spots’ If so, what is 

the extent of this concentration, what causes such 

concentration and which areas are particularly at risk These 

are important questions for policy analysts because of their 

implications for strategies to combat poverty and promote 

social inclusion. 

These are the questions addressed by this report on the 

spatial distribution of poverty in Ireland. The report was 

commissioned by the Combat Poverty Agency to look at the 

following questions: 

• Is poverty concentrated in particular areas If so, what is 

the extent of that concentration 

• Where are the concentrations of poverty 

• Why are there concentrations of poverty 

• What are the characteristics of such concentrations of 

poverty 

• What can be done about reducing such concentrations of 

poverty 

The study brings together data from three national sources: 

the 2002 Census, the 2000 Living in Ireland Survey, with very 

xxxi

AA.Prelims 7/7/05 6:53 am Page xxxii 


detailed information on income and living standards for a 

sample of 3,400 households, and the 2001/2002 National 

Survey of Housing Quality, with a sample of over 40,000 

households. 

A key feature of the report is the concern with the causal 

processes underlying any association between area and 

poverty. Targeting programmes to combat poverty need 

to be fundamentally concerned with the causes of poverty 

rather than with patterns that are accidental. This is 

particularly important from the point of view of area-based 

programmes. Without an understanding of why people in 

certain areas experience a greater risk of poverty, 

programmes to improve their situation cannot be properly 

targeted to those in need. 

Key Findings 

There are clear regional and local differences in both the risk 

of poverty and in the levels of key related social indicators 

such as unemployment, education and social class. Further, 

the differences in risk become larger as we move to lower 

levels of aggregation, e.g. from region to county level. The 

highest poverty risk is found in counties Donegal, Leitrim and 

Mayo and the lowest in the counties around Dublin. However, 

there is considerable diversity within administrative planning 

regions. Louth and Sligo, for instance, show much lower 

relative poverty risk than the other counties in the Border 

region. This points to the importance of avoiding broad 

generalisations based on administrative units which are 

based on historical and political rather than socio-economic 

categories. 

xxxii

AA.Prelims 7/7/05 6:53 am Page xxxiii 


However, the scale of differences based on geographic 

location is rather modest compared to the differences 

between socio-economic groups or people with different 

housing tenures. Furthermore, the pattern varies depending 

on the dimension of deprivation being considered. For 

instance, Dublin, which has a low income poverty risk, scores 

high in terms of housing and environmental deprivation. On 

the other hand, Donegal has a much higher income poverty 

rate than the national average, but it is less likely than 

average to suffer deprivation in terms of housing deterioration 

or environmental deprivation. 

There are modest differences between urban and rural areas, 

but large differences by housing tenure. Local authority 

renters fare worst in terms of almost all measures of 

deprivation. This is largely due to a selection effect, whereby 

the dynamics of the housing market and public housing 

policy result in public and social rented housing being 

reserved for low-income households. 

Evidence of multiple deprivations structured along spatial 

lines is extremely weak. Over time, income poverty rates for 

local authority tenants have increased sharply, and while their 

consistent poverty rate (which takes account of living 

standard as well as income) has shown some improvement, 

they have fallen farther behind homeowners. The decline in 

their numbers, however, means that they constitute a smaller 

proportion of the poor. 

What accounts for the differences between areas at risk of 

poverty and deprivation The key factors accounting for 

poverty and deprivation are socio-economic: unemployment, 

non-participation in the labour force due to old age or illness, 

lone parenthood, low levels of education and social class. 

xxxiii

AA.Prelims 7/7/05 6:53 am Page xxxiv 


However, the study does indicate an additional effect of 

tenure type. Differences between geographic areas in terms 

of poverty and deprivation are largely due to differences 

between these areas in the socio-economic composition of 

their populations. 

Implications 

The study clearly demonstrates that poverty is a structural 

rather than a spatial phenomenon. It arises from socioeconomic 

processes such as unemployment and low-paid 

work, low educational attainment, old age, child dependency 

and lone parenthood. The spatial distribution of poverty 

largely reflects spatial variations in these variables. Policies to 

tackle poverty must prioritise these structural issues. 

Area-based policies have little role in targeting poor 

households – most poor households do not live in clearly 

identifiable geographically concentrated areas. However, 

there may be a role for area-based initiatives in local authority 

estates, where they may enhance service delivery through 

promoting greater integration, community access and user 

involvement. The emphasis of such programmes should be 

on the efficient delivery of services and mobilisation of 

community resources rather than targeting poor households. 

Such locally based programmes may also be used to provide 

an outreach mechanism to meet the multiple needs of 

vulnerable populations, such as long-term unemployed, exprisoners, 

former drug users and early school-leavers. More 

attention should be focused on the added value of areabased 

interventions in relation to these objectives rather than 

as a simple targeting mechanism. 

xxxiv

Chapter 01 7/7/05 6:53 am Page 1 

Chapter 1 

Introduction 

1.1 Background to the Study 

In 1998 the Combat Poverty Agency published Where are 

Poor Households The Spatial Distribution of Poverty in 

Ireland by Nolan, Whelan and Williams. The study employed 

data for the early to mid-1990s. It was conducted in the 

context of a growing concern about the extent of 

geographical concentration of poverty alongside increasing 

fears that such concentration was associated with processes 

of cumulative disadvantage of a vicious cycle nature. It 

addressed the extent to which poverty in Ireland is 

concentrated in particular areas or types of areas, the nature 

of the causal processes underlying the spatial distribution of 

poverty and the implications for policy. 

The key questions of the study were: 

• Is poverty concentrated in particular areas If so, what is 

the extent of that concentration 

• Where are the concentrations of poverty 

• Why are there concentrations of poverty 

• What are the characteristics of such concentrations of 

poverty 

1



• What can be done about reducing such concentrations of 

poverty 

Ireland changed greatly over the second half of the 1990s, 

during which an unprecedented economic boom was 

experienced. Data are now coming on stream to allow these 

issues, which still remain highly relevant, to be addressed 

once again in light of these changes. These data comprise 

not only updated census data but a national survey that 

enables us to address these issues on a sample of 

households whose number goes well beyond anything that 

has been available to us in the past. 

Much of the debate about the impact of economic 

restructuring, industrial modernisation and infrastructural 

development in Ireland has a strong spatial dimension. 

Concern is expressed that the effects of socio-economic 

change vary not only by socio-economic group but also by 

location. In popular discourse in Ireland, certain places (often 

quite vaguely defined) are often said to have fared badly, 

such as ‘the West’, ‘remote rural areas’, ‘new suburbs’ and 

‘poverty black spots in the city’. More recently, concern in a 

number of European countries with vicious cycle processes 

of cumulative disadvantage and the emergence of an urban 

‘underclass’ has led to issues relating to spatial segregation 

figuring prominently in popular and academic debates. 

Since the early 1990s there has been a growing emphasis on 

policy options involving spatial programmes aimed at tackling 

unemployment, poverty and social exclusion. Following a 

number of pilot schemes, area-based programmes have 

become a significant part of government policy aimed at 

tackling poverty. As Walsh (1999: 279) notes, this is clearly 

evidenced in the National Anti-Poverty Strategy (1997), which 

2


Introduction 

exhibits an explicit spatial dimension in two of its five priority 

themes: disadvantaged urban areas and marginalised rural 

communities. In 1992 the Irish government, in collaboration 

with the European Commission, established Area 

Development Management (ADM). Its primary objective was 

to promote social inclusion, reconciliation and equality and to 

counter disadvantage through local, social and economic 

development. As Haase, McKeown and Rourke (1996) 

document, the initiative was representative of a concern to 

develop area-based initiatives in response to emerging 

evidence relating to unemployment black spots. 

The number and range of projects managed by ADM has 

evolved considerably since 1992. Among its current 

programmes are the Local Development Social Inclusion 

(LDSIP) and Revitalising Areas by Planning Investment and 

Development (RAPID) programmes. The former is a National 

Development Programme aimed specifically at addressing 

social inclusion issues at local level. The LDSIP provides 

funding to Partnership Community Groups and Employment 

Pacts that adopt a partnership approach to tackling local 

issues on the basis of comprehensive, integrated local action 

plans. The RAPID programme is intended as a response to 

the need for more and better-targeted investment in 

disadvantaged areas. The focus in terms of service delivery is 

on integrating services more efficiently and tailoring them to 

community needs. It is also intended to encourage 

investment in new facilities and services. Among the 

objectives identified as fundamental to the RAPID programme 

are the development of an integrated focus on social groups 

experiencing cumulative disadvantage, reduction in spatial 

concentration of poverty, unemployment and social exclusion 

and the mobilisation of social capital and capacity for 

3



economic and social development. Other programmes 

include the Equal Opportunities Childcare Programme (EOCP) 

and the Millennium Partnership Fund for Disadvantage. 

One of the important considerations, but by no means the 

only consideration, in the allocation of funding under these 

programmes is the socio-demographic profile of the 

geographical areas. This can involve the use of indices of 

deprivation, which rank district electoral divisions (DEDs) 

according to indicators such as unemployment, education, 

class composition and housing quality. 

This discussion of ADM activities illustrates that while area 

interventions all appear to share the objective of targeting 

scarce resources, the rationales or justifications associated 

with such interventions are variable and can take on a good 

deal of complexity. In our earlier report, Where Are Poor 

Households (Nolan, Whelan and Williams 1998), we pointed 

to the need to consider a number of related, but still relatively 

distinct, arguments for focusing on the spatial distribution of 

poverty. Some of these involve specific assumptions 

regarding the causal basis of poverty, while others take a 

more pragmatic form. The most straightforward justification is 

based on the assumption that if poor households are highly 

concentrated in specific areas, then it is possible to target 

resources on these areas in order to maximise the number of 

households reached. Thus, the argument outlined in the 

NESC (1990) document Strategies for the Nineties was based 

on the assumption that general measures to improve 

employment creation would not be sufficient to have an 

impact on those experiencing long-term unemployment. 

Special employment measures were required which would be 

‘targeted in an integrated fashion in the context of local areabased’ 

strategies. Increased polarisation, and its spatial 

4


Introduction 

manifestations, provides an important component of the 

rationale for spatial interventions. 

1.2 Rationales for Spatial Intervention 

A recent review of area-based targeting by Tunstall and 

Lupton (2003) distinguishes between five different rationales. 

• The first focuses on ‘efficiency’ and completeness in 

reaching poor individuals and derives its logic from the 

concentration of deprivation and disadvantage. Thus, as 

Walsh (1999: 283) notes, developments within mainstream 

welfare policy have encouraged a greater local focus on 

the design and delivery of services. In the context of a 

growing desire among policy makers for better targeting 

of resources, spatial programmes seem to offer an 

attractive means of responding to social needs. Here, 

evidence relating to the distribution of unemployment, 

unskilled manual work, lack of educational qualifications 

and low income plays a crucial role. 

• The second rationale is based on the argument that 

concentrated poverty may have cumulative and 

qualitatively different effects on individuals, organisations 

and infrastructure than less concentrated poverty. Poverty 

‘black spots’ could result in a qualitatively different 

experience of poverty in terms of factors such as physical 

and mental health, degree of economic strain and 

alienation from social and political participation, ranging 

from social isolation to declining church attendance and 

participation and confidence in the political process. Thus, 

NESC (1993) referred to a localised process of interaction 

between labour market, education, housing and 

5



environmental factors, which is most likely to be 

addressed effectively by an area-based strategy. This 

rationale could provide the justification for interventions 

offering support targeted not just at individuals, but 

organisations and infrastructure. A crucial objective 

from this perspective is to provide a focus for 

enhancing service provision in responses to multiple 

deprivation. Local co-ordination is a potentially effective 

means of improving the quality of services. Initiatives of 

this kind have been directed at service co-ordination 

relating to areas such as labour market and training 

access and indebtedness and access to credit. 

• The third justification relates to the choice of areas as a 

form of rationing by taking advantage of the fact that 

targeting areas may be a good deal simpler than 

targeting individuals. As Tunstall and Lupton (2003: 4–5) 

observe, in some cases governments make new money 

available for a specific purpose which will only be 

targeted towards areas with specific needs, though 

usually fewer than the total number of areas with such 

needs are reached on account of limited resources. 

• The fourth relates to the use of area-based initiatives as 

a form of piloting. In assessing such schemes’ 

potential, capacity to innovate or deliver may also be 

taken into account. Thus, Walsh (1999: 288) notes that 

initiatives such as the Area-Based Response to Long- 

Term Unemployment (ABR) and its linked programme, 

the Global Grant for Local Development (GGLD), 

operated on a pilot basis between 1991 and 1995 

before being subsumed into the Local Development 

Programme. 

6


Introduction 

• The fifth rationale focuses on additional benefits deriving 

from area initiatives such as community involvement and 

the development of partnerships. Among other things, this 

justification draws attention to the particular importance of 

co-ordination and more accurate identification of needs in 

deprived areas. Thus, the integration and co-ordination of 

existing services on the basis of partnership arrangements 

may be seen as of particular value in areas where 

deprivation takes a multidimensional form. The NESC 

document, referred to earlier, argues that intensive coordinated 

spatial interventions containing elements of 

housing and environment improvements, as well as 

retraining and employment schemes, could have an 

impact over and above their separate effects. In addition, 

it was argued that participation of local communities in 

the planning and delivery of area-based projects would 

help ensure that they more accurately reflected local 

needs and priorities. The community development 

rationales for spatial intervention attempt to develop the 

argument that involvement of local communities in the 

process of economic and social change has an intrinsic 

value. Thus, rather than such involvement being peripheral 

to a process of economic change, it is seen as a crucial 

means by which marginalised groups can come to share 

in the benefits of economic progress. This thesis relies not 

just on arguments relating to effectiveness and efficiency, 

but rather bases its arguments on an understanding of 

social exclusion which identifies loss of skills, selfconfidence 

and motivation as crucial elements in the 

process. 

A particular justification for targeted intervention does not 

necessarily involve a commitment to a specific understanding 

7



of the causal processes. However, almost inevitably, some 

such set of assumptions underlies spatial interventions. It is 

possible, though, to see spatial concentration of deprivation 

as having no causal significance or distinctive consequences. 

Instead, concentration could be seen to arise simply as a 

consequence of variations across other genuinely causal 

variables such as human capital. Alternatively, location could 

be thought to play a potentially independent role in a number 

of ways. Thus, employers’ hiring behaviour could mean that 

merely residing at a particular address could increase one’s 

risk of unemployment and poverty at a given level of human 

capital. More broadly, those coming from areas where the 

resource stock, in terms of access to training, education, 

financial institutions, etc., is poorer could be seen as being 

additionally disadvantaged. Much more controversially, at the 

centre of recent debates concerning the creation of a 

spatially concentrated underclass is the highly contested idea 

that persistent poverty is transmitted through a fundamental 

altering of norms and ‘tastes’ in relation to welfare 

dependency, employment commitment and non-marital 

fertility. 

It is useful to keep in mind that area-based responses to 

poverty are not an entirely new phenomenon. Pringle (1999) 

notes that many of the initiatives pursued in Ireland in recent 

years, while innovative in the Irish context, are not entirely 

dissimilar to those introduced in others countries in the 1960s 

and 1970s. We should also acknowledge that there is a 

significant literature that is sceptical regarding the extent of 

the contribution that spatial strategies can make to reducing 

or ameliorating poverty. However, it seems more sensible to 

defer consideration of that debate until we have set the 

evidence for the nature and significance of spatial variation in 

poverty and deprivation. 

8


Introduction 

1.3 Data Sources 

In the analysis that follows we shall seek to provide a 

relatively up-to-date account of the spatial distribution of 

poverty in Ireland and critically assess its implications for 

the range of issues that we have outlined. In so doing we 

draw on a number of different data sources. The 1998 study 

used several distinct data sources. The first was the 1994 

Living in Ireland Survey, where results could be compared 

with those from a comparable household survey conducted 

by the ESRI in 1987. The second source was the Small Area 

Population Statistics (SAPS) derived from the Census of the 

Population and available for 1986 and 1991, constituting the 

most geographically disaggregated and comprehensive 

information on the socio-demographic structure of the 

country. In seeking to look once again at the spatial 

concentration of poverty, we can now make use of the 

SAPS data from the 2002 Census, the Living in Ireland 

Survey conducted in 2000 and a major new source of 

information that has become available, namely the 

2001/2002 National Survey of Housing Quality (NSHQ). The 

NSHQ was conducted by the ESRI on behalf of the 

Department of the Environment and Local Government in 

over 40,000 households. By combining these sources we 

hope to provide a reasonably comprehensive account of the 

spatial distribution of poverty and deprivation, making use of 

both direct and indirect measures of deprivation. 

Furthermore, the fact that the NSHQ combines a very large 

sample with detailed household information will allow us to 

go beyond the analysis reported in the previous study in 

some important respects. While the NSHQ does not allow 

for the same level of disaggregation as the SAPS, crucially it 

contains measures of income and deprivation for the 

9



participating households, which can be analysed at 

household level. It is precisely the absence of such 

information in the data made available from SAPS that leads 

to the use of composite indicators for small areas as 

indicators of socio-economic deprivation, about which 

researchers have expressed serious reservations. This issue 

is taken up in detail in the analysis of the 2002 SAPS that 

follows in Chapter 2. 

1.4 Outline of the Study 

In Chapter 2 we consider geographical variations in some of 

the underlying socio-demographic correlates of poverty and 

deprivation risk as contained in the 2002 Census of 

Population. Findings are reported at the level of the 34 

counties and county boroughs in the Republic of Ireland. In 

Chapter 3 we focus on measurement of poverty and 

deprivation at the household level. In so doing we draw 

primarily on the series of Living in Ireland Surveys but also 

make use of the National Survey of Housing Quality. Chapter 

4 documents variation in poverty and deprivation by regional 

authority and local authority areas. In Chapter 5 we deal with 

variations in poverty by area type in the sense of population 

density and type of housing tenure. In Chapter 6 we focus on 

non-monetary deprivation. In particular we address the issue 

of whether different types of deprivation tend to vary in a 

similar manner across spatial categories. 

A further question that is explored is the extent to which there 

is evidence of concentrations of multiple disadvantage within 

particular areas. 

10


Chapter 2 

Spatial Trends in Deprivation 

Surrogates 

2.1 Introduction 

In this chapter we consider geographical variations in some 

of the underlying socio-demographic correlates of poverty 

and deprivation risk as contained in the 2002 Census of 

Population. The variables considered are largely consistent 

with those identified in other studies, e.g. Nolan, Whelan and 

Williams 1998; Nolan et al. 2002, as the main drivers of 

poverty and deprivation. By presenting this type of analysis 

we can provide an understanding of spatial variations in 

socio-demographic structures throughout the country and, 

accordingly, an interpretational framework for much of the 

subsequent analysis. 

Four main areas are considered below, as follows: 

• age structures 

• economic status and activity (including levels of 

farming) 

• levels of educational attainment 

• social class structures. 

11



In the analysis that follows, we present figures and maps at 

the level of the 34 counties and County Boroughs in the 

Republic of Ireland. The maps presented are based on 

quintile distributions at county level, which explains why each 

map has five categories in its legend. Each of the five 

categories contains approximately seven counties. Therefore, 

the numeric range for each category is different from one 

map to the other. The legends distribute counties in five 

groups from the lowest to highest incidence of risk of the 

specific indicators under consideration, but they do not 

purport to have an equal distribution in terms of average rates 

or levels. Hence the maps provide a visual, graphic 

representation of a ranking of the county-level data from 

highest to lowest across the country in terms of the five 

categories. Each distribution must, of course, have a high 

and low point. To properly interpret the maps, the reader 

must take account of the difference between the highest 

figure in the top quintile and the lowest in the bottom quintile. 

The authors would also point out that we are acutely aware 

that the county distributions may mask very substantial subcounty 

variations at, for example, a rural/urban district 

(RD/UD) or district electoral division (DED) level. There are 

approximately 220 RD/UDs in the country and 3,400 DEDs. 

While it would clearly be preferable to undertake the analysis 

contained in this chapter at that level of spatial 

disaggregation, it would be well beyond the scope of the 

current report to do so. 

2.2 Age Structures 

Regional variations in age structures are outlined in Table 2.1. 

(Further county variations are shown in Appendix 1, Table 

A1.1.) Details are presented on the percentage of persons in 

12


Spatial Trends in Deprivation Surrogates 

Table 2.1: Percentage of persons in each region according to age cohort 

Persons 

65+ 

Age living 

0–14 15–29 30–44 45–64 65+ dependent alone 

Region % Rank % Rank % Rank % Rank % Rank % Rank % Rank 

Border 22.5 3 22.3 8 21.0 7 21.7 5 12.5 2 35.0 1 3.5 1 

Midlands 23.0 2 22.5 6 21.8 4 21.0 6 11.6 5 34.6 2 3.1 3 

West 21.1 6 23.0 5 20.9 8 22.0 1 12.9 1 34.0 3 3.4 2 

Dublin 19.2 8 27.7 1 22.7 2 20.2 8 10.2 7 29.3 8 2.6 7 

Mid-East 23.4 1 23.6 3 24.2 1 20.6 7 8.2 8 31.6 7 1.9 8 

Mid-West 21.2 5 23.8 2 21.4 6 21.9 3 11.6 5 32.8 5 3.0 5 

South-East 22.1 4 22.4 7 21.8 4 21.9 3 11.8 4 33.9 4 3.0 5 

South-West 20.8 7 23.3 4 21.9 3 22.0 1 12.0 3 32.7 6 3.0 3 

Total 21.1 24.4 22.1 21.2 11.1 32.3 2.9 

Source: Census 2002 

13



each of five age cohorts, i.e. 0–14 years, 15–29 years, 30–44 

years, 45–64 years and 65 years and over. At a broad regional 

level one can see that the West, Border and South-West 

regions have the oldest age profiles, with 12–13 per cent of 

their populations aged over 65 years. At the other extreme, 

the Mid-East region has the lowest proportion (8.2 per cent) 

aged over 65, followed by the Dublin region with 10.2 per 

cent. 

From the details of Appendix 1, Table A1.1 one can see that 

the counties with the oldest population structures include 

Leitrim (16.1 per cent aged 65 and over), Roscommon (15.5 

per cent), Mayo (14.7 per cent) and Cavan (13.8 per cent). 

An alternative, and more succinct, way of considering 

imbalances in age structures is with the concept of age 

dependency. This is expressed as the ratio of the combined 

population aged 0–14 years and 65+ years to the total 

population. As such, it gives a summary measure of the ratio 

of the dependent to the non-dependent population in a 

region. The summary age dependency figures are shown in 

Map 2.1 as well as Tables 2.1 and Appendix 1, Table A1.1. 

Map 2.1 graphically illustrates that there is a much higher 

level of age dependency in the Border and Western regions, 

especially in Donegal, Mayo, Roscommon, Leitrim, Longford 

and Cavan. It is equally obvious that the lowest levels are 

found in the Dublin and Mid-East regions. 

Details on the percentage of persons aged 65 years or more 

and living alone are outlined in Map 2.2. The concentrations 

of elderly persons in parts of the Border, Midland and 

Western (BMW) Region is very clear from the map, 

particularly in Leitrim, Roscommon, Mayo, Cavan and 

Longford. Rates of 4.0 to 5.1 per cent are obvious in each of 

14



Map 2.1: Per cent age-dependent by local authority area 

these counties, compared with a national total of 2.9 per 

cent. This implies, for example, that Leitrim, with 5.1 per cent 

of its population aged 65 years and living alone, is ‘overrepresented’ 

to the order of 76 per cent in terms of this 

elderly segment of its population. 

15



Map 2.2: Per cent of persons age 65 and over and living alone 

 

In summary, therefore, Maps 2.1 and 2.2 indicate a 

substantial over-representation of older persons in the Border 

and West regions. 

2.3 Economic Status and Activity 

Commonly used measures of disadvantage, deprivation and 

socio-economic imbalance are labour force participation 

rates, unemployment rates and levels of economic 

dependency. 

16



Table 2.2 and Appendix 1, Table A1.2 along with Maps 2.3 and 

2.4 present details on labour force participation and 

unemployment rates. One can see from these that participation 

rates are highest in the Mid-East, Mid-West and Dublin. For 

example, counties Fingal and South Dublin have the highest 

levels of labour force participation with rates of 64.8 and 64.6 

per cent, respectively. 

Map 2.3: Labour force participation rate by local authority area 

17



The lowest rates are in Cork County Borough, Leitrim, 

Roscommon, Mayo, Donegal and Longford, all in the range of 

52–54 per cent – nearly six percentage points below the 

national average. Generally, these are the areas identified in 

the previous section as having an above average percentage 

of their population in the older cohorts, most of whom are 

retired rather than economically active. 

Map 2.4: Unemployment rate by local authority area 

18



Unemployment rates among labour force participants are also 

outlined in Table 2.2 and are plotted in Map 2.4. These show 

that the national rate was 8.8 per cent. Rates are the highest 

in the Border region (12 per cent) followed by the South-East 

(9.6 per cent), with Dublin and the West having the lowest per 

cent (5.2 per cent). 

Table 2.2: Regional labour force participation and 

unemployment rates 

Labour force Unemployment 

participation rate 

rate 

Rate Rank Rate Rank 

Border 56.1 6 12.0 1 

Midlands 57.4 4 9.1 3 

West 55.6 8 5.2 7 

Dublin 61.2 3 5.2 7 

Mid-East 61.5 1 6.8 6 

Mid-West 61.5 1 8.2 4 

South-East 56.6 5 9.6 2 

South-West 56.0 7 8.2 4 

Total 58.3 8.8 


Some of these regional trends are strongly influenced by 

individual counties. For example, the unemployment rates in 

Donegal and Louth are significantly high compared to the 

national average of 8.8 percent, being 15.6 per cent and 13.2 

per cent, respectively. The levels in the county boroughs of 

Limerick and Cork are also particularly high. 

An alternative way of considering the spatial distribution of 

the unemployed is presented in Table 2.3 (the regional level) 

19



and Appendix 1, Table A1.3 (the county level). The tables 

show the percentage of the total number of unemployed 

persons according to their distribution across regions and 

counties. For comparative purposes the table also shows the 

distribution of all persons aged 15 years and over who are in 

the labour force. If the unemployed were distributed on a pro 

rata basis across the counties in line with the total labour 

force, the percentage figures in Columns A and B of the table 

would be identical. The disparity between the figure in 

Columns A and B for any county is a measure of the over- or 

under-concentration of the unemployed in a given county. In 

Column C we provide details of the ratio of the percentage of 

the unemployed to the percentage of the labour force 

contained in each county. A figure which is greater than 1 in 

Column C indicates an over-representation of the 

unemployed in a given county. A figure of less than 1 in 

Column C indicates an under-representation of the 

unemployed in a county. 

If one considers the figures for Donegal as shown in 

Appendix 1, Table A1.3, for example, one can see that it 

contained 5.6 per cent of the unemployed, as recorded in the 

2002 Census of Population. The county contained 3.2 per 

cent of the labour force as a whole. This implies an overconcentration 

of the unemployed of the order of 76 per cent 

in the county. In order words, the proportion of unemployed 

in the county is 1.76 times the level it would be if the 

unemployed were distributed on a pro rata basis with the 

total labour force. The counties and regions are ranked in 

Column D in terms of their level of over- or underconcentration 

of the unemployed, using the total labour force 

as a comparator. This shows that Donegal has the highest 

level of over-concentration (1.76 times the proportion if 

distributed on a pro rata basis). This is followed by Limerick 

20



Table 2.3: Distribution of the unemployed across regions 

% of the % 15+ in Ratio of 

Region unemployment labour force unemployment 

to labour force 

A B C D 

Ratio Rank 

Border 14.1 10.4 1.35 1 

Midlands 5.7 5.5 1.03 4 

West 9.8 9.3 1.05 3 

Dublin 29.7 30.8 0.96 5 

Mid-East 8.3 10.8 0.77 8 

Mid-West 7.9 8.5 0.93 6 

South-East 11.2 10.4 1.08 2 

South-West 13.3 14.3 0.93 6 

Total 100.0 100.0 


County Borough (1.57 times), Louth (1.50 times), Cork and 

Waterford County Boroughs (each approximately 1.37–1.40 

times) and so on. At the other extreme, Dún Laoghaire- 

Rathdown and Kildare are as low as 0.65 and 0.70, 

respectively, followed by Cork county, Meath and Limerick 

county (each in the region of 0.72–0.74). 

Given the importance of unemployment as a determinant of 

poverty and deprivation, it would be tempting to focus on it in 

isolation as an all-encompassing proxy for disadvantage. 

Accordingly, given the very low levels of unemployment in the 

Border counties mentioned above, one may feel that the risk 

of poverty and disadvantage in these areas may be 

correspondingly below the national average. 

21



In interpreting the figures, however, one must take account 

of the importance of farming in determining regional 

unemployment rates. Tables 2.4 and Appendix 1, 

Table A1.4 along with Maps 2.5 to 2.8 provide details 

on regional variations in the extent of farming in Ireland. 

Table 2.4: Percentage of all persons engaged in farming by 

number of acres farmed 

Region % of 15+ Farmers Farmers Farmers 

persons



Map 2.5: Per cent of persons in farming 

From Map 2.5 one can see that a high percentage of persons 

aged 15 years and over engaged in farming is found in the 

counties of Cavan, Roscommon and Leitrim – each with rates 

of the order of 10 to 12 per cent compared with the national 

total of 4.4 per cent. 

23



In Maps 2.6 to 2.8 we present details on the scale of farming 

in question. These clearly indicate that counties with the 

greatest extent of small-scale farming activity are 

substantially located in the Border and West regions. 

Appendix 1, Table A1.4 indicates that at a national level 12.3 

per cent of all farmers work less than 30 acres. Apart from 

Dublin, where less than 1 per cent of the population is 

Map 2.6: Per cent of farmers farming less than 30 acres 

24



engaged in farming, the regional rates of farming less than 30 

acres are highest in the Border and the West regions (18.4 

and 18.2 per cent, respectively). At a county level, excluding 

the county boroughs, we find that the figure for Mayo is 23.4 

per cent, 21.8 per cent in Donegal and 20.4 per cent in 

Monaghan. Similarly, from Map 2.7 one can see a high 

proportion of farmers in the range of 30–49 acres in these 

Map 2.7: Per cent of farmers farming 30–49 acres 

25



same counties. In contrast, Map 2.8 shows the much higher 

percentages of larger commercial farming enterprises in the 

South-East region. A total of 74 per cent of all farmers in the 

country were recorded as farming 50 or more acres. In 

Kilkenny the figure was 86.9 per cent, with high rates also 

found in Waterford County (86.8 per cent) and North 

Tipperary (86.3 per cent). 

Map 2.8: Per cent of farmers farming 50+ acres 

26



The pattern of agriculture in Ireland, therefore, can be 

characterised as being dominated by small-scale activity in 

the Western and Border counties with larger-scale 

enterprises being much more generally found in the South- 

East regions. Although no accurate figures are available, 

relatively higher levels of under-employment are often 

associated with small-scale farming. Accordingly, when 

considering regional trends in unemployment rates as 

shown in Map 2.4, one must take account of the role played 

by small-scale farming in determining measured 

unemployment rates. In general, therefore, although it would 

appear that unemployment rates are relatively low in many 

Border and Western counties, one must bear in mind that 

many of these same counties contain the highest levels of 

small-scale farming activity in the country. This latter may, 

at least to some degree, be masking recorded 

unemployment in the form of under-employment. 

In Table 2.5, Appendix 1, Table A1.5 and Map 2.9 we 

present details on a summary measure of economic 

dependency. This can be defined in various ways. For the 

purposes of the current report we have derived a 

dependency measure based on the ratio of those who are 

economically inactive to those who are economically active. 

The former includes the following categories: persons aged 

0–14 years, students, unemployed, first job seekers, those 

on home duties, retired, unable to work and ‘other’ labour 

force status categories. The economically active are simply 

those classified as being ‘at work’. The figures in Table 2.5 

and Map 2.9, therefore, provide details on the extent to 

which those who are economically active in each region or 

county are economically ‘supporting’ those who are 

inactive. 

27



Table 2.5: Index of economic dependency in each region 

Region 

Economic Dependency 

Index 

Rank 

Border 1.61 1 

Midlands 1.49 4 

West 1.51 2 

Dublin 1.21 8 

Mid-East 1.28 7 

Mid-West 1.43 6 

South-East 1.51 2 

South-West 1.46 5 

Total 1.39 

From Table 2.5 one can see that at a national level we had 

1.39 inactive persons for every person who was ‘at work’ in 

the 2002 Census. In general terms the figures suggest that 

levels of economic dependency are quite mixed across the 

country. 

Highest levels are found in the Border region (1.61), but these 

are largely driven by Donegal (1.84 per cent). This reflects the 

relatively elderly age structure of the county, combined with 

its very high level of recorded unemployment. The South-East 

and West also have high levels of economic dependency 

(1.51 per cent each). This is followed by the Midlands and the 

South-West having levels of economic dependency higher 

than the national average (1.49 and 1.46 per cent, 

respectively). This variability across the country and lack of a 

well-defined regional pattern reflects the fact that economic 

dependency is substantially affected by a number of factors, 

including regional age structure, unemployment rates and the 

related issues of agricultural activity, especially small-scale 

farming, as discussed above. 

28



Map 2.9: Economic dependency rates by local authority area 

In summary, therefore, in this section we saw that the lowest 

levels of labour force participation were in counties such as 

Leitrim, Cork County Borough, Roscommon, Mayo, Donegal 

and Longford. Unemployment rates were highest in parts of 

the Border region (largely driven by the very high rate 

recorded in Donegal) as well as the South-East. The level of 

29



unemployment in some Border and Western counties seemed 

to be lower than national levels. For example, Cavan and 

Roscommon both had recorded unemployment rates below 

8 per cent compared with a national figure of 8.8 per cent. 

However, we found that the Border and Western counties 

which were associated with low levels of recorded 

unemployment were equally characterised with high levels of 

small-scale farming. One can surmise that this latter may 

reflect elderly age structures and may, at least to some 

degree, mask recorded unemployment by under-employment. 

2.4 Variations in Levels of Educational Attainment 

Given the demonstrated relationship between poverty and 

disadvantage on the one hand and level of educational 

attainment on the other, it is clearly important to consider 

regional variations in the level of education as an aid to 

understanding the spatial patterning of disadvantage. The 

relevant figures are outlined in Table 2.6, Appendix 1, Table 

A1.6 and Maps 2.10 to 2.13. 

It is clear that the Border and West regions stand out as 

having an above average percentage of persons with lower 

levels of attainment. The national figure of 22.2 per cent of 

persons who are recorded as having left education with no 

qualifications or primary level only compares with a figure of 

29 per cent in the Border counties and 26 per cent for the 

Western region. One can see that 34 per cent of persons in 

Donegal, 31 per cent in Cavan, 30 per cent in Mayo and 29 

per cent each in Leitrim and Monaghan have left full-time 

education with, at most, primary-level education. Outside the 

counties in the Border and West regions only Longford, 

Offaly, Laois and Wexford (each with 26–30 per cent) have a 

30



Table 2.6: Distribution of persons 15 years and over whose 

education has ceased, classified by highest level of 

educational attainment 

Region None/ Lower Upper Third 

primary secondary secondary level 

% Rank % Rank % Rank % Rank 

Border 29.3 1 25.2 3 25.7 8 19.9 6 

Midlands 25.5 3 25.4 2 29.9 3 19.3 8 

West 25.9 2 21.2 7 29.1 6 23.7 4 

Dublin 18.7 7 19.2 8 28.7 7 33.5 1 

Mid-East 18.5 8 23.1 6 31.3 1 27.2 2 

Mid-West 22.3 5 23.8 4 30.6 2 23.3 5 

South-East 24.0 4 26.7 1 29.5 4 19.8 7 

South-West 21.3 6.0 23.8 4.0 29.4 5.0 25.5 3.0 

Total 22.2 22.7 29.1 26.0 


particularly high percentage of persons with at most 

primary-level education (see Map 2.10). In contrast, the 

counties with the highest percentages of third-level 

graduates are generally in Dublin and the Mid-East regions 

(see Map 2.13). 

In interpreting these figures it is important to remember that 

educational structures are to a large extent influenced by age 

structures of a region. The areas identified above as having 

lower than average levels of attainment are generally regions 

of the country which have an over-concentration of older 

persons. Overall, therefore, Maps 2.10 to 2.13 would seem to 

confirm a relatively higher level of educational disadvantage 

(much of which is related to age structures) in the counties of 

the Border and Western regions. 

31



Map 2.10: Per cent of persons age 15+ who have left school 

with no education or primary education 

In Table 2.7 we provide information on the levels of over- or 

under-concentration of those with the lowest and highest 

levels of educational attainment at a regional level. In Column 

A of the table we show the percentage of persons aged 15 

years and over in each county who have left full-time 

32




with lower secondary education 

education with no formal education or primary level only. 

Column B presents a comparable figure in respect of those 

who leave with third level. Column C provides details on the 

percentage of all persons who have left education in the 

region. This is the most appropriate comparator of the 

33




with upper secondary education 

distributions in Columns A and B. The ratio of the figures in 

Columns A and B to those in Column C provides a measure 

of the level of over- or under-concentration of persons into 

low and high levels of education. From Column D of the table 

34




with third-level education 

one can see that over-concentration of persons with lower 

levels of educational attainment is characteristic of Border, 

West and Midland regions to the extent of 33 per cent, 18 per 

cent and 15 per cent, respectively. In contrast, one can see 

35



that only in the Mid-East and Dublin regions do we have an 

over-representation of those with third-level qualifications by 

5 per cent and 27 per cent, respectively. 

Table 2.7: Distribution of persons whose education has ceased 

in highest and lowest levels of educational attainment across 

regions 

A B C D E 

Region % None/ % Third % Pop. Ratio Ratio 

primary level who left none/ third level 

education primary to to total 

total left left 

education education 

Ratio Rank Ratio Rank 

Border 14.6 8.4 11.0 1.33 1 0.77 6 

Midlands 6.5 4.2 5.7 1.15 3 0.74 8 

West 11.3 8.8 9.6 1.18 2 0.92 4 

Dublin 24.0 36.8 29.0 0.83 8 1.27 1 

Mid-East 8.7 10.9 10.3 0.84 7 1.05 2 

Mid-West 8.7 7.8 8.6 1.01 5 0.91 5 

South-East 11.8 8.3 10.9 1.09 4 0.77 6 

South-West 14.3 14.7 14.9 0.96 6 0.99 3 

Total 100.0 100.0 100.0 


At a county level one can see from Appendix 1, Table A1.7 

that counties such as Donegal, Cavan, Mayo and Monaghan 

have over-concentrations of persons with low levels of 

education to the extent of 32–53 per cent. At the other 

extreme, Dún Laoghaire-Rathdown is over-represented by 

persons with higher level of attainment by 76 per cent. Also in 

36



the Dublin region, Fingal is over-represented by 28 per cent. 

Galway County Borough also has a high over-representation 

of persons with higher levels of educational attainment by 55 

per cent. 

2.5 Variations in Levels of Social Class 

One of the more direct proxies of advantage and 

disadvantage available from census data is social class. The 

importance of class in determining poverty and disadvantage 

is well established. The census classifies all persons into five 

broad class categories, as follows: 

• higher professional workers/managerial/technical 

• non-manual 

• skilled manual 

• semi-skilled manual 

• unskilled manual. 

The classification is based on current or former occupation. 

Persons engaged in home duties, those classified as 

‘students’ and those who are otherwise economically inactive 

are assigned the class status of their family head. 1 

The percentage of persons in each social class is shown in 

Table 2.8, Appendix 1, Table A1.8 and Maps 2.14 to 2.18. If 

one concentrates initially on the percentage of persons in 

1 There is a residual ‘unknown’ category. This is used for persons who 

have no occupations in their own right and who live in households 

headed by someone who has never worked. 

37



Professional/Managerial and Technical classes, one can see 

that the regions with the highest rates are in Dublin (44.4 per 

cent) and the Mid-East region (41 per cent). Counties such as 

Dún Laoghaire-Rathdown have an extremely high proportion 

of their population classified as Higher/Lower Professional – 

60 per cent compared with a national total of 39 per cent. 

This is followed by Fingal and Galway County Borough (both 

47 per cent). At the other extreme, counties in the Border and 

West regions have a much lower percentage of population 

classified in the highest class category. In Donegal only 30 

per cent are classified as Professional/Managerial or 

Technical. The rates are also low in Monaghan and Cavan 

(31 per cent each) and Offaly and Mayo (32 per cent each) 

and in some of the main urban areas such as Limerick 

County Borough, with only 31 per cent of the population 

classified as being Professional/Managerial/Technical. 

Table 2.8: Percentage of persons in each social class 

Professional/ 

Managerial/ Non- Skilled Semi- 

Region Technical manual manual skilled Unskilled 

% Rank % Rank % Rank % Rank % Rank 

Border 32.2 8 19.7 6 23.9 1 15.8 1 8.4 2 

Midlands 33.9 7 20.2 2 23.1 2 14.5 3 8.3 3 

West 36.5 5 19.9 4 22.0 4 14.4 4 7.1 5 

Dublin 44.4 1 21.3 1 18.1 8 11.0 8 5.2 8 

Mid-East 41.0 2 19.5 7 21.1 6 11.8 7 6.6 7 

Mid-West 37.4 4 20.2 2 21.0 7 14.2 5 7.2 4 

South-East 34.5 6 19.0 8 23.1 2 14.7 2 8.8 1 

South-West 37.8 3 19.8 5 21.2 5 14.2 5 6.9 6 

Total 38.6 20.2 21.0 13.3 6.9 


38



Map 2.14: Per cent of persons classified as Professional/ 

Managerial/Technical 

In broad terms one can see from Map 2.14 that the overall 

pattern across the country clearly indicates that the areas 

with the lowest percentage of higher/lower professionals are 

the Border and Western counties while Dublin, the Mid-East 

and Cork County have the highest proportions in the country. 

39



Map 2.15: Per cent of persons classified as ‘Other Non-Manual’ 

In contrast, one can see from Map 2.18 that at the other end 

of the social class spectrum, Donegal (10.3 per cent) and 

several counties in the South-East (South Tipperary 9.7 per 

cent, Wexford 9.3 per cent and Carlow 9.2 per cent) have the 

highest percentages of persons in the Unskilled Manual 

category. 

40



Map 2.16: Per cent of persons classified as Skilled Manual 

In interpreting regional distributions of social class, the reader 

should note that farmers are assigned to social class 

categories according to acreage farmed. Small farmers (less 

than 30 acres) are assigned to the Semi-skilled Manual 

category. Large farmers (with 200 or more acres) are assigned 

to the Higher Professional category. No farmers are assigned 

to the unskilled manual category. 

41



Map 2.17: Per cent of persons classified as Semi-skilled 

Manual 

Table 2.9 provides details on the over- or under-concentration 

of persons from the top and bottom social class categories 

across the regions. The details in the table are comparable to 

those presented in Table 2.7 in respect of level of educational 

attainment. From Columns D and E in the table one can see 

42



Map 2.18: Per cent of persons classified as Unskilled Manual 

that the largest over-concentration of the Professional social 

class category is in the Dublin region (13 per cent), followed 

by the Mid-East (10 per cent). In contrast, we find the 

greatest over-concentration of the Unskilled Manual category 

in the South-East region (29 per cent). At the county level 

43



(see Appendix 1, Table A1.9), the highest relative 

concentration of this class group is found in Donegal, 

Wexford and South Tipperary (each in the range of 41–49 

per cent). 

Table 2.9: Regional distribution of Professional and Unskilled 

Manual class categories 

Region 

A B C D E 

Professional/ Un- % Ratio Ratio 

Managerial/ skilled Total profes- Unskilled 

Technical Manual popula- sional- Manual 

tion to total to total 

ratio 

Ratio Rank Ratio Rank 

Border 9.2 13.5 11.0 0.84 8 1.23 2 

Midlands 5.0 6.9 5.8 0.87 7 1.19 3 

West 9.0 9.9 9.7 0.93 5 1.02 5 

Dublin 32.4 21.2 28.7 1.13 1 0.74 8 

Mid-East 11.6 10.4 10.5 1.10 2 0.99 7 

Mid-West 8.4 9.1 8.7 0.97 4 1.04 4 

South-East 9.8 14.0 10.8 0.90 6 1.29 1 

South-West 14.6 15.0 14.8 0.99 3 1.02 5 

Total 100.0 100.0 100.0 

In general terms, therefore, it would appear that Dublin, the 

Mid-East and parts of the South-West region (specifically 

Cork) have the highest levels of their populations in the 

Professional/Managerial/Technical category. In terms of the 

Unskilled Manual group, although some counties in the 

South-East of the country have among the highest rates, 

there would generally appear to be a less clearly defined and 

more mixed pattern than is apparent with some of the other 

distributions considered. 

44



2.6 Conclusions 

In this chapter we considered spatial variations in sociodemographic 

variables related to poverty and deprivation 

from the 2002 Census of Population. The somewhat 

aggregated county level of the analysis undoubtedly masked 

a substantial degree of sub-county variation. Nonetheless, 

the chapter provides the reader with a clear picture of 

regional variations in the relevant indicators and in so doing 

adds considerably to the interpretative context for much of 

our discussion in later chapters of regional trends in direct 

measures of poverty and deprivation. 

Throughout the chapter we saw that the Border and West 

regions contained counties with the highest percentages of 

their population who were elderly, with the lowest levels of 

educational attainment, highest incidence of small farming 

activity and high levels of economic dependency. 

Throughout the chapter we emphasised that what we were 

examining were spatial patterns and outcomes. The 

underlying causal processes and drivers will be explored in 

later chapters, particularly in Chapter 7. These drivers may 

well be individual- or household-level characteristics such as 

age and housing tenure. 

45


Chapter 3 

Measurement of Poverty at the 

Household Level 


In this chapter we focus on the measurement of poverty and 

deprivation at the household level. In doing so we draw 

primarily on the series of Living in Ireland Surveys (LII) 

conducted in 1987, 1994 and 2000 but also on the National 

Survey of Housing Quality (NSHQ), which was conducted in 

2001/2002. All four surveys were designed to provide a 

nationally representative sample of the population resident in 

private households. The technical properties of these surveys 

have been described in detail elsewhere 1 and a summary of 

their major features is contained in Appendix 2. 

The major advantage of the LII surveys is that they contain 

the key indicators on which national reports monitoring 

poverty and deprivation are based and they allow us to track 

change over time using the same set of indicators. The major 

limitation of the LII surveys for our present purposes is that 

despite the fact that they constitute very large samples by 

1 Full descriptions are provided in Callan et al. (1989 and 1996) and 

Nolan et al. (2002), respectively, and Watson and Williams (2003). 

46


Measurement of Poverty at the Household Level 

social science standards, the extent to which we can 

disaggregate by spatial units is strictly limited. 

The NSHQ, which was conducted by the ESRI on behalf of 

the Department of the Environment and Local Government, 

has a substantially larger sample with 40,000 households. 

The NSHQ contains measures of income and deprivation for 

the participating households, which can be analysed at 

household level. However, as will become clear, these 

measures do not correspond precisely to those employed in 

the LII surveys. Fortunately, as we explain below, steps can 

be taken to ensure that while statements concerning absolute 

levels of deprivation and poverty are not possible using the 

NSHQ data, conclusions concerning relativities are robust. 

This allows us to take advantage of the much larger sample 

size in the NSHQ to achieve a much greater degree of 

geographical disaggregation when considering such 

relativities. 

3.2 Measuring Poverty and Deprivation in the LII 

Surveys 

We have extensively discussed elsewhere the many issues 

involved in the definition and measurement of poverty 

(most recently in Layte et al. 2001 and Nolan et al. 2002). 

Without duplicating that discussion, for the present study 

we take as our starting point that the concept of poverty is 

an explicitly relative one, relating to exclusion from the 

ordinary life of society due to lack of resources. We 

continue to stress the importance of acknowledging, where 

relevant, any uncertainty and absence of robustness in 

results. We therefore employ several different relative 

income poverty lines and a measure that takes into 

47



account both income and non-monetary indicators of 

deprivation. 

Relative income poverty lines are calculated as a particular 

percentage of average household income, taking differences 

in household size and composition into account. We take into 

account the fact that adults have greater needs than children 

and that there are economies of scale in consumption by 

using equivalence scales to adjust household income. Here 

we report results employing an equivalence scale of 0.66 for 

additional adults and 0.33 for children. The two relative 

income lines employed in this study are 50 and 60 per cent of 

mean equivalent income. For an adult living alone this 

produced poverty lines in 2000 of about £113/€144 and 

£136/€173 per week. The corresponding lines for 1994 were 

£65/€83 and £78/€99, respectively, and for 1987 they were 

£43/€55 and £51/€65, respectively. 

Reliance on income alone is open to the criticism that it may 

not be a good measure of low consumption and therefore 

deprivation. If poverty is defined as exclusion from the 

ordinary life of society due to lack of resources – understood 

as a state of generalised deprivation – it should be 

characterised by both a low standard of consumption and a 

low level of income. We therefore also make use of the 

consistent measure of poverty that has been employed in the 

National Poverty Strategy (Callan, Nolan and Whelan 1993; 

Nolan and Whelan 1996). This combines information on 

relative income with that relating to basic deprivation. This 

comprises a set of eight indicators (see Table 3.1) that relate 

to the enforced absence of such items as food, clothing and 

heat and going into debt to meet ordinary living expenses. 

Previous analysis has shown that these indicators reflect 

rather basic aspects of current material deprivation of items 

48



that are possessed by most people, are considered as 

necessities by most people and cluster together, providing 

support for the view that they serve as valid indicators of 

underlying generalised deprivation. The combined income 

and deprivation standard that we employ here requires that a 

household is counted as poor where it reports enforced lack 

of at least one of these basic items and also falls below 60 

per cent of mean income. This combined measure is one that 

we felt best captures our underlying conceptualisation of 

poverty as involving enforced absence of socially defined 

necessities. In the current context it is important to stress that 

unlike items such as cars, telephones and holidays, there is 

no evidence for significant urban-rural difference in the extent 

to which the items included in the basic deprivation index are 

considered to be necessities. 

3.3 Measuring Poverty and Deprivation in the NSHQ 

The discussion above outlined the issues involved in 

measuring poverty and deprivation using the LII surveys, 

which had been specifically designed for this purpose. In the 

NSHQ we are more limited in terms of what is possible. 

Income in the National Survey of Housing Quality is measured 

by a single item, which asks for the approximate level of net 

household income, and records the answer into one of 16 

categories. The wording is as follows: 

Finally, a few questions about how you are able to 

manage financially. Could I ask about the 

approximate level of net household income This 

means the total income, after tax and PRSI, of ALL 

MEMBERS of the household. It includes ALL TYPES 

49



of income: income from employment, social welfare 

payments, Child Benefit, rents, interest, pensions etc. 

We would just like to know into which broad group 

the total income of your household falls. I’d like to 

assure you once again that all information you give 

me is entirely confidential. 

Respondents were first of all presented with a card showing 

four broad income categories. Then they were presented with 

a second card that broke down each of these four broad 

categories into four more detailed categories. The result was 

a 16-category variable for total household income. This item 

had a reasonably good response rate, with 87.7 per cent of 

respondents providing information on the initial four-category 

breakdown and 85.3 per cent providing information on the 

more detailed 16-category breakdown. Income category was 

imputed for the 12.3 per cent of households for whom the 

information was missing using information on household size, 

number of persons at work, social class, local authority area 

and sample cluster. Further details on the imputation of 

missing information in the NSHQ are given in Appendix 2. 

The NSHQ single-item measure of income will tend to 

understate total household income, particularly in larger 

households, in comparison with the Living in Ireland Survey. 

The understatement arises for a number of reasons: 

incomplete information on the part of the householder 

regarding earnings and income of other people in the 

household and a tendency to forget some components (such 

as Child Benefit and irregular payments) when responding to 

a single question. 

Data from the 2000 Living in Ireland Survey was used to 

develop a correction for the NSHQ single-item income 

50



measure. Details of how this was done are provided in 

Appendix 2. Nevertheless, because of the differences in the 

measurement of income compared to the LII, it is not 

possible to reproduce a measure with the same overall level. 2 

What we are relying on in this analysis is that the measure of 

poverty risk derived from the NSHQ will faithfully reproduce 

the same patterns of variation across sub-groups of 

households as the LII measure. To the extent that it was 

possible to compare the distribution of risk across regions, 

tenure, household size categories, economic status and 

social class, the NSHQ measure compares very favourably to 

the LII measure 3 of income. 

The non-monetary indicators of deprivation available on the 

NSHQ are also different from those available in the Living in 

Ireland Survey. Some of the items likely to capture an 

absence of very basic necessities (such as shoes, coat and a 

hot meal) are not available. There are also some important 

differences in how ‘basic’ deprivation items are measured. In 

the NSHQ, we ask whether the household can afford each 

item if it wanted it. In the LII, a two-step approach is taken to 

most deprivation items. The household is first asked if it 

possesses an item. If the answer is no, the household is 

asked if this is something it would like to have but cannot 

afford. ‘Enforced lack’ in the LII context, then, refers to a 

household (a) lacking an item (b) unable to afford it and (c) 

would like it. In the NSHQ, the last condition is missing, so 

some households who would not want an item would be 

counted as experiencing an enforced lack in the sense that 

2 In fact, the measure of average income in the NSHQ tends to be higher 

than in the LII and the ‘poverty’ levels about 5–6 percentage points 

lower. 

3 Detailed comparisons are provided in Appendix 2. 

51



they could not afford the item, even if they did want it. For 

this reason, the levels on the NSHQ for enforced lack of 

comparable items tend to be higher. 

Table 3.1 shows the items used in each deprivation 

dimension and the percentage of households that lack them 

on the NSHQ. Note that the items are measured differently to 

those in the Living in Ireland Survey, as noted earlier, so that 

the levels are not directly comparable. A factor analysis 4 was 

undertaken to check if the same type of underlying factor 

structure emerged in the NSHQ data as in the LII. The 

factor loadings set out in Table 3.1 show which items figure 

prominently on each of the dimensions. The structure 

represented by the set of dimensions identified is very similar 

to that identified using the LII data, although there are some 

differences in terms of the factors on which specific items 

load most highly. The five-factor solution resulted in a basic 

dimension, a secondary dimension, two housing dimensions 

(housing amenities and housing deterioration) and an 

environment dimension. 

The basic dimension reflects an inability to afford items such 

as a meal with meat every second day, new clothes, 

adequate heating, an ability to replace worn furniture, to 

socialise with family or friends once a month and to go on a 

week’s annual holiday. It also includes an item on whether the 

household experienced arrears in the past 12 months on 

housing or utility bills. Note that this measure does not 

contain the same set of items as were used in the analysis of 

the LII data – the same set of items is not available on the 

NSHQ. However, this set of items emerged in analyses of the 

4 The method used was principal component extraction and oblique 

rotation. 

52



internationally comparable European Community Household 

Panel Survey (ECHP) as forming a coherent dimension of 

deprivation across European countries (Whelan et al. 2001). 

The secondary dimension includes an item on inability to 

afford a car, as well as items on whether the household would 

like but cannot afford a range of household appliances 

(microwave, dishwasher, colour TV, video recorder and 

telephone). The levels on these household appliance items 

are lower than on some of the basic items because of the use 

of the two-step method in measuring them. The household 

was first asked if it had the item. If not, it was asked whether 

the item was something it would like to have but could not 

afford. Unlike the basic items, then, households who would 

not like the item are not counted as deprived. 

The housing amenities dimension captures an absence of 

basic housing amenities: a flush toilet, bath or shower and 

hot running water. 

The housing deterioration index captures the presence of 

problems in the accommodation that are regarded as 

moderate or major problems by the householder. The 

problems are leaks (in roof/doors/windows), dampness (rising 

damp, condensation dampness or dampness of unknown 

origin), rot (in doors, windows or other timbers) and noise 

from neighbouring houses. The index also includes an item 

on the adequacy of the space available. It also measures the 

householder’s judgement as to whether the accommodation 

is too small for the household’s needs. 

The final dimension, labelled environment, refers more to the 

social environment than to the physical environment. It 

includes problems of public order such as vandalism, graffiti, 

rubbish or litter lying about, homes or gardens in bad 

53



Table 3.1: Items used in the deprivation scales in the NSHQ 

Factor loading 

(principal component 

extraction; oblique rotation) 

Basic 

Secondary 

Housing 

amenities 

Housing 

deterioration 

Environment 

% 

Deprived 

Cannot afford: 

Meal with meat, etc. 0.55 4.5 

New clothes 0.66 8.0 

Adequate heating 0.60 7.9 

Socialising once a month 0.74 24.2 

Replacing worn furniture 0.76 27.3 

Week’s holiday per year 0.77 31.2 

Arrears on housing/utility bills 0.26 8.9 

Cannot afford car or van 0.24 22.0 

Would like but cannot afford: 

Microwave 0.66 4.1 

Dishwasher 0.51 14.7 

Colour TV 0.51 0.4 

Video recorder 0.70 2.9 

Telephone, incl. mobile 0.56 2.3 

Accommodation lacks: 

Flush toilet 0.65 1.6 

Bath/shower 0.83 1.4 

Hot running water 0.79 1.6 

Moderate or major problem with: 

Leaking roof/doors/windows 0.77 5.8 

Dampness 0.78 6.5 

Rot in doors/windows/floors 0.69 3.4 

Noise from neighbours 0.21 2.8 

Accommodation too small for needs 0.30 13.5 

Very common in area: 

Vandalism 0.79 2.8 

Graffiti 0.76 2.6 

Rubbish/litter 0.73 5.0 

Homes/gardens in bad condition 0.74 1.8 

Public drunkenness 0.70 3.1 

Reliability (alpha) 0.76 0.46 0.66 0.50 0.79 – 

Source: Irish National Survey of Housing Quality 2001/2002 

54



condition or people being drunk in public. The items capture 

the householder’s judgement that these problems are very 

common in the area where the accommodation is located. 

The reliabilities of the scales, i.e. the extent to which the items 

are tapping a common underlying dimension, are shown in the 

last row of Table 3.1. The values of the indices of reliability are 

very respectable for the basic and environment dimensions 

and satisfactory for the housing amenity dimension. The 

reliability of the secondary dimension and of the housing 

deterioration dimension is lower than we would like. 

In constructing indices, the items were weighted by the 

proportion that does not lack the item. For instance, 96 per 

cent of households can afford a meal with meat, chicken or fish 

every second day while 69 per cent can afford a week’s annual 

holiday away from home. Enforced lack of a meal with meat, 

etc. is given a weight of 0.96, while enforced lack of a holiday 

is given a lower weight of 0.69. Such weighting by the 

proportion not lacking an item captures the intuition that 

enforced lack of something possessed by most people is more 

‘serious’ than enforced lack of something that fewer people 

possess. 

3.4 The National Context 

Before proceeding to a spatial analysis of poverty and 

deprivation, we wish to put our later results in context by 

presenting results from earlier Living in Ireland Surveys. This 

is done in Table 3.2. In 2000 over one in four respondents 

were below the 50 per cent line, in comparison with just less 

than one in five in 1994 and one in six in 1987. 5 The picture at 

5 The most up-to-date national poverty rate trends can be found in 

Whelan et al. (2003). 

55



the 60 per cent line is somewhat different. By 2000 the 

poverty rate was lower that that observed in 1994 but higher 

than that for 1987. The actual number of households below 

the 50 and 60 per cent lines in 2000 was 26 per cent and 33 

per cent, respectively. 

Significant changes in the composition of households below 

the relative income lines took place during the period under 

study. Among the most striking findings was the increasing 

risk for those in single-person households, in households 

where the reference person is ill/disabled or retired and for 

those who are themselves aged 65 or over. Those in 

households where the reference person is unemployed still 

faced a relatively high risk of falling below the income 

thresholds but continued to decline as a proportion of all 

those below the lines. Conversely, those in households where 

the reference person is an employee still faced the lowest risk 

by far but are becoming more important among those below 

the thresholds (as the numbers of unemployed continued to 

fall and the number of employees continued to rise from 1998 

to 2000). Those aged 65 or over faced a much higher risk of 

falling below 60 or 70 per cent of median income than those 

aged 18–65, with children then facing an intermediate level of 

risk. Women faced a higher risk of falling below those lines 

than men, but this gap was most marked among the elderly. 

As well as household income, non-monetary indicators of 

deprivation have also been employed in attempting to identify 

those in each of the LII surveys who were experiencing 

exclusion from ordinary living patterns due to lack of 

resources. Analysis of available indicators led us to focus on 

the eight indicators included in the basic deprivation index. 

The combined income and deprivation consistent poverty 

standard employed here is that a household is counted as 

56



poor where it reports enforced lack of at least one of these 

basic items and falls below 60 per cent of mean income. 

About 16 per cent of households were in that position in 

1987. By 1994 the percentage had fallen marginally to 15 per 

cent. In 2000 we observed a substantial reduction to 6 per 

cent. 

Table 3.2: Trends in percentage of households below poverty 

lines 

(Equivalence Scale A – 1.0/0.66/0.33) 

1987 1994 2000 

Percentage of households 

below line: 

50% line 16.3 18.6 25.8 

60% line 28.5 34.2 32.9 

60% consistent line 16.0 15.1 6.2 

Source: Living in Ireland Surveys, 1987, 1994 and 2000 


In this chapter we have dealt with the manner in which 

income and deprivation are measured at the household level 

in the major data sources on which we have based our 

analysis. As we have pointed out, the manner of measuring 

both variables is sufficiently different in the NSHQ survey that 

we feel it would be confusing rather than helpful to report 

poverty rates from that survey. However, our exploratory 

analysis indicates that very similar conclusions emerge from 

both sources when we focus not on levels but on differentials 

or disparities and that is the strategy we will pursue when 

57



utilising this data in our subsequent analysis. This will allow 

us to take advantage of the fact that the much greater 

sample size of the NSHQ allows us to make much finer 

distinctions in terms of units of analysis. 

58


Chapter 4 

Poverty and Deprivation by Region and 

Local Authority Area 


In this chapter we present results on poverty and deprivation 

by regional authority and by local authority area. Two different 

sources of data are used for this purpose: the Living in 

Ireland Survey (LII) for 2000 and the National Survey of 

Housing Quality (NSHQ) for 2001/2002. The LII Surveys were 

designed to provide nationally representative samples of the 

population resident in private households. While the numbers 

of households in the achieved samples 1 are relatively large, 

both sample size and the sampling designs limit the extent to 

which they can be disaggregated by geographical area. Thus, 

while it would be possible to derive poverty rates by county 

from the samples, it would not be sensible to do so because 

of the wide margin of error associated with the figures. 

Analysis at the level of planning region or regional authority, 

each comprising a number of counties, is possible. In this 

chapter we focus on the latter. 2 59 

1 The completed sample was 3,467 in 2000. 

2 However, in order to allow trends over time to be considered, we 

report figures for planning region by year in Appendix 1.



On the other hand, at over 40,000 households, the NSHQ 

sample is large enough to provide reliable estimates at the 

level of local authority. The NSHQ data are used for this 

purpose in this chapter. In using the NSHQ data, we focus on 

disparities – or variations across local authorities – in poverty 

and deprivation. As noted in the previous chapter, it is not 

possible to use the NSHQ to replicate the levels of income 

and deprivation used in the LII Survey, but the patterns of 

variation across regions are very similar. 

In the final section of the chapter, the regional distribution of 

poverty and deprivation is compared to the distribution of 

characteristics of individuals and households from the 2002 

Census, as presented in Chapter 2. 

4.2 Poverty Risk and Incidence by Regional 

Authority in the LII Survey 

In Table 4.1 we show the distribution of risk of poverty across 

regional authority for all three poverty lines. The risk of 

poverty, or poverty rate, is the percentage of households with 

equivalised income below the relevant poverty line. 3 The 

regional authorities are made up as follows. 

• Border: Cavan, Donegal, Leitrim, Louth, Monaghan, Sligo. 

• Dublin: Dublin City, Fingal, Dún Laoghaire-Rathdown, 

Dublin South. 

• Mid-East: Kildare, Meath, Wicklow. 

3 Equivalised income, or income per adult equivalent, is calculated by 

allowing a ‘weight’ of 1 for the first adult in the household, 0.66 for 

each subsequent adult and 0.33 for each child under age 15. The 

poverty line is expressed in terms of a percentage, e.g. 50 or 60 per 

cent, of the average equivalised income. 

60


Poverty and Deprivation by Region and Local Authority Area 

• Midlands: Laois, Longford, Offaly, Westmeath. 

• Mid-West: Clare, Limerick, Tipperary North Riding. 

• South-East: Carlow, Kilkenny, Tipperary South Riding, 

Waterford, Wexford. 

• South-West: Cork, Kerry. 

• West: Galway, Mayo, Roscommon. 

At the 50 per cent line the highest poverty rate of 36 per cent 

was observed for the Border region. However, it was closely 

followed by the South-West and West with rates of 32 per 

cent. Dublin has the lowest rate by far at 17 per cent. The 

remaining regions are found in the range running from 22 to 

30 per cent, with the Mid-East at the lower end of the range 

and the Midlands and South-East at the upper end. The 

highest rate at the 60 per cent line was for the Border region 

and the West, with rates of 46 per cent and 44 per cent, 

respectively. Once again, the lowest rate of 21 per cent was 

found for Dublin, followed by the Mid-East with a rate of 26 

per cent. The poverty risk for the remaining regions was 

about 35 to 40 per cent. 

The highest values for the consistent poverty line, as shown 

in Table 4.1 and Map 4.1, are found in the Border and Mid- 

West regions with rates of about 10 per cent. The lowest rate 

of 3 per cent is observed in Dublin. The values for the other 

regions range between 4 per cent and 8 per cent. 

The bottom panel of Table 4.1 aggregates the regional 

authorities into the two regional assembly areas: the Border, 

Midlands and West (BMW) and the South and East (Dublin, 

Mid-East, Mid-West, South-East and South-West). The 

greater risk of poverty in the BMW region compared to the 

South and East is very evident on all three measures. 

61



Table 4.1: Poverty risk in 2000 by regional authority and 

regional assembly 

Region 50% 60% 60% line+ 

income income deprivation 

line risk line risk risk 

Border 35.5 46.4 10.1 

Dublin 17.1 21.3 3.3 

Mid-East 22.2 25.9 5.8 

Midlands 29.7 34.5 4.2 

Mid-West 25.9 35.2 10.4 

South-East 26.9 35.4 7.7 

South-West 32.4 40.0 7.5 

West 31.6 43.5 5.0 

Regional Assembly 

BMW 32.9 42.7 7.1 

South & East 23.0 29.0 5.8 

Total 25.8 32.9 6.2 

Source: Living in Ireland Survey, 2000 

Table 4.2 sets out the incidence figures broken down by 

regional authority and regional assembly. The incidence of 

poverty refers to the percentage of all poor households that 

are located in each region. At all three lines approximately 

one in five to one in six of the poor are located in each of the 

Border regions, Dublin and the South-West; one in seven to 

one in eight are in the South-East. The figure for the 

consistent line is somewhat lower for Dublin and higher for 

the South-East. For the other regions the modal figure is 7 

per cent. However, the concentration of consistent poverty is 

rather higher in the Mid-West, at 13 per cent, and low in the 

Midlands, at 4 per cent. 

The figures in the bottom panel of Table 4.2 indicate that 

close to two-thirds of the poor are located in the South and 

62



Map 4.1: Consistent poverty risk by region 

63



East, in spite of the higher risk of poverty (as seen in Table 

4.1) in the BMW region. This reflects the higher population of 

the South and East, as can be seen in the final column of the 

table: 72 per cent of households are located in the South and 

East, compared to 28 per cent in the BMW region. 

Table 4.2: Poverty incidence in 2000: Per cent of poor in each 

regional authority and regional assembly area (relative income 

poverty and consistent poverty) 

Region 50% relative 60% relative Consistent Per cent 

income income poverty of all 

poverty poverty 60% line+ households 

line line deprivation 

Border 17.4 17.8 20.5 12.6 

Dublin 20.7 20.3 16.7 31.3 

Mid-East 6.9 6.3 7.6 8.0 

Midlands 7.6 7.0 4.4 6.6 

Mid-West 7.7 8.2 13.1 7.7 

South-East 11.7 12.1 14.1 11.2 

South-West 16.8 16.3 16.3 13.4 

West 11.2 12.1 7.4 9.2 

Regional Assembly 

BMW 36.2 36.9 32.3 28.4 

South & East 63.8 63.1 67.7 71.6 

Total 100.0 100.0 100.0 100.0 


4.3 Disparities in Income Poverty Risk by Local 

Authority Area from the NSHQ 

At this point we turn to data from the NSHQ to examine 

regional variation in poverty risk at a more detailed level. 

64



Table 4.3: Disparities in income poverty risk by local authority 

area 

Region County Income poverty Income poverty 

risk (50%) risk (60%) 

Border Cavan 1.3 1.3 

Donegal 1.7 1.6 

Leitrim 1.3 1.4 

Louth 1.1 1.1 

Monaghan 1.1 1.0 

Sligo 1.2 1.1 

Midlands Laois 1.0 1.0 

Longford 1.5 1.4 

Offaly 1.3 1.2 

Westmeath 1.0 1.1 

West Galway City 0.7 0.7 

Galway County 1.2 1.2 

Mayo 1.5 1.4 

Roscommon 1.2 1.1 

Dublin Dublin City Council 1.0 1.0 

Dublin Fingal 0.4 0.5 

Dublin South 0.5 0.6 

Dún Laoghaire-Rathdown 0.4 0.5 

Mid-East Kildare 0.7 0.7 

Meath 0.7 0.8 

Wicklow 0.9 1.0 

Mid-West Clare 1.0 1.1 

Limerick City 1.5 1.3 

Limerick County 1.1 1.1 

Tipperary North Riding 0.9 0.9 

South-East Carlow 1.3 1.2 

Kilkenny 0.9 0.9 

Tipperary South Riding 1.3 1.2 

Waterford City 1.2 1.1 

Waterford County 1.1 1.1 

Wexford 1.2 1.2 

South-West Cork City 1.4 1.3 

Cork County 1.0 0.9 

Kerry 1.3 1.3 

Total 1.0 1.0 

Source: National Survey of Housing Quality, 2001/2002 

65



Table 4.3 and Maps 4.2 and 4.3 show the disparities in 

poverty risk by local authority area using data from the 

NSHQ. The figures in the table show the ratio of the poverty 

risk in each local authority area to the national figure. Thus, 

for instance, the risk of poverty in Cavan – using the 50 per 

cent mean income line – is 30 per cent higher than the 

national average. On the other hand, it is 60 per cent lower 

than the national average in Dublin Fingal. 

It is clear from Table 4.3 that there are considerable 

differences within regions. For example, in the Border region, 

all of the counties (except Monaghan using the 60 per cent 

line) tend to have poverty risk levels above average. However, 

the risk of poverty is 60–70 per cent greater than the national 

average in Donegal, 30–40 per cent above that national figure 

in Cavan and Leitrim and drops to 10–20 per cent above the 

average in Louth, Monaghan and Sligo. There is a somewhat 

smaller level of variation among the Midland counties, with 

Longford and Offaly faring 20–50 per cent worse than 

average in terms of poverty risk, while Laois and Westmeath 

are close to average. 

The differences are even more dramatic in the West: Galway 

City has a poverty risk level that is well below average (30 per 

cent), while the risk in Mayo is 40–50 per cent above average. 

Galway County and Roscommon occupy an intermediate 

position, with risk levels 10–20 per cent above average. The 

risk of poverty is just at the national average in Dublin City, 

but is lowest of all the local authority areas nationally in the 

surrounding Dublin counties at 40–60 per cent below the 

average. 

In the Mid-East, Kildare and Meath have risk levels that tend 

to be 20–30 per cent below the national level, but the risk 

level in Wicklow is closer to the national average. 

66



Map 4.2: Disparities in income poverty risk by local authority 

area (50 per cent poverty line) 

Among the counties in the Mid-West, the risk is highest in 

Limerick City (30–50 per cent above average). In fact, the 

poverty risk level at the 50 per cent line in Limerick City is the 

highest among the five county boroughs (Dublin, Cork, 

Limerick, Galway and Waterford cities). The risk level is 

somewhat below average in Tipperary North (10 per cent 

67



Map 4.3: Disparities in income poverty risk by local authority 

area (60 per cent poverty line) 

below average), while Clare and Limerick County show risk 

levels that are slightly above average. 

In the South-Eastern region, only Kilkenny fares better than 

the nation as a whole (10 per cent below the average) in 

terms of poverty risk, while Tipperary South, Wexford and 

Carlow (20–30 per cent above) fare worst. 

68



In the South-West, the level of poverty risk is 30–40 per 

cent above the national level in Cork City and in Kerry, but 

is at or slightly below the overall average in Cork County 

(0–10 per cent below). 

4.4 Modified Consistent Poverty (MCP) 

As noted earlier, the NSHQ does not contain the set of 

basic deprivation items used in the construction of the 

measure of consistent poverty based on the LII data, so 

even if we could ignore the differences in how income is 

measured, it would not be possible to reproduce the 

measure of consistent poverty that is used in monitoring 

the National Anti-Poverty Strategy. Nevertheless, we can 

construct a ‘modified’ consistent poverty index which takes 

account of both income and living standard and examines 

the way in which it varies by region. 

Modified consistent poverty (MCP) is defined as 

households below 60 per cent of the mean equivalised 

household income (using the NSHQ income measure) and 

experiencing basic deprivation, i.e. lacking one or more of 

the items in the basic deprivation index. The basic 

deprivation items in the NSHQ include an inability to afford 

adequate food, clothing and heating for the home, as well 

as socialising, an annual holiday and the presence of 

arrears on utility bills. 

As before, we do not present levels of MCP in the following 

tables to avoid confusion with the full measure of 

consistent poverty based on the LII data. Instead, the 

results are presented in terms of ratios of the national 

average figure. 

69



4.5 MCP by Local Authority Area 

Table 4.4 and Map 4.4 show the disparities in MCP by local 

authority area. Earlier tables revealed considerable variation in 

income poverty by local authority area within regions. As can 

be seen in Table 4.4, the differences persist when the 

combined measure of disadvantage is used, taking account 

of both income and basic deprivation. Overall, the highest 

level of MCP is found in Donegal (1.9 times the average) and 

the lowest level in Dún-Laoghaire-Rathdown (0.60 times the 

average). Nevertheless, within the Border region the range of 

MCP is from Monaghan’s low rate at 10 per cent below the 

average to Donegal’s high figure of 90 per cent above. The 

figures in the Midlands are all above average, but range from 

a modest 10 per cent above average in Laois, Offaly and 

Westmeath to 50 per cent above average in Longford. 

Galway City stands apart in the Western region with an MCP 

30 per cent below the average, while in the other counties in 

this region the figure ranges from 1.0 to 1.5. 

Dublin City has an MCP which is on a par with the national 

average, but it is well below the national figure in the 

surrounding counties of Fingal, Dublin South and Dún 

Laoghaire-Rathdown. The Mid-East evinces a similar, though 

less marked, pattern of variation: the level in Wicklow is the 

same as the national level but it is 20–30 per cent lower in 

Kildare and Meath. The MCP in the Mid-West ranges from a 

figure equal to the national average in Limerick County to 1.5 

in Limerick City. 

In the South-East, Kilkenny maintains its relatively favourable 

position with an MCP 20 per cent below the national figure, 

while high figures (30 per cent above the average) are found 

in Waterford City and Carlow. Cork City and Kerry have MCP 

70



Table 4.4: Disparities in risk of modified consistent poverty by 

local authority area 

Region County Modified consistent poverty 

Border Cavan 1.2 

Donegal 1.9 

Leitrim 1.5 

Louth 1.2 

Monaghan 0.9 

Sligo 1.3 

Midlands Laois 1.1 

Longford 1.5 

Offaly 1.1 

Westmeath 1.1 

West Galway City 0.7 

Galway County 1.0 

Mayo 1.5 

Roscommon 1.1 

Dublin Dublin City Council 1.0 

Dublin Fingal 0.5 

Dublin South 0.6 

Dún Laoghaire-Rathdown 0.4 

Mid-East Kildare 0.7 

Meath 0.8 

Wicklow 1.0 

South-West Clare 1.2 

Limerick City 1.5 

Limerick County 1.0 

Tipperary North Riding 1.1 

South-East Carlow 1.3 

Kilkenny 0.8 

Tipperary South Riding 1.2 

Waterford City 1.3 

Waterford County 1.0 

Wexford 1.0 

South-West Cork City 1.3 

Cork County 0.9 

Kerry 1.3 

Total 1.0 

Source: Irish National Survey of Housing Quality, 2001/2002 

71



levels 30 per cent above the national figure, while the level in 

Cork County is about 10 per cent lower. 

Map 4.4: Disparities in risk of modified consistent poverty by 

local authority area 

72



One pattern that is becoming clearer at this stage is a 

tendency for levels of deprivation to be higher in the large 

cities themselves (Dublin, Cork, Limerick, Waterford) than in 

the surrounding counties. This pattern fits with a model where 

better-off households move to newer and more spacious 

accommodation in the suburbs, which are often located 

outside the city boundaries or even in neighbouring counties. 

This leaves poorer households trapped in often older and less 

attractive accommodation in the cities. Galway is an 

important exception to this pattern: the city itself is in a more 

favourable situation than the county with respect to 

deprivation indicators. 

4.6 Direct and Indirect Measures of Poverty and 

Deprivation 

Having examined the direct measures of poverty and 

deprivation from the LII Survey and the NSHQ, we now turn 

to the relationship between these direct measures at an 

aggregate level and the indirect measures based on census 

data presented in Chapter 2. 

We saw earlier in this chapter that counties in the Border and 

Western regions emerge as having the highest risk in terms of 

relative income poverty. As previously shown in Chapter 2, 

these were areas with (a) the oldest and most age-dependent 

population structures (b) the lowest levels of educational 

achievement (c) the lowest levels of professional classes and 

(d) the highest levels of small farm activities. In contrast, at 

the other extreme, most of the Dublin and Mid-East region 

along with commuter counties such as Meath, Kildare and 

counties such as Kilkenny, North Tipperary and Cork County 

are among the lowest in terms of relative income poverty risk. 

73



These were also identified in Chapter 2 as being among the 

counties with (a) the youngest age profiles (b) the highest 

levels of large-scale farming (c) the lowest unemployment 

rates and (d) the highest social class and educational 

attainment structures in the country. 

A useful tool in examining the relationship between direct and 

indirect measures of poverty is correlation analysis. A 

correlation coefficient ranges from –1.0 to +1.0 depending on 

the strength of a relationship between two variables. If the 

two variables in question move together in perfect unison, i.e. 

as one increases/decreases the other increases/decreases at 

exactly the same rate, the correlation coefficient will be +1.0. 

If one increases as the other decreases (and vice versa), the 

coefficient will be negative. 

The figures in Table 4.5 present correlation coefficients of the 

socio-demographic variables used throughout Chapter 2 with 

three key measures of poverty discussed in the present 

chapter. These figures provide a quantifiable measure of the 

strength of the relationship at the county level of trends in the 

poverty measures with the underlying characteristics. 

From the table one can see that there is a strong relationship 

between county-level poverty risk (using the 50 per cent, 60 

per cent and modified consistent lines) and several of the 

socio-demographic variables considered. For example, one 

can see that the percentage of the county population 

classified as being 65 years and over, having no formal 

education/primary certificate only and also being in the 

professional class categories are strongly correlated (some 

negatively) to risk of poverty. 

The reader should note that the correlation coefficients 

presented here refer to characteristics measured at the level 

74



Table 4.5: Correlation of socio-demographic characteristics of 

county with risk of poverty 

Poverty Poverty MCP 

50% 60% line 

Per cent persons age 65+ 

and living alone 0.693 0.718 0.653 

Age dependency index 0.576 0.628 0.536 

Unemployment rate 0.650 0.607 0.686 

Per cent all persons at work who 

are small farmers (




This chapter drew on data from the 2000 LII Survey and the 

2001/2002 NSHQ to examine patterns of variation in income 

poverty and deprivation by area. At the level of regional 

authority, the Border area emerges as the most clearly 

disadvantaged area. The West and South-West tend to be 

next in line. At the other extreme is Dublin, and to a slightly 

lesser extent the Mid-East, with low levels of poverty and 

deprivation. 

When we differentiate within regions by identifying local 

authority areas, Donegal, Cavan, Leitrim, Longford, Mayo and 

the major urban centres outside Dublin emerge as 

experiencing the highest levels of disadvantage. 

One finding to emerge rather clearly is that there are big 

differences between the counties comprising a regional 

authority area. In other words, spatial contiguity does not 

correspond to a common fate in terms of income poverty and 

deprivation. The boundaries between regions are drawn on 

the basis of political and pragmatic considerations and do not 

necessarily reflect a confluence of similar causal processes. 

This has implications for area-based strategies to combat 

poverty in that any remedies applied at the broad level of 

regional authority will target areas that are quite different from 

each other in terms of the level and intensity of deprivation. 

In this chapter we complemented the direct measures of 

poverty risk and deprivation with a consideration of spatial 

variations in relevant socio-demographic variables from the 

Census of Population. The somewhat aggregated county 

level of the analysis undoubtedly masked a substantial 

degree of sub-county variation. Nonetheless, the correlation 

coefficients indicated that at the county level there is a 

76



reasonably strong relationship between the direct measures 

of poverty on the one hand and indirect measures such as 

age dependency, unemployment rate, rate of educational 

disadvantage and membership of the unskilled manual social 

class on the other. 

In the next chapter we turn to two other aspects of spatial 

variation: size of place (or urban-rural location) and tenure to 

ask to what extent these are associated with significant 

differences in income poverty and deprivation. 

77


Chapter 5 

Poverty by Area and Tenure Type 


In this chapter we draw on data from a number of sources to 

examine variations in poverty by area type, tenure and a 

combination of these two variables. As it is used here, ‘area 

type’ refers to the population density of the area where the 

household is located. 

The National Survey of Housing Quality (NSHQ) results 

consistently pointed to the importance of housing tenure as a 

marker of disadvantage. Local authority renters in particular 

tended to live in accommodation that was in poorer 

condition, to have lower incomes and to report experiencing 

more financial strain than those in other tenures (Watson and 

Williams 2003). The disadvantage experienced by local 

authority renters is largely a selection effect: households have 

to be low-income in order to qualify for local authority 

housing in the first place and if their income position 

improves they have traditionally bought out their rented 

dwellings, becoming homeowners. 

The data analysed in this chapter comes from the Living in 

Ireland (LII) Surveys of 1987, 1994 and 2000 and the National 

Survey of Housing Quality 2001/2002 (NSHQ). The inclusion 

78



of the 1987 and 1994 LII Surveys permits us to examine 

trends over time. The larger sample size of the NSHQ allows 

us to examine the patterns in greater detail. We begin by 

describing how area type and tenure are measured. Next we 

examine the risk and incidence of poverty by area type and 

tenure using data from the LII surveys. Finally, we examine 

disparities in income poverty and in modified consistent 

poverty (MCP) by area type and tenure using data from the 

NSHQ. As in the previous chapter, since the measures of 

income poverty and consistent poverty derived from the LII 

cannot be replicated using the more limited variables 

available on the NSHQ, we focus on disparities in poverty 

rather than levels of poverty when analysing data from the 

NSHQ. 

5.2 Measuring Area Type and Tenure 

We distinguish the following five categories for area type: 

• Open country 

• Village or town with population of less than 3,000 

• Town with population of 3,000 or over 

• Cork, Limerick, Waterford or Galway cities 

• Dublin City or County. 

In both the LII Survey and the NSHQ we will make use of a 

variable that captures the nature of the household’s tenure 

arrangement. It is reasonable to expect that local authority 

housing, and in particular rented local authority housing, will 

account for a large proportion of concentrated pockets of 

poverty. In establishing the extent to which the poor are 

79



located in such households, we will provide one estimate of 

the extent to which poverty is spatially concentrated. 

For the LII Survey we distinguish between the following 

categories in relation to tenure arrangements: 

• Property owned outright 

• Owned with mortgage 

• Local authority purchaser 

• Local authority tenant 

• Private rented. 

For the NSHQ data we further distinguish within the house 

owner categories (the first and second points above) between 

those who purchase through local authority tenant purchase 

schemes or mortgages on the one hand and those who 

purchase through a private lending agency on the other. 

5.3 Poverty and Deprivation by Area Type 

In Table 5.1 we examine the manner in which risk varies 

across area type and how this is patterned over time. As we 

have seen earlier, there was a substantial increase between 

1994 and 2000 in the numbers falling below the 50 per cent 

poverty line. This trend is evident in Table 5.1 for each of the 

types of area we identified. However, it is particularly true for 

open country areas where the poverty rate rose from 20 per 

cent in 1994 to 32 per cent in 2000 and for villages and 

towns with a population of less than 3,000 where the 

corresponding figures were 26 per cent and 36 per cent. For 

the other categories the rise was a good deal more modest, 

80



being of the order of 6 per cent for towns greater than 3,000, 

3 per cent for the major urban centres outside Dublin and 2 

per cent for Dublin City and County. 

As in earlier years. the highest poverty rate at the 50 per cent 

line in 2000 (36 per cent) was observed in villages and towns 

of less than 3,000 followed by open country, where the rate 

was 32 per cent. In towns greater than 3,000 the rate fell to 

one in four; for urban centres outside Dublin it was just above 

one in five and for Dublin one in six. Thus the disparities 

between areas increased in the second half of the 1990s. 

At the 60 per cent line, on the other hand, the poverty rate fell 

slightly between 1994 and 2000. This occurred fairly evenly 

across areas, apart from an increase in the rate (from 37 to 41 

per cent) in open country areas. However, the ordering of 

areas remained constant over time and identical to that at the 

50 per cent line. The highest risk of 46 per cent was in 

villages and towns less than 3,000 followed by open country, 

with a rate of 42 per cent. The rate then fell significantly to 32 

per cent for towns less than 3,000, to 28 per cent for cities 

outside Dublin and finally to 21 per cent for Dublin. 

Turning to the consistent poverty line, involving falling below 

the 60 per cent income line and the enforced lack of at least 

one basic item, we can see that in earlier years variation by 

type of area was a good deal less evident than in the case of 

income poverty lines. In particular, the situation of those living 

in open country was more favourable. This remained true in 

2000. Between 1994 and 2000 there was a very sharp decline 

in consistent poverty. Substantial reductions were observed 

for all types of areas, but the most dramatic decline was for 

Dublin, where the rate fell from 19 per cent to 3 per cent. In 

the other urban centres it declined from 19 per cent to 7 per 

81



Table 5.1: Risk of poverty by type of area for 1987, 1994 and 2000 

50% income line 60% income line 60% income line 

and basic deprivation 

1987 1994 2000 1987 1994 2000 1987 1994 2000 

Area: % of households 

Open country 21.6 19.6 32.4 34.7 37.4 41.5 15.5 9.6 5.3 

Village/town 3,000 14.2 19.2 25.2 31.5 37.2 32.4 18.1 18.7 10.2 

Waterford, Cork, Galway and 

Limerick cities 18.0 18.4 21.3 28.0 32.0 27.7 21.0 19.2 6.8 

Dublin City and County 8.7 15.1 17.1 16.7 27.4 21.3 12.0 19.2 3.3 

All 17.0 18.6 25.8 29.1 34.8 32.9 16.4 14.9 6.2 


82



Table 5.2: Incidence of poverty by type of area for 1987, 1994 and 2000 

50% 60% 60% Percentage of 

income line income line income line and all households 

basic deprivation 

1987 1994 2000 1987 1994 2000 1987 1994 2000 1987 1994 2000 

Area: % of households 

Open country 46.5 34.3 40.0 42.7 35.6 40.2 33.9 21.3 27.3 35.8 33.1 31.9 

Village/town 3,000 15.1 19.0 19.1 19.5 19.4 19.3 19.8 22.8 32.4 18.0 18.1 19.6 

Waterford, Cork, Galway, 

Limerick cities 9.1 8.5 5.7 8.3 8.0 5.8 11.0 11.3 7.7 8.6 8.7 6.9 

Dublin City and County 13.5 23.5 20.7 15.1 22.9 20.3 19.2 29.5 16.7 26.3 29.5 31.3 

All 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 


83



cent and in towns of more than 3,000 from 19 per cent to 10 

per cent. In towns less than 3,000 the rate goes from 22 per 

cent to 10 per cent and in open country from 10 per cent to 5 

per cent. 

The highest level of consistent poverty in 2000 (10 per cent) 

is observed for villages and towns outside the country’s major 

urban centres. They are followed by urban centres outside 

Dublin with a rate of 7 per cent, open country (5 per cent) and 

Dublin (4 per cent). 

In Table 5.2 we turn from the risk of poverty to the incidence 

of poverty, i.e. the percentage of all poor people who are 

located in each type of area. At the 50 per cent line we 

observe a partial reversal between 1994 and 2000 of a trend 

observed between 1987 and 1994. During the earlier period 

there had been a sharp increase in the proportion of the poor 

located in Dublin (from 14 to 24 per cent), while there was a 

corresponding decline (47 to 34 per cent) in open country 

areas. The figures for the other areas remained relatively 

constant with 15 per cent located in towns and villages less 

than 3,000, 19 per cent in towns greater than 3,000 and 6 per 

cent in the county boroughs outside Dublin (Cork, Galway, 

Limerick and Waterford cities). 

A similar but less pronounced pattern was observed at the 60 

per cent line, once again giving a distribution by 2000 that 

was very close to that observed in 1987 and one just about 

identical to that just described for the 50 per cent line. For 

the consistent poverty line the change over time was uneven. 

In 1987 one-third were located in open country areas. This fell 

to one-fifth in 1994 but increased to over one-quarter in 2000. 

For Dublin, on the other hand, there was a sharp increase 

between 1987 and 1994 and a corresponding decrease by 

84



2000. By that time one in six of the consistent poor were found 

in Dublin. For other urban areas a modest decrease was 

observed between 1994 and 2000, while for villages and towns 

there was no significant change. The clearest trend involved a 

steady increase over time for villages and towns with 

populations less than 3,000 from one in five to one in three. 

Overall, villages and towns of less than 3,000 emerge as 

having a disproportionate number of the poor for each of the 

income poverty lines but not for the consistent poverty 

measure. Towns greater than 3,000, on the other hand, are 

over-represented in terms of consistent poverty. Urban centres 

outside Dublin are slightly under-represented at the income 

lines but not for the consistent poverty line. Dublin, on the 

other hand, is significantly under-represented for all three lines. 

This pattern, whereby rural areas appear less disadvantaged 

according to the consistent poverty measure that takes 

account of standard of living as well as income, is largely due 

to the situation of farm households (see Whelan et al. 2003, 

Tables 4.11 and 5.10). On average, these households emerge 

as having a higher standard of living than their current income 

level would lead us to predict. This may reflect a number of 

factors, including the accumulation of wealth over time in the 

form of savings, the capacity to sell off assets and a lower cost 

of housing than in urban areas. 

5.4 Poverty and Deprivation by Tenure Type 

One of the limitations of the analysis conducted to this point is 

that it is perfectly possible that a more detailed disaggregation 

would reveal pockets of deprivation within the areas we have 

been able to distinguish. We will deal with this in two ways. 

85



Later in the chapter we will take advantage of the 

opportunities provided by the NSHQ. In this section we focus 

on the relationship between poverty and tenure type and in the 

section that follows we combine information about type of 

area and type of tenure. 

In Table 5.3 we look at variation in poverty risk level by tenure 

across all three poverty standards and for each of the three 

years in our analysis. Between 1987 and 1994 there was a 

sharp increase in poverty risk at the 50 per cent line for local 

authority tenants from 37 per cent to 50 per cent, even though 

the overall increase in the poverty rate was rather modest. 

Between 1994 and 2000, in the context of a much sharper 

increase in the overall poverty rate, the poverty rate for local 

authority tenants rose to 66 per cent. At the same time, the 

rate for outright owners increased from 18 per cent to 33 per 

cent. There was also a more modest increase for private 

renters from 15 to 22 per cent. The other categories remained 

relatively stable. 

At the 60 per cent line the picture remained a good deal more 

stable between 1994 and 2000. A slight rise was noted for 

outright owners and local authority tenants and a slight decline 

for all other categories. The risk rate reaches a remarkably 

high level of 78 per cent for local authority tenants. The nexthighest 

rate was observed for outright owners, with a figure of 

42 per cent. This rather high figure is undoubtedly related to 

the relatively high proportion of elderly in this category. Local 

authority tenant purchasers display a risk rate of 33 per cent 

and private renters one of 28 per cent. By far the lowest rate 

is observed for mortgage holders, at 12 per cent. 

Turning our attention to the consistent poverty lines, we find 

that the rate for local authority tenants has fallen from 52 per 

86



cent in 1994 to 32 per cent in 2000. However, given the 

marked decline in risk levels for all of the other categories, 

the relative position of local authority tenants actually 

deteriorated. Thus, for outright owners the rate went from 

11 to 4 per cent, for mortgage holders from 6 to 2 per 

cent and for local authority tenant purchasers from 25 to 5 

per cent. The decline was least for private tenants, where it 

went from 15 to 13 per cent. It is clear that the distribution 

of risk levels is rather different for the consistent poverty 

measure than for the income poverty lines. The greatest 

discrepancy is observed for private renters. Their relative 

position is much worse for the consistent poverty measures. 

Thus, a relatively clear hierarchy of disadvantage emerges 

for consistent poverty, with local authority tenants worst 

off by a significant margin, followed at a distance by private 

renters and then, at a further distance, by the remaining 

groups. 

In Table 5.4 we document the incidence of poverty by tenure 

type. This outcome depends on both the risk levels for each 

type of tenure and the distribution of households across 

tenure types. 

From the last two columns of Table 5.4 we can see that the 

main change in the distribution of households between 1994 

and 2000 involved the continuation of two trends already 

evident in the earlier period. The first involved a further 

decline in the number of local authority tenants from 12 to 7 

per cent. The second involved a continued increase in the 

number of private renters from 8 to 10 per cent. One 

consequence of these patterns, taken together with the 

overall increase in the poverty rate at the 50 per cent line, is 

that despite the sharp increase in the poverty risk for local 

authority tenants, by 2000 they constituted a substantially 

lower proportion of the poor than was previously the case. In 

87



Table 5.3: Risk of poverty by tenure type for 1987, 1994 and 2000 

50% income line 60% income line 60% income line 

and basic deprivation 

1987 1994 2000 1987 1994 2000 1987 1994 2000 

% % % 

Owned outright 16.8 18.1 32.9 30.0 37.8 42.3 12.6 10.5 3.9 

Owned with mortgage 6.7 8.7 8.6 2.5 14.6 11.5 6.3 5.5 1.8 

Local Authority Tenant 

Purchase (LATP) 17.8 21.8 24.7 2.5 41.6 33.9 15.1 24.5 4.5 

Local authority rented 37.4 49.8 66.1 59.1 74.6 77.7 46.8 52.0 32.4 

Other rented* 14.4 15.1 21.9 27.7 34.0 27.5 21.3 15.3 12.6 

All 17.0 18.8 25.9 29.1 34.6 33.0 16.4 14.9 6.2 


*Most ‘Other rented’ dwellings are privately rented, but a small number are rented from voluntary housing 

associations. 

88



1987 and 1994 they made up almost one-third of those below 

the 50 per cent line, but by 2000 this figure had fallen to less 

than one in five. At the same time, the corresponding figure for 

outright owners rose from 41 per cent in 1994 to 56 per cent in 

2000, while the corresponding figures for private renters went 

from 7 to 9 per cent. At the 60 per cent line a similar, although 

less pronounced, pattern was observed. Local authority 

tenants constituted 30 per cent of those below the 60 per cent 

line in 1987. By 1994 this had fallen to 25 per cent, and by 

2000 to 17 per cent. 

There was a corresponding increase between 1994 and 2000 

in the figure for outright owners from 46 to 57 per cent and for 

private renters from 8 to 9 per cent. 

For the consistent poverty line the changes were somewhat 

less dramatic, except in the case of private renters. In 1994 

this group made up 8 per cent of the consistent poor, but by 

2000 this had risen to 21 per cent. The most substantial 

decline was observed for local authority purchasers, with the 

rate going from 10 to 5 per cent. In all other cases the trend 

was downward but of very modest proportions. 

5.5 Disparities in Poverty Risk by Size of Place and 

Tenure 

In Tables 5.5 and 5.6 we present comparable figures from the 

NSHQ. In a later section we will move to a further level of 

disaggregation in this regard. Table 5.5 presents disparities in 

the risk of income poverty by size of place using a more 

detailed breakdown. The figures in the table show the relative 

risk in each category compared to the national average. Thus, 

at the 50 per cent line, the poverty risk in open country areas 

is 20 per cent higher than (or 1.2 times) the national average. 

89



Table 5.4: Incidence of poverty by tenure type for 1987, 1994 and 2000 

50% 60% 60% Percentage of 

income line income line income line and all households 

basic deprivation 

1987 1994 2000 1987 1994 2000 1987 1994 2000 1987 1994 2000 

% % % % 

Owned outright 44.2 40.5 56.2 45.9 45.9 56.7 34.0 29.4 28.1 44.4 42.0 44.2 

Owned with mortgage 11.1 14.5 10.6 12.2 13.5 11.1 10.8 11.8 9.4 28.1 32.0 31.9 

Local Authority Tenant 

Purchase (LATP) 8.1 7.3 6.2 7.3 7.6 6.6 7.1 10.2 4.7 7.7 6.3 6.4 

Local Authority rented 32.1 30.9 18.5 29.6 25.1 17.1 41.3 40.4 37.1 14.5 11.7 7.3 

Other rented* 4.5 6.5 8.6 5.0 7.9 8.5 6.8 8.2 20.8 5.2 8.0 10.2 

All 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 


*Most ‘Other rented’ dwellings are privately rented, but a small number are rented from voluntary housing 

associations. 

90



Table 5.5: Disparities in income poverty risk by urban/rural 

location 

Urban/rural Income Income 

location poverty poverty 

risk 

risk 

(50%) (60%) 

Open country 1.2 1.2 

Town 10,000 1.0 1.0 

Cork, Galway, 

Limerick, Waterford 1.1 1.0 

Dublin City/County 0.7 0.8 

Total 1.0 1.0 


The figures indicate that although the risk tends to be highest 

in rural areas and small towns (population under 3,000) and 

lowest in Dublin City and County, there is not a clear trend by 

size of place at intermediate population densities. The rate is 

below average in towns with a population between 3,000 and 

5,000, about 20 per cent above average in towns with a 

population between 5,000 and 10,000, at the average in 

towns over 10,000 and about 10 per cent above the average 

(but only at the 50 per cent line) in the group comprising the 

cities of Cork, Galway, Limerick and Waterford. 

Table 5.6 shows the disparities in income poverty risk by 

tenure. In this table, the group of householders who own their 

accommodation outright is split in two: those who purchased 

91



through a Local Authority Tenant Purchase scheme (LATP) 

and others. The risk is by far the highest for local authority 

renters, at over twice the average. It is also considerably 

above average (70–90 per cent) for those who now own the 

home outright, but who purchased through a LATP scheme. 

Other households who own the home outright are above 

average in terms of poverty risk, but by a smaller amount (30 

per cent). Private sector renters and local authority tenant 

purchasers have income poverty risk levels that are 10–20 per 

cent below the average across households. Those purchasing 

their accommodation through a private mortgage with a 

lending institution fare best, with risk levels that are 20–30 per 

cent of the overall level. 

To a large extent, the variations across tenure types are due 

to selection issues. In order to obtain a mortgage, a 

household must demonstrate that it has the resources to 

make repayments. Similarly, with increasing rent levels, the 

Table 5.6: Disparities in income poverty risk by tenure 

Tenure Income Income 

poverty risk poverty risk 

(50%) (60%) 

Own outright, LATP 1.9 1.7 

Own outright, other 1.3 1.3 

Purchasing, LATP 0.8 0.9 

Purchasing, other 0.2 0.3 

Local authority renter 2.4 2.2 

Other renter 0.8 0.8 

Rent free 1.2 1.1 

Total 1.0 1.0 


92



private rental market has become increasingly unaffordable 

for households on low incomes. The local authority rented 

sector is a select group by definition: households are eligible 

for local authority housing by virtue of being unable to afford 

to provide adequate accommodation from their own 

resources. Traditionally, as the circumstances of a local 

authority tenant improve, they have become tenant 

purchasers. 

We will turn to a more complete examination of the factors 

accounting for these variations by size of place and tenure in 

the multivariate analysis of Chapter 7. 

5.6 Modified Consistent Poverty by Area Type 

and Tenure 

In the following section, we turn once more to the measure of 

modified consistent poverty (MCP) in the NSHQ data. As 

noted in the previous chapter, MCP involves having an 

equivalised household income below 60 per cent of the 

average and an enforced lack of at least one of the ‘basic’ 

items. 

Table 5.7 presents the disparities in MCP by size of place. 

As we saw before, there is no clear trend by population 

density, and when both income and deprivation are taken into 

account the range of variation by type of area tends to be 

narrower. It is still the case that people living in the open 

countryside are more likely to experience MCP than those 

living in Dublin, but the pattern at intermediate population 

densities is mixed. The smallest towns (population under 

3,000) emerge as most disadvantaged (20 per cent above the 

national level). 

93



Table 5.7: Disparities in risk of MCP by urban/rural location 

MCP 

Open country 1.1 

Town 10,000 1.0 

Cork, Galway, Limerick, Waterford 1.1 

Dublin City/County 0.7 

Total 1.0 


Table 5.8 examines disparities in MCP by tenure. Not 

surprisingly, given their disadvantaged position on virtually all 

the measures examined so far, local authority renters emerge 

as most likely to experience MCP, with a rate 2.8 times the 

national average. 

It is also not surprising that those purchasing on a mortgage 

are least likely to experience MCP, with a rate only 30 per 

cent of the national figure. Among owners, there is a clear 

distinction between those who purchased through the local 

authority (80 per cent above average) and other owners 

(10 per cent above average). Those currently purchasing from 

the local authority are at the average level and private sector 

renters are slightly below the average in terms of MCP levels. 

The small group comprising people occupying the 

accommodation rent free have a MCP rate that is 10 per cent 

above the national figure. 

In order to interpret these patterns it is important to see if 

they can be accounted for by patterns of employment, 

unemployment and social class in the areas or tenure groups. 

94



Table 5.8: Disparities in risk of MCP by tenure 

MCP 

Own outright, LATP 1.8 

Own outright, other 1.1 

Purchasing, LATP 1.0 

Purchasing, other 0.3 

Local authority renter 2.8 

Other renter 0.9 

Rent free 1.1 

Total 1.0 


This is the type of analysis that we will turn to in Chapter 7. 

Before that, in the next chapter we consider the extent to 

which the different dimensions of non-monetary deprivation 

tend to occur together. 


Over time the different measures of poverty behave quite 

differently. There was a sharp increase at the 50 per cent line, 

a modest decrease at the 60 per cent line and a sharp 

decrease for the consistent poverty measure. The increase at 

the 50 per cent line was much sharper in the less populated 

areas. As a consequence, urban-rural differentials, which had 

narrowed between 1987 and 1994, increased substantially 

between 1994 and 2000. At the 60 per cent line, differentials 

remained relatively stable. For the consistent poverty 

measure, differentials, which were already modest, narrowed 

even further. 

Overall, less-populated areas have a disproportionate risk of 

poverty for each type of measure. Analysis using the NSHQ 

95



also indicates that poverty risk tends to be highest in rural 

areas and lowest in Dublin, although the pattern at 

intermediate levels tends to be mixed. However, differences 

connected with size of location are modest in comparison 

with those associated with type of tenure. As before, local 

authority tenants emerge as a particularly disadvantaged 

group. Over time, as they have declined as a proportion of 

the total number of households, they account for a smaller 

proportion of poor households. Their poverty rates for both 

income lines have increased sharply, such that two-thirds are 

below the 50 per cent line and three-quarters are below the 

60 per cent line. While their consistent poverty rate has 

shown some recent decline, the disparities with other groups 

have increased as the level has reached five times the 

average rate. However, the decline in their numbers means 

that they constitute a declining proportion of the income poor 

at each line and a relatively constant proportion of the 

consistent poor. Analysis based on the NSHQ and using a 

more differentiated measure of tenure confirms the major 

disadvantages experienced by local authority tenants and, 

indeed, local authority owners. These are issues to which we 

will return in our multivariate analysis in Chapter 7. 

96


Chapter 6 

Disparities in Deprivation 


In this chapter we move the focus from income poverty to nonmonetary 

indicators of deprivation, again using data from the 

NSHQ. Five dimensions of deprivation are described in the 

next section, and the disparities in deprivation by spatial 

characteristics of households are analysed in the remainder of 

the chapter. The focus is on the extent to which the different 

dimensions of deprivation tend to vary across spatial 

categories such as region, local authority, area type and 

tenure group. A second focus in this chapter is on whether 

there is evidence of multiple deprivation, i.e. where relative 

disadvantage along one dimension of deprivation is 

associated with relative disadvantage along another. 

The dimensions of deprivation are as set out in Table 6.1. 

6.2 Methodology 

In the following tables, we will focus on disparities in 

deprivation across sub-groups in the population rather than on 

levels of deprivation. Five distinct dimensions of deprivation 

are discussed: basic, secondary, housing amenities, housing 

97



Table 6.1: Dimensions of deprivation (cannot afford items) 

Basic 

Meal with meat, etc. 

New clothes 

Adequate heating 

Socialising once a month 

Replacing worn furniture 

Week’s holiday per year 

Arrears on housing/utility bills 

Housing amenities 

Flush toilet 

Bath/shower 

Hot running water 

Secondary 

Microwave 

Dishwasher 

Colour TV 

Video recorder 

Telephone, incl. mobile 

Housing deterioration 

Leaking roof/doors/windows 

Dampness 

Rot in doors/windows/floors 

Noise from neighbours 

Accommodation too small for needs 

Environment 

Vandalism 

Graffiti 

Rubbish/litter 

Homes/gardens in bad condition 

Public drunkenness 

deterioration and environment. The items included in each of 

the measures of deprivation are shown in Table 3.1 in Chapter 

3. As in the previous chapter, the results will be presented in 

terms of the ratio of the level in a sub-group, e.g. a local 

authority area, to the overall level nationally. 

6.3 Disparities in Deprivation by Region and Local 

Authority Area 

Table 6.2 presents the disparities in deprivation by regional 

authority area. The figures are also shown in Maps 6.1a to 

98



6.1e. Basic deprivation involves an inability to afford basic 

necessities such as adequate food, clothing and heating for 

the home, as well as socialising, an annual holiday and the 

presence of arrears on utility bills. The Border and Midland 

regions fare worst in terms of basic deprivation, while Dublin 

and the South-East fare most favourably. The Dublin region 

fares best, with levels of basic deprivation 20 per cent lower 

than the average. Earlier we showed that the South-East 

and the West had higher levels of income poverty than the 

country as a whole, but this is not apparent for basic 

deprivation. The Western region is about average in terms of 

basic deprivation, while the South-East is somewhat below 

average. 

The pattern in terms of secondary deprivation is less 

variable across regions. Secondary deprivation consists of 

an inability to afford a car and an inability to afford various 

common household appliances that the householder would 

like to have (microwave, dishwasher, colour TV, video 

recorder and telephone). Levels tend to be below average in 

Dublin and the South-West and are highest in the West. 

The housing amenities items (absence of flush toilet, bath or 

shower and hot running water) will mainly reflect an older 

housing stock. The levels are highest in the Border (40 per 

cent above average), the South-East, South-West and Mid- 

West (20–30 per cent above average). The levels are lowest 

in the Mid-East (40 per cent below average) and Dublin (30 

per cent below), and are also below average in the Midlands 

and West. 

Housing deterioration includes items such as a leaking 

roof/doors/windows, dampness, rot in window or 

doorframes or other timbers and noise from neighbouring 

99



Table 6.2: Disparities in risk of basic, secondary, housing and 

environmental deprivation by regional authority area 

Region Basic Secondary Housing Housing Environamenities 

deterioration ment 

Border 1.4 1.0 1.4 1.0 0.5 

Dublin 0.8 0.9 0.7 1.2 2.1 

Mid-East 1.0 1.0 0.6 0.9 0.8 

Midlands 1.2 1.1 0.9 1.1 0.4 

Mid-West 1.0 1.0 1.2 1.0 0.7 

South-East 0.9 1.0 1.3 1.0 0.7 

South-West 1.0 0.9 1.3 0.9 0.6 

West 1.0 1.2 0.9 0.8 0.2 

Total 1.0 1.0 1.0 1.0 1.0 


houses. The items are based on the householder identifying 

these problems as present in the accommodation and as 

constituting a moderate or major problem. It follows a 

different pattern by region than the housing amenities index, 

with the greatest level of problems in the Dublin area (20 per 

cent above average) and the lowest levels in the West (20 per 

cent below average). 

Finally, the environmental dimension of deprivation consists 

of various problems of public order being very common in the 

area, such as graffiti, vandalism and litter. The main contrast 

in terms of the environment index is between Dublin, with a 

level over twice the average, and the other regions. The 

lowest level of environmental problems is reported in the 

West (20 per cent of the average level). 

We saw in the last chapter that regional patterns may mask 

considerable levels of variation between local authority areas 

100



Map 6.1a: Disparities in risk of deprivation by regional authority 

area – basic deprivation 

101



Map 6.1b: Disparities in risk of deprivation by regional authority 

area – secondary deprivation 

102



Map 6.1c: Disparities in risk of deprivation by regional authority 

area – housing amenities 

103



Map 6.1d: Disparities in risk of deprivation by regional authority 

area – housing deterioration 

104



Map 6.1e: Disparities in risk of deprivation by regional authority 

area – environmental deprivation 

105



within a region. Table 6.2 and Maps 6.2a to 6.2e examine the 

disparities in deprivation by local authority area to see if this 

is also the case for the non-monetary items. 

This is clearly the case for basic deprivation within the Border 

region. The level is highest in Donegal, at 80 per cent above 

the average nationally, and is lowest in Monaghan, at 20 per 

cent below the national average. The variation is more muted 

for secondary deprivation, but some counties have a level 

below average (Monaghan, Leitrim and Louth), while the 

others have a level above the national average. Apart from 

Louth, all of the Border counties evince above average levels 

of deprivation in terms of housing amenities, rising to double 

the average in Cavan and Leitrim. 

In the Midland region, basic deprivation is highest in Laois 

and Longford (30–40 per cent above average) and is at the 

national average in Westmeath. Differences between the 

counties of the South-East in terms of basic deprivation are 

also quite marked: Carlow has a level 30 per cent above 

average while Kilkenny has a figure 40 per cent below 

average. 

Housing deterioration shows a more mixed pattern, but with 

levels in the Border counties tending to be close to or below 

the average nationally. The final column in Table 6.3 indicates 

that the Border counties are all considerably less likely than 

average to face problems in terms of public disorder in the 

environment. 

A similar set of points can be made in the case of the pattern 

by county in the other regions. 

• The patterns of deprivation do not necessarily correspond 

to the patterns in terms of income poverty. Counties 

106



Table 6.3: Disparities in risk of basic deprivation, secondary 

deprivation, housing deprivation and environmental deprivation 

by local authority 

County Basic Secondary Housing Housing Environamenities 

deterioration ment 

Cavan 1.2 1.0 2.0 1.0 0.4 

Donegal 1.8 1.2 1.2 0.8 0.5 

Leitrim 1.1 0.9 2.1 0.5 0.2 

Louth 1.4 0.9 1.0 1.1 0.7 

Monaghan 0.8 0.7 1.1 1.0 0.3 

Sligo 1.6 1.3 1.7 1.0 0.4 

All Border region 1.4 1.0 1.4 1.0 0.5 

Laois 1.4 1.1 0.9 1.0 0.4 

Longford 1.3 1.2 1.4 0.9 0.2 

Offaly 1.2 0.9 0.6 1.0 0.4 

Westmeath 1.0 1.3 1.0 1.3 0.4 

All Midlands region 1.2 1.1 0.9 1.1 0.4 

Galway City 0.9 1.2 0.4 0.8 0.6 

Galway County 0.9 0.9 1.0 0.9 0.1 

Mayo 1.2 1.6 0.9 0.8 0.1 

Roscommon 1.2 1.1 1.6 0.7 0.1 

All West region 1.0 1.2 0.9 0.8 0.2 

Dublin City Council 0.9 1.3 1.0 1.3 3.0 

Dublin Fingal 0.7 0.6 0.3 0.9 0.9 

Dublin South 0.6 0.6 0.3 1.1 1.6 

Dún Laoghaire-Rathdown 0.7 0.7 0.4 1.0 1.0 

All Dublin region 0.8 0.9 0.7 1.2 2.1 

Kildare 0.8 0.9 0.5 1.0 0.9 

Meath 1.1 1.1 0.5 0.8 0.4 

Wicklow 1.2 1.0 1.0 1.0 1.2 

All Mid-East region 1.0 1.0 0.6 0.9 0.8 

Clare 1.2 1.2 1.1 0.7 0.6 

Limerick City 1.1 1.1 0.9 1.5 2.0 

Limerick County 0.7 0.8 1.4 0.9 0.3 

Tipperary North Riding 1.3 1.1 1.5 1.0 0.5 

All Mid-West region 1.0 1.0 1.2 1.0 0.7 

Carlow 1.3 1.6 1.4 1.2 1.2 

Kilkenny 0.6 0.9 1.4 0.8 0.4 

Tipperary South Riding 1.1 1.2 1.3 1.0 0.5 

Waterford City 1.1 1.5 0.8 1.0 1.5 

Waterford County 0.9 0.7 1.5 0.7 0.4 

Wexford 0.8 0.8 1.3 1.1 0.5 

All South-East region 0.9 1.0 1.3 1.0 0.7 

Cork City 1.0 1.0 1.4 1.1 1.0 

Cork County 1.0 0.8 1.1 0.9 0.5 

Kerry 1.1 1.0 1.9 0.8 0.2 

All South-West region 1.0 0.9 1.3 0.9 0.6 

Total 1.0 1.0 1.0 1.0 1.0 


107



Map 6.2a: Disparities in risk of deprivation by local authority 

area – basic deprivation 

108



Map 6.2b: Disparities in risk of deprivation by local authority 

area – secondary deprivation 

109



Map 6.2c: Disparities in risk of deprivation by local authority 

area – housing amenities 

110



Map 6.2d: Disparities in risk of deprivation by local authority 

area – housing deterioration 

111



Map 6.2e: Disparities in risk of deprivation by local authority 

area – environment 

112



which rank least favourably in terms of income poverty 

within a region (see Table 4.3 in Chapter 4) do not 

necessarily rank at the bottom, or even as below average, 

in terms of deprivation. 

• Basic and secondary deprivation also appear to be 

patterned differently, particularly in the counties of the 

Border, Midlands and West regions. 

• Housing amenities and housing deprivation appear to be 

capturing phenomena with a very different distribution 

across areas. Counties which are below average on one 

dimension will often be above average on the other. 

• The environment dimension, which captures problems of 

public order, is clearly capturing problems which are more 

common in urban than in rural areas. 

6.4 Disparities in Deprivation by Size of Place 

and Tenure 

At this point we turn to disparities in the dimensions of 

deprivation by size of place and tenure. Table 6.4 presents 

disparities in the dimensions of deprivation by size of place. 

As with income poverty, basic deprivation tends to be more 

prevalent in rural areas and less prevalent in Dublin, but it is 

the smaller towns (under 3,000 and 3,000–5,000) rather than 

the open countryside that fare worst. The pattern in terms of 

secondary deprivation is different: the open countryside and 

the Dublin region are most advantaged in this respect, but 

differ from the average by only 10 per cent. An absence of 

housing amenities is clearly a problem of open country areas, 

where the prevalence is 70 per cent higher than nationally. On 

the other hand, dwellings in the open countryside fare better 

113



than average in terms of housing deterioration, while the 

Dublin region fares worst. The final column again confirms 

that problems in the social environment are related to 

urbanisation: they are rarest in the open countryside and 

increasingly prevalent as population density increases. 


environmental deprivation by urban/rural area 

Urban/ Basic Second- Housing Housing Environrural 

area ary amenities deteriora- ment 

tion 

Open country 1.1 0.9 1.7 0.8 0.2 

Town 10,000 1.0 1.1 0.8 1.0 0.9 

Cork, Galway, Limerick, 

Waterford 1.1 1.2 0.8 0.9 1.1 

Dublin City/County 0.8 0.9 0.7 1.2 2.1 

Total 1.0 1.0 1.0 1.0 1.0 


Table 6.5 shows that the disparities in deprivation by tenure 

tend to be greater than by size of place. Apart from housing 

amenities, local authority renters fare worst on all dimensions, 

although private sector renters are not far behind them in 

terms of secondary deprivation. Those who own outright 

having purchased through a local authority have levels of 

basic and secondary deprivation that are about 30 per cent 

above average, but they are closer to average in terms of the 

114



housing and environment items. Other households who own 

the accommodation outright are somewhat below average in 

terms of basic deprivation and secondary deprivation, 

housing deterioration and problems in the environment. 

However, it is among this group and the much smaller group 

occupying the accommodation rent free that housing 

amenities are most likely to be lacking. 


environmental deprivation by tenure 

Basic Second- Housing Housing Environary 

amenities deteriora- ment 

tion 

Own outright, LATP 1.3 1.3 1.1 0.9 0.9 

Own outright, other 0.9 0.8 1.6 0.7 0.6 

Purchasing, LATP 1.1 1.0 0.4 1.4 1.8 

Purchasing, other 0.6 0.5 0.2 0.7 0.7 

Local authority renter 2.6 2.8 1.3 2.6 4.0 

Other renter 1.3 2.0 0.9 1.5 1.2 

Rent free 1.1 1.3 2.4 1.0 0.3 

Total 1.0 1.0 1.0 1.0 1.0 


Tenant purchasers and those purchasing on a private 

mortgage looked rather similar in terms of income poverty 

risk, at least in so far as both groups had a lower risk than 

average. However, the deprivation items point to some 

important differences between these groups. While private 

purchasers are much less likely than average to experience 

any of the types of deprivation shown in the table and are 

115



least likely to experience basic or secondary deprivation or 

to lack housing amenities, tenant purchasers fare worse 

than average on a number of dimensions. They are slightly 

above average in terms of basic deprivation and 

considerably above average in terms of housing 

deterioration (40 per cent) and problems of public disorder 

(80 per cent). 

Private sector renters rank second to local authority renters 

in terms of levels of basic and secondary deprivation and 

housing deterioration and are also less favourably placed 

than average in terms of problems of public order, but they 

fare slightly better than average in terms of housing 

amenities. 


Regional disparities differ across the five dimensions of 

deprivation. For the basic deprivation items, the by now 

familiar pattern whereby the Border region is at one end of 

the continuum and Dublin at the other is observed. For the 

secondary deprivation index, which incorporates consumer 

items that can be accumulated over a period of time, spatial 

variation is extremely modest. When we focus on housing 

amenities, which cover the presence of rather basic 

facilities, Dublin and the Mid-East fare best. In contrast, 

housing deterioration is at its worst in Dublin, although 

regional variation is rather modest. Environmental problems, 

which are primarily of a social character, serve to sharply 

differentiate Dublin from all other regions. 

When we disaggregate the regions and look at the individual 

local authority it is clear that, as with income poverty, there 

116



is considerable variation within regions. Further, the two 

housing deprivation sets of items – housing amenities and 

housing deterioration – are clearly seen to be independent 

phenomenon. 

The implications of urban or rural location vary by dimension 

of deprivation, as was clear from the analysis by area type. 

For the basic and secondary dimensions such variation is 

modest, although Dublin occupies the most favourable 

position on both dimensions. Variation is similarly modest for 

housing deterioration, but on this occasion Dublin occupies 

the least favourable position. Problems in the social 

environment display a clear urban-rural gradient, being 

considerably more prevalent in urban areas. 

The pattern by tenure shows both substantially higher 

disparities and more striking patterns of overlapping or 

multiple deprivation. Thus, local authority tenants are by 

some distance the most disadvantaged group on four of the 

five dimensions. The exception is housing amenities, where 

the least favourable circumstances are associated with being 

a private tenant. 

Overall, then, the main findings in this chapter were that: 

• The disparities in deprivation by tenure tend to be much 

greater than the differences between regions, local 

authority areas or by size of place. 

• As was the case for income poverty, there are substantial 

differences in deprivation level between the counties 

within a given regional authority area. 

• Apart from local authority renters who fare worst on four 

of the five dimensions of deprivation, there is only a weak 

association between the different types of deprivation. 

117



Areas or tenures that fare badly on one dimension do not 

necessarily fare badly on the others. Evidence of multiple 

deprivation that is structured along spatial lines, then, is 

relatively weak. 

118


Chapter 7 

Understanding Variation in Poverty by 

Area and Tenure Type 


In addition to documenting variation in poverty rates by 

location and tenure, we also need to get some sense of how 

the magnitude of such variation compares with the scale of 

variation within areas and tenure type. Following on from this, 

we seek to assess the extent to which such variation is 

causal in nature. To what extent could the patterns we 

observe arise solely because poverty and tenure/location are 

jointly associated with other factors that are the true 

determinants of poverty 

7.2 Assessing the Net Effect of Location and Tenure 

The fact that a specific type of area or tenure has a relatively 

high poverty rate does not in itself indicate anything about the 

impact of location or tenure per se on poverty – such effects 

could be entirely attributable to the socio-economic 

composition of the households involved. Households renting 

local authority housing could be at a high risk of poverty 

because applicants for such housing tend to be drawn from 

119



the most vulnerable sectors of the population: the 

unemployed, lone parents, the elderly and those unable to 

work. Cross-tabulations can only take us so far in 

understanding variations in poverty risk by area/tenure type at 

a point in time or the changes in this pattern observed over 

time. We therefore need to systematically examine the extent 

to which the observed variation in risk of poverty by area and 

housing tenure may be attributable to differences in the 

measured characteristics of those located in different areas or 

tenure groups, using logistic regressions. This procedure 

allows one to assess the effect of any particular factor on the 

odds of being poor while holding the influence of other 

factors constant. 

7.3 Methodology 

The notion of odds is one familiar to all gamblers. Suppose 

20 per cent of the population are poor. Instead of saying that 

the probability of being poor is 0.2 and of not being poor is 

0.8, we can say that the odds of being poor is 0.25, i.e. 

0.2/0.8, and the odds on being non-poor is 4:1 (0.8/0.2). An 

actual example will help to further illustrate the relationship 

between risk of poverty and odds ratios. In 2000, 3.9 per cent 

of outright owners fell below the combined 60 per cent mean 

income and basic deprivation line, giving them a 0.04 

(3.9/96.1) odds of being poor. On the other hand, 32.4 per 

cent of local authority tenants were found to be below the 

consistent poverty line, giving them a 0.48 (32.4/67.6) odds of 

being poor. The disparity in risk of being consistently poor 

between these two groups (local authority tenants and 

outright owners) can be indexed by the ratio of these odds 

(.48/.04), which gives an odds ratio of 12:1. In other words, 

120


Understanding Variation in Poverty by Area and Tenure Type 

the odds on local authority tenants being poor are 12 times 

higher than the odds for outright owners. By choosing any 

group as a reference point, we can summarise the set of 

inequalities between that group and all others. It is more 

convenient to work with odds rather than probabilities 

because it allows us to make explicit comparisons between 

groups. 

In the analysis that follows we focus on the NSHQ data in 

estimating gross and net effects, although we also report 

supplementary analysis relating the 2000 LII Survey in 

Appendix 1. We concentrate on the NSHQ because of the 

advantages conferred by its substantially larger sample size. 

This makes it possible to employ a variable relating to 

housing tenure that is a good deal more differentiated than 

the version we have used with LII data. Earlier, and in 

Appendix 2, we noted the similar patterns of disparities 

revealed by the LII Survey and NSHQ data sets. In the 

analysis that follows we would expect some differences to 

emerge. First, where a more differentiated tenure variable is 

employed, we would anticipate that larger disparities would 

be observed. In addition, since the range of sociodemographic 

variables for which we can control is narrower 

in the NSHQ, we would expect the net effects to be larger in 

this survey. However, the broad picture painted by both 

surveys is extremely consistent. 

The analysis proceeds as follows. First we will present the 

odds ratios for being poor rather than non-poor by type of 

region. We will then consider the impact on such ratios of 

controlling for a range of socio-demographic variables, 

including employment status, social class, age group of the 

householder, number at work in the household and household 

type. When implementing such a procedure it is impossible to 

121



be sure that we have controlled for all relevant factors. 

However, the survey information available to us allows us 

to control for a comprehensive set of variables, and as the 

number of variables increases, the impact of such controls 

will substantially diminish because of the correlation of 

later control variables with earlier ones. We then proceed 

to consider the impact of size of location and tenure and, 

crucially, the manner in which they interact. 

7.4 Gross and Net Effects of Region on Poverty in 

the NSHQ 

In Table 7.1 we show the gross and net odds ratios for 

being poor at 50 per cent and 60 per cent of mean income 

and for the modified consistent poverty (MCP) measure 

broken down by regional authority, with Dublin taken as a 

reference point. The gross odds ratios summarise the 

differences between groups before taking account of the 

socio-demographic controls, whereas the net ratios refer 

to the remaining differences once the effects of the sociodemographic 

variables have been statistically controlled. 

We focus first on the gross effects. At the 50 per cent 

income line, Dublin is sharply differentiated from all other 

regions. The most disadvantaged regions are the West and 

the Border, with odds ratios of 1.78 and 1.75, respectively. 

The values for the remaining regions are located in the 

range running from 1.38 to 1.46. When we control for the 

socio-demographic characteristics of the householder, the 

range of differentials is slightly narrowed, with the West 

and the Border region now having odds ratios of 1.64 to 

1.66. 

122



An explanation of this finding is provided when we examine 

the level of variance explained with and without the inclusion 

of socio-demographic controls. In order to assess the impact 

of variables such as region, we distinguish between the 

variance that is unique to each influence and that which is 

shared between them. We identify these components by 

varying the order of entry of each type of variable. Thus, the 

variance that is unique to the variable region is that which is 

added after socio-demographic variables have been taken 

into account and vice versa. The shared component is that 

which cannot be uniquely allocated to either variable. Unlike 

ordinary least squares regression, there is no universally 

accepted measure of explained variance for logistic 

regression. We have reported the Nagelkerke R 2 ; however, 

our conclusions relating the relative importance of the types 

of effects we are considering would not be substantially 

affected by opting for another measure. 

From Table 7.1 we can see that identifying the regional 

authority does very little to enhance our understanding of the 

factors leading to respondents being exposed to poverty at 

the 50 per cent income line. In fact, region uniquely accounts 

for effectively none of the variance. Poverty risk at this level is 

being predicted almost entirely by the socio-demographic 

characteristics of the householder. This indicates that rather 

than it being possible to explain how regional differences vary 

by taking socio-demographic differences into account, the 

scale of such difference is actually tiny in comparison with 

those being produced within regions by differences between 

households in socio-demographic background. 

When we focus on the 60 per cent line we find a similar but 

slightly less extreme position. In terms of gross effects, the 

Border region is now clearly the most disadvantaged area, 

123



with a ratio exceeding two to one. At the other end of the 

continuum, the South-East is close to Dublin, with a ratio of 

1.23. All the remaining regions are found in the narrow range 

running from 1.71 to 1.87. Introducing socio-demographic 

controls actually increases the coefficients for all regions 

relative to Dublin, with the Border, Midlands, South-East, 

South-West and West regions also displaying odds ratios 

above two. This indicates that rather than it being possible to 

explain how regional differences vary by taking sociodemographic 

differences into account, the scale of such 

difference is actually less than we would have expected on 

the basis of socio-demographic variation across region. Once 

again, though, the explanatory power of region is modest, 

with less than 1 per cent of the explained variance being 

uniquely attributable to region, 2 per cent being shared 

between it and socio-demographic characteristics and 97 per 

cent being uniquely attributable to the latter. 

When we turn our attention to the MCP measure, regional 

differences are again found to be modest in terms of their net 

contribution to explained variation. The Border region is the 

only case where the odds ratio exceeds two, and with the 

exception of the Mid-East, which is closest to Dublin, all 

other regions are located in the range running from 1.77 to 

1.84. The pattern of gross coefficients is again very close to 

that of their net counterparts. Again, however, only 1 per cent 

of the explained variance is uniquely attributable to region. 

Thus, a clear picture emerges across poverty lines that while 

Dublin and the Mid-East are consistently at one end of the 

continuum of disadvantage and the Border region at the 

other, exposure to poverty has a great deal to do with sociodemographic 

differences between households within a region 

and very little to do with other characteristics of the regions. 

124



Table 7.1: Gross and net odds ratios of being poor versus nonpoor 

by regional authority 

50% income 60% income MCP 

line 

line 

Region Gross Net Gross Net Gross Net 

Dublin 1.00 1.00 1.00 1.00 1.00 1.00 

Border 1.75 1.66 2.06 2.53 2.23 2.39 

Mid-East 1.46 1.23 1.23 1.62 1.34 1.60 

Midlands 1.46 1.31 1.72 2.23 1.84 2.11 

Mid-West 1.38 1.43 1.71 1.83 1.83 1.85 

South-East 1.54 1.59 1.87 2.13 1.78 1.70 

South-West 1.47 1.56 1.80 2.02 1.84 1.76 

West 1.78 1.64 1.82 2.11 1.77 1.80 

% unique to 

region 0.00 0.94 1.18 

Nagelkerke R 2 1 0.005 0.551 0.014 0.640 0.014 0.509 

% common 0.09 2.26 1.57 


7.5 Disparities in Poverty by Local Authority Areas 

(LAA) 

It might be argued that the relatively small number of regional 

units may lead us to underestimate the importance of spatial 

variation since a great deal of such variation may occur within 

rather than between regions. In Chapter 4 we observed 

1 Nagelkerke R 2 is a modification of Coz and Snell’s R 2 . The latter is an 

attempt to imitate the interpretation of the multiple R 2 in OLS. 

However, its maximum can be, and usually is, less than one. 

Nagelkerke R 2 involves a modification of the Cox and Snell coefficient 

to ensure that it can vary from 0 to 1. 

125



that the risk of income poverty and of MCP could vary 

substantially between the counties within a region. In Table 

7.2 we extend our analysis to take variation across 34 local 

authority areas into account. From row one in Table 7.2, it 

is clear that spatial variation in poverty risk continues to be 

modest, even with a greatly differentiated set of spatial 

units. The Nagelkerke R 2 estimate varies from 0.02 at the 

50 per cent line to .039 at the 60 per cent line. When our 

set of socio-demographic controls is added, the range runs 

from .529 for the MCP line to .666 for the 60 per cent line. 

The unique contribution of the local authority variable 

ranges from 1.7 per cent of the explained variance at the 

60 per cent line to 2.6 per cent at the MCP. In contrast, the 

socio-demographic variables that we use as controls 

explain a low of 92.8 per cent of the explained variance for 

the MCP measure and a high of 96.4 per cent of such 

variance in the case of the 50 per cent line. Even with the 

substantially more differentiated measure, variation across 

geographical areas plays a very modest role in explaining 

exposure to risk of household poverty in comparison with 

variation in the socio-demographic characteristics of 

households within such areas. 

Table 7.2: Poverty variation between and within local authority 

areas 



7.6 Gross and Net Effects of Size of Area and Tenure 

on Risk of Poverty in the NSHQ 

In Figure 7.1 we address the issue of the combined impact on 

risk of poverty of urban-rural location and tenure at the 50 per 

cent level. Urban is defined as being located in Dublin, Cork, 

Galway, Limerick or Waterford. For tenure the reference 

category is private purchasers and the remaining categories 

distinguish between local authority owners, private owners, 

local authority purchasers and local authority tenants. It is 

immediately evident that variation in risk of poverty by type of 

tenure within urban-rural location is a good deal sharper than 

by regional authority. There are two countervailing influences 

that it is necessary to take into account. In the first place, 

rural poverty rates are on average higher than urban rates. 

This is clearly the case with private purchasers, where the 

odds ratio for the rural group is almost twice that of their 

urban counterparts. However, as we shall see, variation in 

poverty risk by tenure is substantially weaker in rural areas. 

As a consequence, urban-rural differences are considerably 

weaker outside the private purchasers group. 

If we focus first on gross effects, we find that in urban areas 

private purchasers occupy a particularly favoured position. 

Private owners have an odds ratio almost five times higher, 

local authority (LA) purchasers six times higher, LA owners 

seven times higher, private tenants nine times higher and LA 

tenants 27 times higher. Thus, in urban areas private 

purchasers and LA tenants are found at opposite ends of the 

spectrum (1 and 27, respectively), while other tenure groups 

are located in a relatively modest range (5 to 9). When we 

compare urban and rural residents by type of tenure, we find 

fairly similar odds ratios. Thus, urban tenants are not more 

vulnerable than rural tenants. However, given the less 

127



favourable position of rural private purchasers, this means 

that the impact of type of tenure is considerably stronger in 

urban areas. To illustrate this point, we focus on urban and 

rural LA tenants. These have almost identical odds ratios of 

27:1 and 28:1, in comparison with urban private purchasers. 

However, when we compare rural tenants with rural private 

purchasers, this reduces to less than 15:1. Similarly, for LA 

owners the rural specific ratio is 4:1 compared to 7:1 for their 

urban counterparts; for private tenants the respective figures 

are 5:1 and 9:1. Therefore, on an overall ranking of 

disadvantage, rural residents on average do slightly worse 

than their urban counterparts. However, this situation is 

brought about by the combined influence of being located in 

a rural area and a rather weaker impact of type of tenure than 

prevails in urban areas. 

When we focus on net odds ratios having taken into account 

a range of socio-demographic variables, we observe a 

substantial reduction in the magnitude of the observed 

coefficients. In particular in the case of LA tenants, we see a 

reduction in the size of the relevant odds ratio from 27:1 to 

4:1 for the urban group, constituting a reduction of 86 per 

cent in the size of the coefficient. For their rural counterparts, 

in comparison with rural private purchasers their relative 

disadvantage falls from just under 15:1 to just below 3:1, 

constituting a reduction of almost 80 per cent in the size of 

the effect. The range of effects for other forms of urban 

tenure decline from 4.6–8.9 to 1.6–3, involving reductions 

running from over two-thirds in the case of private tenants to 

just under a half in the case of LA purchasers. The 

corresponding rural relativities decline from 5:1–3:1 to 

1.8:1–1.4:1. Thus, the vast bulk of the impact of the 

combination of urban-rural location and tenure is accounted 

128



for by the socio-demographic control variables. The 

remaining effects may be a consequence of the independent 

effect of location or tenure or it may, at least in part, be 

accounted for by selection effects relating to unmeasured 

variables. One further indication of the relative importance of 

different factors is provided by an examination of the division 

of the explained variances into its unique and common 

components. Our analysis shows that that 1.8 per cent of the 

variance can be uniquely attributed to the combination of 

location and tenure, 16.6 per cent is shared between these 

factors and the socio-demographic controls and finally, 81.6 

per cent can be uniquely attributed to the latter. 

Figure 7.1: Gross and net ratios of being poor versus non-poor 

at the 50 per cent income line by tenure and urban-rural 

location 

It could be argued that the limited importance of location in 

accounting for poverty risk arises because the urban category 

combines cities with different levels of poverty risk. 

129



In Chapter 4 it was indeed clear that the risk of poverty in 

Dublin was quite different from that in the other large cities. 

In Figure 7.2 we report a comparable analysis that now 

distinguishes between Dublin and all other locations. Again, 

the introduction of socio-demographic controls 

substantially reduces the size of the coefficients. The 

net coefficients showing the impact of tenure within the 

Dublin area are not very different for those reported for 

urban areas generally in Figure 7.1. However, the net 

coefficient for LA tenants rises to almost six (from 4.4 in 

Figure 7.1). 

However, the comparable rural coefficients all show an 

increase indicating a sharper pattern of differentiation by 

tenure. The net non-Dublin odds ratios, showing the level of 

disadvantage in relation to Dublin private purchasers, are all 

(apart from LA purchasers) higher than the corresponding 

Dublin figure. Thus, exposure to risk of poverty is higher for 

those outside Dublin in each tenure category, including LA 

tenants. However, when internal Dublin and non-Dublin 

relativities are compared, the former are consistently higher. 

Thus, Dublin LA tenants exhibit a net odds ratio that is six 

times higher than private purchasers in Dublin, while the 

corresponding rural figure is only about half this size 

(6.8/2.12 = 3.4). The Dublin/non-Dublin gap is a good deal 

less sharp in net terms for other categories of tenure, but in 

every case the relative degree of disadvantage is clearly 

greater in the case of Dublin. The major factor contributing 

to this more substantial disparity in the case of Dublin is 

not a distinctive level of disadvantage experienced by 

urban LA tenants but the uniquely favourable situation 

associated with being a private purchaser in Dublin and the 

consequent disparity between the latter and the former. 

130




at the 50 per cent income line by tenure and Dublin versus 

elsewhere 

Overall we can see that, with the exception of the urban local 

authority tenant effect, the net impact of tenure is relatively 

modest. Taken together, tenure and location in Dublin or 

outside Dublin uniquely account for just over 2 per cent of the 

explained variance and jointly account for 16 per cent 

together with the socio-demographic controls. The remaining 

80 per cent plus can be attributed uniquely to the latter. 

In Figure 7.3 we turn our attention to the 60 per cent income 

poverty line. Here we observe a pattern not very different 

from that observed for the 50 per cent line. Although the 

gross and net rural coefficients tend to be higher, relativities 

remain larger in the urban area. Once again, the introduction 

of socio-demographic controls substantially reduces the size 

of the coefficients. However, for urban LA tenants the 

reduction is somewhat smaller than was the case for the 50 

per cent line and the fall from 27:1 to just over 7:1 constitutes 

131



a reduction of 74 per cent compared to one of 84 per cent in 

the case of the 50 per cent line. An almost identical degree of 

reduction is observed in the rural case, which once again is 

on a somewhat smaller scale than was the case for the 50 

per cent line. 


at the 60 per cent income line by tenure and urban-rural 

location 

Figure 7.4 presents a comparable analysis for the 60 per cent 

line that now distinguishes between Dublin and elsewhere. 

Here we observe higher coefficients, both gross and net, for 

those outside Dublin than was the case at the 50 per cent 

line. LA owners, both Dublin and other, are more 

disadvantaged than was the case at the 50 per cent line. The 

odds ratio relating to the comparison of rural local authority 

tenants with Dublin private purchasers now comes close to 

50:1. However, when the comparison is made to rural private 

purchasers, the figure falls to 16:1, a level of inequality that is 

just above half that observed when the corresponding Dublin 

132



groups are compared. Once again we see the pattern 

whereby rural LA tenants constitute by some distance the 

most disadvantaged group, but their distinctive position 

comes from a combination of location and tenure effects. If 

on the other hand we focus on tenure effects per se within 

urban and rural locations, the relative disadvantage enjoyed 

by Dublin LA tenants is substantially greater. Similar contrasts 

exist between absolute and relative disadvantage for other 

tenure groups. 


at the 60 per cent income line by tenure and Dublin versus 

elsewhere 

Once again, the net coefficients are of a considerably smaller 

magnitude than their gross counterparts, although in each 

case they are substantially larger than those observed at the 

50 per cent line. However, despite such increases, the unique 

variance accounted for by location and tenure remains a 

modest 2.5 per cent compared to the corresponding figure of 

almost 80 per cent for socio-demographic factors. 

133



Finally, we turn our attention to the modified consistent 

poverty (MCP) measure. Figure 7.5 shows the breakdown of 

odds ratios by urban-rural location and tenure. While the 

pattern here is broadly similar to that observed at the 60 per 

cent line, the position of rural local authority tenants appears 

to be rather worse than was the case in that instance. In both 

the urban and rural cases the gross odds ratio is higher for 

the MCP. In the former case it exceeds 30 and in the latter it 

exceeds 40. For the 60 per cent line, a focus on internal 

urban and rural relativities showed the latter to be 

approximately twice the former; for the MCP they come 

closer to 60 per cent higher. For both the urban and rural 

cases the net coefficients involve a sizeable reduction of 70 

per cent on their gross. 


at the MCP line by tenure and urban-rural location 

The combination of urban-rural location and tenure uniquely 

accounts for 4.7 per cent of the explained variance, which is 

rather higher than for either of the income lines. The shared 

134



variance of 22 per cent is also slightly higher than in the 

earlier cases. The remaining 70 per cent plus is attributable 

solely to the socio-demographic controls. 

Figure 7.6 provides a comparable analysis for the MCP 

measure, distinguishing between Dublin and elsewhere. This 

analysis produces the highest odds ratios to date, particularly 

in the case of respondents from outside Dublin. However, we 

must take care in interpreting the substantive meaning of 

these coefficients. We start by focusing on the gross 

coefficients. A value of 54:1 is reported for rural local 

authority tenants. This implies that the odds of such tenants 

being below rather than above the MCP line is 54 times 

greater than is the case for private purchasers in Dublin. This 

rate is a good deal higher than for the comparison of rural LA 

tenants with urban private purchasers. However, the main 

reason for this higher effect is not the effect of tenure as such 

but the fact that the overall contrast between Dublin and 

elsewhere is sharper than in the case of the urban-rural 

contrast. 

To illustrate we will proceed to calculate the gross and net 

disadvantages of LA tenancy within rural areas. The gross 

effect has a value of 18.3 (54/2.95) and the corresponding net 

figure is 5.4. These figures are actually identical to those 

observed for the corresponding urban-rural comparisons. 

Thus, the contrast between the two sets of results is entirely 

a consequence of the fact that Dublin is more sharply 

differentiated from other areas than is the case for the urbanrural 

comparison. For both urban and rural local authority 

tenure, a reduction of 70 per cent is observed between the 

gross and the net coefficients. Taken together, location in 

Dublin versus elsewhere and tenure uniquely account for 5.5 

per cent of the explained variance and share a further 21.2 

135




at the 50 per cent MCP line by tenure and Dublin versus 

elsewhere 

per cent with socio-demographic controls, leaving almost 

three-quarters uniquely accounted for by the latter. 


Our analysis in this chapter shows that the impact of spatial 

variation per se, whether defined in terms of regional 

authority or local authority area, is extremely modest in 

comparison with that of the socio-demographic 

characteristics of householders. In this case we must stress 

that the most important finding relates not to differences 

between gross and net effects but to the modest level of both 

compared to the impact of household differences within areas 

associated with the socio-demographic characteristics of 

households. It is not that spatial effects are being explained 

by socio-demographic factors but that they are remarkably 

modest in comparison with the latter. 

136



In contrast to our spatial findings, we do observe substantial 

gross effects that are then very significantly reduced when we 

control for socio-demographic factors. This suggests that in 

large part, tenure/type of area effects are a consequence of 

the socio-demographic profiles of households within each of 

the combinations of tenure and location. We observe 

significant net effects for the combination of tenure and type 

of area. Although we cannot entirely rule out the possibility 

that such differences are a consequence of unmeasured 

variables for which we cannot control, taken at face value 

they point to the operation of additional disadvantages 

associated with being a local authority tenant in particular. 

The combined effect of size of location and housing tenure is 

clearly more substantial than that of location, although still 

modest in comparison with factors such as unemployment, 

class and household type. The strength of the former arises 

from two distinct but interacting influences. The first relates to 

the disadvantage experienced by rural respondents and more 

generally to those outside Dublin. The second relates to the 

impact of tenure, which is generally considerably stronger in 

urban rather than rural areas. The consequence of this is that 

if we take urban purchasers as the reference category, the 

rural groups and, in particular, LA tenants are most exposed 

to poverty risk. However, the level of disadvantage is less 

than if the impact of location and tenure were operating in a 

straightforward cumulative matter. However, if we make our 

comparison within the urban and rural sectors and thus focus 

on tenure per se, the level of disadvantage associated with 

types of tenure, and most particularly with being a local 

authority tenant, is significantly greater in urban areas. These 

conclusions also hold if the location variable on which we 

focus involves Dublin versus elsewhere rather than urban 

137



versus rural. Comparing our findings with those arising from 

earlier work, it would appear that there has been a significant 

reduction in the additional impact of tenure type in urban 

areas and in Dublin, although the differential remains 

substantial. 

138


Chapter 8 

Conclusions 


In the earlier chapters we have sought to document the 

spatial distribution of poverty and deprivation in Ireland. In 

doing so we have made use of a variety of sources. Each of 

these has its own strengths and weaknesses. Thus, census 

data enables us to achieve a much greater degree of 

geographical differentiation than the survey data, but it lacks 

direct measures of poverty and deprivation. Survey data 

contains such measures, but there are strict limits to the 

range of geographical variation that we can encompass in our 

analysis. By using a variety of sources we hope to provide a 

rounded picture of the extent and implication of spatial 

variation in poverty and deprivation. 

As we have noted, there is no one rationale for spatial 

interventions. The justifications for area-based strategies can 

range from relatively straightforward efforts at rationing of 

scarce resources to much more complex arguments relating 

to efficient delivery of services and mobilisation of community 

resources. However, in every case the issue of the extent of 

geographical concentration of poverty and deprivation is 

relevant. Our first effort to address this issue involved analysis 

of CSO data relating to administrative counties and county 

139



boroughs and focusing on indirect or surrogate measures of 

deprivation. 

8.2 Surrogate Measures of Deprivation 

In the absence of suitable income data from the census, one 

is obliged to use surrogate measures of deprivation such as 

age structure, economic status and activity, levels of farming 

and social class. In pursuing such analysis, a decision must 

be made whether to present results separately for a number 

of dimensions or to calculate a composite measure of 

multiple deprivation. In line with our earlier work and in 

agreement with the arguments made by Pringle et al. (2000), 

we have opted for the former strategy. There are a number of 

reasons underlying this choice, including the fact that there is 

no clear consensus on which indicators to include in such an 

index and the fact that crucial differences may be obscured 

by such aggregation. The most important factor underlying 

our decision, however, is the absence of any objective basis 

on which to assign weights to dimensions. While simply 

assigning equal weights to each dimension is obviously 

arbitrary, it emerges that apparently more sophisticated 

statistical approaches can be equally arbitrary. 

The most commonly used statistical technique for this 

purpose is factor analysis. This procedure can identify 

dimensions of socio-demographic characteristics that we 

have reason to expect to be associated with poverty and 

deprivation. However, as Nolan et al. (1998) observe, in the 

absence of direct measures of the latter, the weight derived 

for such an analysis can be informative only with regard to 

the relationship between the socio-demographic variables but 

not concerning the association between such variables and 

140


Conclusions 

deprivation and poverty. Thus, as Pringle et al. (2000: 11) 

conclude, it is questionable whether the fact that a variable 

happens to be correlated highly with another variable 

provides a sound basis for according it more importance. 

Pursuing this strategy will give disproportionate weight to 

dimensions that have been tapped using multiple but similar 

indicators. 

8.3 Overview of CSO Results 

In Chapter 2 we considered spatial variations in sociodemographic 

variables related to poverty and deprivation 

from the Census of Population aggregated to the level of 

local authority area. Throughout the chapter we saw that the 

Border and West regions contained counties with the highest 

percentages of population who were elderly, with lowest 

levels of educational attainment, highest incidence of small 

farming activity and high levels of economic dependency. At 

the other end of the spectrum, Dublin and the Mid-East 

emerged as the regions with the lowest levels of economic 

dependency and the highest levels of labour force 

participation, percentage of the population with third-level 

education and percentage of the labour force in the higher 

social classes. 

8.4 Geographical Variation: Evidence from Analysis 

of Indirect Measures of Deprivation 

Our analysis of geographical variation in direct measures of 

poverty and deprivation based on household data from the LII 

Surveys and the NSHQ were broadly consistent with the 

analysis based on census data. At the level of regional 

141



authority, the Border area clearly exhibits the highest levels of 

poverty and deprivation. The sharpest contrast is with Dublin 

and the Mid-East area. However, it should be kept clearly in 

mind that there is substantial variation between counties 

within regions. The boundaries between regions are drawn on 

the basis of political and pragmatic considerations and do not 

necessarily capture a confluence of similar causal processes. 

Further differentiation within regions reveals particularly higher 

levels of disadvantage in counties such as Donegal, Cavan, 

Leitrim, Longford and Mayo and urban centres other than 

Dublin. 

When we focus on dimensions of deprivation we find that 

regional disparities vary across the five dimensions of 

deprivation that we have identified. For the basic deprivation 

there is once again a continuum, with the Border region at 

one end and Dublin at the other. However, for other 

dimensions variation does not follow this familiar pattern. In 

the case of secondary deprivation, involving a range of 

consumer items, geographical location has little impact. 

Absence of basic housing facilities is a problem least 

frequently in Dublin, but physical deterioration is worst in 

Dublin, as is exposure to social problems. Once again, we 

observe very clear differentiation by county within regions. As 

one would expect from our summary thus far, the impact of 

urban-rural location is highly dependent on the particular 

dimension on which one focuses. Indeed, the only evidence 

for overlapping patterns of multiple disadvantage comes in 

the case of local authority renters, who fare worst on four of 

the five dimensions. Otherwise, geographical locations or 

types of areas or forms of tenure that fare worst on one 

dimension do not necessarily do so on the reminder. 

Evidence of multiple deprivation structured along spatial lines 

is extremely weak. 

142


Conclusions 

8.5 Variation by Size of Area and Tenure 

While there is clear variation in risk of poverty and deprivation 

by geographical location, such differentiation is modest in 

comparison with that associated with size of area and, more 

particularly, tenure. Overall, less populated areas are more 

disadvantaged while Dublin is most favoured. The picture at 

intermediate levels is somewhat mixed. However, although 

variation by size of area is larger than that associated with 

geographical location per se, it is extremely modest in 

comparison with the impact of housing tenure. Local authority 

tenants are a particularly disadvantaged group. Over time 

their income poverty rates have increased sharply, and while 

their consistent poverty rate has shown some decline, their 

relative positions have shown further deterioration. The 

decline in their numbers, however, means that they constitute 

a smaller proportion of the poor than heretofore. 

8.6 Policy Implications 

In order to draw sensible policy conclusions from the results 

of our analysis it is necessary to go beyond descriptions of 

differences by location, type of area or tenure and address 

the issue of the underlying processes. In order to do so, in 

Chapter 7 we looked not only at the gross or observed 

effects of geographical location and the combination of size 

of area and location but also at the net effect after controlling 

for a range of socio-demographic household characteristics. 

In doing so we sought to address issues relating to the 

distinction made by such authors as Pringle and Walsh (1999) 

between geographical causal processes that increase the 

relative risks of people living in one area experiencing poverty 

or deprivation compared with individuals living in other areas. 

143



In other words, the risk of poverty is a function of 

geographical distribution processes that influence the 

likelihood of individuals who experience poverty or 

deprivation being located or concentrated in particular areas, 

but these geographical processes do not directly influence 

the likelihood of an individual experiencing poverty or 

deprivation. Such processes influence only the spatial 

distribution of poverty – they do not have a major influence 

on the total number of poor. 

In relation to the impact of geographical location, we found 

that such effects were not in any sense statistically spurious. 

Even after controlling for a variety of household factors, the 

magnitude of such effects remains largely unaffected. The 

crucial point that needs to be made in relation to 

geographical effects is not that the differences we observe 

are in any sense misleading, but that they are extremely 

modest when placed in the context of overall variation in risk 

between households. Furthermore, the regional and local 

authority spatial units that we have employed in our analysis 

do not in any way constitute homogeneous blocks in relation 

to risk of poverty and deprivation. Finally, as we have 

observed, different dimensions of deprivation have rather 

different spatial distributions. 

It is of course possible to conduct a more fine-grained 

analysis of spatial location than it was possible for us to 

undertake on this occasion by going to the level of a district 

electoral division (DED). However, while such an analysis 

would inevitably reveal a more marked pattern of spatial 

differentiation, the regional and local authority patterns that 

we have documented will also break up, yielding a more 

scattered overall picture of spatial differentiation. Thus, using 

1996 SAPS data, Nolan and Fahey (2002: 240–41) concluded 

144


Conclusions 

that the national picture is something akin to a partially 

patterned mosaic: the DEDs with the highest unemployment 

rates are far worse off than those with the lowest 

unemployment rates and there is some tendency towards a 

clustering of the worst with the worst and the best with the 

best. However, the clustering effect is relatively modest when 

placed in the context of overall variation, so that the effect is 

not of sharply delineated black spots in which the bulk of the 

poor or unemployed are concentrated. The most reasonable 

conclusion remains that while some geographical 

concentration of disadvantage exists, poverty and deprivation 

are spatially pervasive and affect almost all parts of the 

country at all levels of geographical disaggregation. Our 

current analysis confirms earlier findings (Nolan, Whelan and 

Williams 1994, 1999; Fahey and Williams 2000) that from a 

pure targeting perspective, a focus on geographical location 

offers the crudest basis for reaching ‘at-risk populations’. As 

Nolan et al. (1999) concluded, area-based strategies cannot 

be the panacea for spatially pervasive problems. 

When we focus on size of area and type of tenure, a 

somewhat different picture emerges. While the effect of the 

former per se is modest, in combination they do produce a 

level of differentiation in terms of poverty and deprivation that 

far exceeds anything arising from geographical location. 

Furthermore, although the magnitude of such effects is still 

modest in comparison with those associated with the sociodemographic 

characteristics of households, unlike those 

relating to geographical location they remain highly significant 

even after we control for socio-demographic characteristics. 

To a very large extent the relative disadvantages suffered by, 

for example, local authority tenants are a consequence of the 

socio-demographic composition of such households. It is 

145



necessary to caution against assuming that such residual 

effects can be entirely attributed to the causal effect of the 

combination of type of location and form of tenure. It is 

impossible to rule out the possibility that such effects arise as 

a consequence of differences relating to unmeasured 

household characteristics for which we have been unable to 

control. However, the significant net effects that we have 

observed do leave open the possibility that such households 

face significant additional disadvantages associated with their 

tenure that cannot be accounted for by selection on the basis 

of socio-demographic characteristics associated with poverty 

and deprivation. 

As we noted in our introduction, such effects could arise as a 

consequence of discriminatory behaviour, a low level of 

community resources or as a consequence of cumulative 

processes of disadvantage that undermine the coping 

capacities of households. However, the methodological 

problems that confront one in attempting to establish 

neighbourhood effects are formidable. Even in the United 

States, where we would expect such effects to be a good 

deal more substantial and where data sources and 

investment in research on the topic go beyond anything we 

can imagine in the Irish situation, it is necessary to adopt a 

cautious stance. Thus, Jencks and Mayer (1990), in a review 

of the US neighbourhood literature, were ultimately 

pessimistic because few studies succeeded in isolating and 

measuring the social processes or mechanisms that could 

count for such effects. While a great deal of subsequent work 

in the US has attempted to rectify these deficiencies and 

progress has been made, Sampson et al. (2002: 473–4) 

conclude that we continue to know relatively little about key 

processes or whether they are responsive to neighbourhood 

146


Conclusions 

policy interventions. Conclusions regarding such effects must 

be qualified by a range of caveats relating to intervening 

mechanisms, identification of appropriate spatial units and 

the role of selection effects (Jargowsky 1996; Brooks-Gunn 

et al. 1997). 

Efforts by Nolan and Whelan (2000) and Layte, Nolan and 

Whelan (2000) to examine the possibility of the emergence of 

vicious cycles of deprivation in urban areas exposed to 

patterns of multiple disadvantage have also resulted in 

cautious conclusions regarding the independent effect of 

neighbourhood or community. This stance is in line with that 

arising out of a review of the evidence, both US and 

European, by Friedrichs (1999). In evaluating the available 

European evidence, Friedrichs (1999) cautions that it is 

necessary to take into account that many of the most 

frequently cited sources are descriptive and small-scale 

studies which are highly selective since only neighbourhoods 

with high poverty rates are sampled. His overall conclusion is 

as follows: ‘The general evidence presented on 

neighbourhood effects indicates low or negligible effects; 

most context effects can be explained by either individual or 

institutional effects’ (Friedrichs 1998: 93). Fahey and Williams 

(2000: 241) suggest that if neighbourhoods matter, it is 

perhaps in a more complex and fine-grained way than can be 

captured by statistical analysis. They suggest an 

understanding of neighbourhood that views deprived areas 

not as large uniform social environments but as complex 

composites of micro areas, each micro area having its own 

character and neighbourhood quality. 

Effects associated with both geographical location and the 

combination of size of area and form of tenure are modest in 

comparison with those associated with the 

147



socio-demographic characteristics. Poverty and deprivation 

are spatially pervasive and it is extremely difficult to justify a 

strategy of spatial intervention on the basis of straightforward 

targeting arguments. It is difficult to avoid the conclusion that 

in discussions of poverty in Ireland, disproportionate attention 

has been focused on spatial and community issues in 

comparison with that devoted to vulnerable social groups. 

However, as we stressed in our introduction, it is possible to 

disentangle arguments for area-based intervention from a 

focus on rationing or targeting. 

Most recent interventions have relied on more complex 

justifications that have encompassed factors such as 

improved service delivery and mobilisation of community 

resources involving a somewhat broader quality of life 

perspective. Thus, as Pringle and Walsh (1999: 339) observe, 

policies aimed at promoting social inclusion involve more 

than simply providing resources and have the objective of 

creating mechanisms whereby the disadvantaged can 

participate more actively in all aspects of life, including 

decision making. 

It goes well beyond the brief of this report to offer an 

evaluation of the impact of such more complex interventions. 

Any assessment of such initiatives would need to be made in 

light of evaluations designed specifically to address the 

stated objectives of the initiatives. However, the evidence we 

have presented relating to the limited role of spatial factors in 

comparison with the socio-demographic composition of 

households would lead us to warn against developing 

unrealistic expectations of area-based initiatives. This is 

particularly true in circumstances where, as Pringle and Walsh 

(1999: 343) argue, area-based funding has been extremely 

modest in comparison to what one would expect to be 

148


Conclusions 

required to have a major impact. It is necessary to strike a 

balance between an unduly deterministic perspective that 

views area-based initiatives as serving no purpose until 

fundamental structural problems are resolved and a 

perspective that encourages unrealistic expectations of the 

extent to which disadvantaged communities can find 

solutions to problems, many of which can be addressed only 

by national policies and mobilisation of substantial resources. 

149

Chapter 08 7/7/05 6:58 am Page 150

xAppendix 1 7/7/05 6:59 am Page 151 

Appendix 1 

Additional Tables 

151



A1.1 Additional Tables from Census 2002 

Table A1.1: Percentage of persons in each county according to age cohort 

County % 0–14 % 15–29 % 30–44 % 45–64 % 65+ % age % age 65+ 

dependent live alone 

% Rank % Rank % Rank % Rank % Rank % Rank % Rank 

Cavan 22.8 7 21.1 26 20.7 25 21.5 19 13.8 4 36.6 3 4.1 3 

Donegal 23.4 3 21.6 22 20.5 29 21.8 17 12.6 12 36.0 6 3.5 8 

Leitrim 21.0 25 18.7 34 20.1 32 24.1 1 16.1 1 37.1 1 5.1 1 

Louth 22.7 9 24.0 12 22.4 8 20.5 27 10.3 28 33.0 21 2.9 21 

Monaghan 22.4 16 23.2 14 20.8 24 21.4 20 12.3 15 34.7 10 3.5 8 

Sligo 20.8 27 22.5 16 20.7 25 22.8 7 13.2 8 34.0 16 3.7 6 

Border 22.5 3 22.3 8 21.0 7 21.7 5 12.5 2 35.0 1 3.5 1 

Laois 23.2 4 22.4 17 22.5 7 20.6 23 11.3 22 34.5 13 2.9 21 

Longford 22.5 12 20.7 29 20.1 32 22.9 6 13.7 5 36.2 4 4.0 5 

Offaly 23.2 4 22.4 17 21.7 15 21.0 21 11.6 21 34.9 9 3.0 19 

Westmeath 22.9 6 23.4 13 22.2 12 20.6 23 11.0 24 33.9 18 3.0 19 

Midlands 23.0 2 22.5 6 21.8 4 21.0 6 11.6 5 34.6 2 3.1 3 

Galway County Borough 16.3 33 37.3 1 20.7 25 17.6 34 8.2 31 24.5 34 1.9 31 

Galway County 22.8 7 20.3 31 21.6 18 22.5 10 12.8 10 35.6 7 3.2 15 

Mayo 21.5 23 20.2 32 20.4 31 23.2 4 14.7 3 36.2 4 4.1 3 

Roscommon 21.2 24 19.2 33 20.6 28 23.5 3 15.5 2 36.7 2 4.4 2 

West 21.1 6 23.0 5 20.9 8 22.0 1 12.9 1 34.0 3 3.4 2 

Dublin County Borough 16.2 34 29.5 2 22.3 9 19.2 33 12.8 10 29.0 31 3.7 6 

Fingal 22.7 9 26.6 7 24.5 2 20.2 29 5.9 34 28.6 33 1.3 33 

South Dublin 22.5 12 27.8 5 22.9 6 20.6 23 6.3 33 28.7 32 1.3 33 

Dún Laoghaire-Rathdown 19.2 31 24.1 11 21.7 15 22.6 9 12.4 14 31.6 25 3.1 17 

Dublin 19.2 8 27.7 1 22.7 2 20.2 8 10.2 7 29.3 8 2.6 7 

Kildare 23.7 1 24.9 9 25.0 1 19.7 32 6.7 32 30.4 29 1.5 32 

152


Appendix 1 

Table A1.1: Percentage of persons in each county according to age cohort (cont’d) 

County % 0–14 % 15–29 % 30–44 % 45–64 % 65+ % age % age 65+ 

dependent live alone 

% Rank % Rank % Rank % Rank % Rank % Rank % Rank 

Meath 23.6 2 23.0 15 24.1 3 20.5 27 8.7 30 32.3 23 2 30 

Wicklow 22.5 12 22.4 17 23.2 4 21.9 16 10.0 29 32.5 22 2.4 29 

Mid-East 23.4 1 23.6 3 24.2 1 20.6 7 8.2 8 31.6 7 1.9 8 

Clare 22.1 19 21.1 26 22.3 9 22.7 8 11.8 17 34.0 16 3.1 17 

Limerick County Borough 19.4 30 28.7 4 20.5 29 19.8 31 11.7 18 31.1 28 3.2 15 

Limerick County 20.9 26 25.3 8 21.3 21 22.0 13 10.5 26 31.4 27 2.7 27 

Tipperary N.R. 21.8 22 21.2 25 21.1 22 22.5 10 13.4 7 35.2 8 3.4 13 

Mid-West 21.2 5 23.8 2 21.4 6 21.9 3 11.6 5 32.8 5 3 5 

Carlow 21.9 21 24.9 9 22.3 9 20.6 23 10.4 27 32.2 24 2.6 28 

Kilkenny 22.5 12 21.8 21 22.0 13 22.0 13 11.7 18 34.2 15 2.9 21 

Tipperary S.R. 22.0 20 21.5 24 21.5 20 22.4 12 12.6 12 34.6 11 3.5 8 

Waterford County Borough 20.2 28 26.7 6 21.6 18 20.2 29 11.3 22 31.5 26 2.8 25 

Waterford County 22.3 18 20.7 29 21.7 15 23.0 5 12.3 15 34.6 11 3.3 14 

Wexford 22.7 11 21.6 22 22.0 13 22.0 13 11.7 18 34.4 14 2.9 21 

South-East 22.1 4 22.4 7 21.8 4 21.9 3 11.8 4 33.9 4 3 5 

Cork County Borough 17.2 32 29.2 3 19.8 34 21.0 21 12.9 9 30.0 30 3.5 8 

Cork County 22.4 16 22.0 20 23.1 5 21.6 18 10.9 25 33.3 20 2.8 25 

Kerry 20.2 28 21.1 26 21.1 22 23.9 2 13.7 5 33.9 18 3.5 8 

South-West 20.8 7.0 23.3 4.0 21.9 3.0 22.0 1.0 12.0 3.0 32.7 6 3.1 3 

Total 21.1 24.4 22.1 21.2 11.1 32.3 2.9 


153



Table A1.2: County-level labour force participation and 

unemployment rates 

County Labour force Unemployment 

participation rate 

rate 

Rate Rank Rate Rank 

Cavan 55.8 26 7.9 24 

Donegal 54.4 30 15.6 1 

Leitrim 53.7 33 8.7 17 

Louth 58.2 9 13.2 3 

Monaghan 57.4 14 9.9 11 

Sligo 56.8 17 8.7 17 

Border 56.1 6 12.0 1 

Laois 58.3 8 9.8 12 

Longford 54.9 29 10.1 9 

Offaly 57.4 14 8.8 16 

Westmeath 57.6 11 8.5 21 

Midlands 57.4 4 9.1 3 

Galway County Borough 56.9 16 10.0 10 

Galway County 56.5 20 8.7 17 

Mayo 54.4 30 10.7 6 

Roscommon 54.0 32 7.0 28 

West 55.6 8 5.2 7 

Dublin County Borough 60.2 5 10.4 8 

Fingal 64.8 1 6.9 29 

South Dublin 64.6 2 8.0 23 

Dún Laoghaire-Rathdown 56.0 25 5.7 34 

Dublin 61.2 3 5.2 7 

Kildare 63.4 3 6.2 33 

Meath 61.7 4 6.6 30 

Wicklow 58.5 7 8.1 22 

Mid-East 61.5 1 6.8 6 

Clare 58.8 6 7.6 25 

Limerick County Borough 55.1 28 13.9 2 

Limerick County 56.5 20 6.6 30 

Tipperary N.R. 56.7 18 7.6 25 

Mid-West 61.5 1 8.2 4 

Carlow 56.4 22 9.5 13 

Kilkenny 57.5 13 7.6 25 

Tipperary S.R. 56.1 23 9.5 13 

Waterford County Borough 57.6 11 12.1 5 

Waterford County 56.6 19 8.7 17 

Wexford 56.1 23 10.5 7 

South-East 56.6 5 9.6 2 

Cork County Borough 52.3 34 12.4 4 

Cork County 57.8 10 6.4 32 

Kerry 55.3 27 9.0 15 

South-West 56.0 7 8.2 4 

Total 58.3 8.8 


154


Appendix 1 

Table A1.3: Distribution of the unemployed across regions 

County % of the % of those Ratio of 

unemployed 15 + in unemployment to 

labour force labour force 

Col. % Col. % Ratio Rank 

Cavan 1.2 1.4 0.90 23 

Donegal 5.6 3.2 1.76 1 

Leitrim 0.6 0.6 0.98 19 

Louth 3.8 2.5 1.50 3 

Monaghan 1.5 1.3 1.12 11 

Sligo 1.4 1.5 0.99 16 

Border 14.1 10.4 1.35 1 

Laois 1.6 1.5 1.11 12 

Longford 0.8 0.7 1.14 9 

Offaly 1.5 1.6 0.99 16 

Westmeath 1.7 1.8 0.96 21 

Midlands 5.7 5.5 1.03 4 

Galway County Borough 2.0 1.7 1.13 10 

Galway County 3.4 3.5 0.99 16 

Mayo 3.4 2.8 1.21 6 

Roscommon 1.0 1.3 0.79 28 

West 9.8 9.3 1.05 3 

Dublin County Borough 16.3 13.9 1.18 7 

Fingal 4.2 5.5 0.77 29 

South Dublin 6.0 6.6 0.90 23 

Dún Laoghaire-Rathdown 3.1 4.8 0.65 34 

Dublin 29.7 30.8 0.96 5 

Kildare 3.1 4.4 0.70 33 

Meath 2.6 3.5 0.74 31 

Wicklow 2.7 2.9 0.92 22 

Mid-East 8.3 10.8 0.77 8 

Clare 2.3 2.6 0.86 25 

Limerick County Borough 2.1 1.3 1.57 2 

Limerick County 2.3 3.0 0.75 30 

Tipperary N.R. 1.3 1.5 0.85 25 

Mid-West 7.9 8.5 0.93 6 

Carlow 1.2 1.1 1.08 13 

Kilkenny 1.7 2.0 0.86 25 

Tipperary S.R. 2.1 1.9 1.07 14 

Waterford County Borough 1.6 1.1 1.37 5 

Waterford County 1.4 1.4 0.98 19 

Wexford 3.3 2.8 1.18 7 

South-East 11.2 10.4 1.08 2 

Cork County Borough 4.2 3.0 1.40 4 

Cork County 5.9 8.1 0.72 32 

Kerry 3.3 3.3 1.01 15 

South-West 13.3 14.3 0.93 6 

Total 100.0 100.0 


155



Table A1.4: Percentage of all persons engaged in agriculture 

classified by number of acres farmed 

County % of persons Farmers Farmers Farmers 50+ 

15 years+


Appendix 1 

Table A1.5: Index of economic dependency in each county 

County 

Economic dependency 

Index 

Rank 

Cavan 1.52 10 

Donegal 1.84 1 

Leitrim 1.58 6 

Louth 1.56 8 

Monaghan 1.49 14 

Sligo 1.43 22 

Border 1.61 1 

Laois 1.48 18 

Longford 1.61 5 

Offaly 1.49 14 

Westmeath 1.46 20 

Midlands 1.49 4 

Galway County Borough 1.34 28 

Galway County 1.51 12 

Mayo 1.62 3 

Roscommon 1.53 9 

West 1.51 2 

Dublin County Borough 1.21 31 

Fingal 1.14 34 

South Dublin 1.17 33 

Dún Laoghaire-Rathdown 1.34 28 

Dublin 1.21 8 

Kildare 1.21 31 

Meath 1.27 30 

Wicklow 1.40 24 

Mid-East 1.28 7 

Clare 1.36 27 

Limerick County Borough 1.62 3 

Limerick County 1.39 25 

Tipperary N.R. 1.44 21 

Mid-West 1.43 6 

Carlow 1.51 12 

Kilkenny 1.43 22 

Tipperary S.R. 1.52 10 

Waterford County Borough 1.48 18 

Waterford County 1.49 14 

Wexford 1.58 6 

South-East 1.51 2 

Cork County Borough 1.64 2 

Cork County 1.38 26 

Kerry 1.49 14 

South-West 1.46 5 

Total 1.39 


157



Table A1.6: Distributions of persons 15 years and over 

classified by highest level of educational attainment 

County None/primary Lower sec. Upper sec. Third level 

% Rank % Rank % Rank % Rank 

Cavan 31.3 2 24.7 13 25.6 31 18.4 26 

Donegal 33.7 1 25.4 11 22.5 34 18.3 28 

Leitrim 28.7 6 24.3 18 27.6 27 19.3 24 

Louth 25.4 12 25.8 9 27.4 29 21.4 21 

Monaghan 29.2 5 28.2 1 25.0 33 17.6 33 

Sligo 23.9 16 21.9 29 29.6 17 24.5 12 

Border 29.3 1 25.2 3 25.7 8 19.9 6 

Laois 25.4 12 26.2 8 30.2 13 18.3 28 

Longford 29.5 4 24.4 14 28.3 26 17.8 31 

Offaly 26.1 10 26.4 6 29.8 16 17.7 32 

Westmeath 23.4 19 24.2 20 30.3 11 22.1 16 

Midlands 25.5 3 25.4 2 29.9 3 19.3 8 

Galway County Borough 13.5 33 15.5 33 30.3 11 40.8 2 

Galway County 27.4 8 21.3 30 28.8 21 22.5 14 

Mayo 29.6 3 22.5 26 28.7 23 19.1 25 

Roscommon 27.7 7 24.4 14 29.6 17 18.3 28 

West 25.9 2 21.2 7 29.1 6 23.7 4 

Dublin County Borough 23.6 18 19.2 32 25.1 32 32.1 4 

Fingal 13.6 32 19.8 31 33.5 1 33.1 3 

South Dublin 18.0 30 22.7 25 32.0 3 27.3 7 

Dún Laoghaire-Rathdown 11.7 34 14.3 34 29.0 19 45.0 1 

Dublin 18.7 7 19.2 8 28.7 7 33.5 1 

Kildare 17.1 31 22.2 28 32.3 2 28.4 5 

Meath 19.1 29 24.3 18 31.3 5 25.3 10 

Wicklow 19.7 27 22.9 24 29.9 15 27.5 6 

Mid-East 18.5 8 23.1 6 31.3 1 27.2 2 

Clare 21.4 25 22.4 27 31.7 4 24.4 13 

Limerick County Borough 26.0 11 24.4 14 27.6 27 22.1 16 

Limerick County 20.7 26 23.7 22 30.6 7 25.0 11 

Tipperary N.R. 23.3 20 25.8 9 31.3 5 19.5 23 

Mid-West 22.3 5 23.8 4 30.6 2 23.3 5 

Carlow 24.1 15 26.8 4 28.8 21 20.3 22 

Kilkenny 22.0 24 26.3 7 30.2 13 21.5 20 

Tipperary S.R. 23.7 17 27.3 3 30.6 7 18.4 26 

Waterford County Borough 22.6 22 26.7 5 29.0 19 21.7 19 

Waterford County 22.2 23 25.1 12 30.5 10 22.1 16 

Wexford 26.7 9 27.4 2 28.4 25 17.5 34 

South-East 24.0 4 26.7 1 29.5 4 19.8 7 

Cork County Borough 22.8 21 24.1 21 27.2 30 25.9 9 

Cork County 19.3 28 23.5 23 30.6 7 26.7 8 

Kerry 24.8 14 24.4 14 28.5 24 22.3 15 

South-West 21.3 6.0 23.8 4.0 29.4 5.0 25.5 3.0 

Total 22.2 22.7 29.1 26.0 


158


Appendix 1 

Table A1.7: Distribution of persons whose education has 

ceased in highest and lowest levels of educational attainment 

across counties 

County A B C D E 

% none/ % third % pop. Ratio Ratio 

primary level left none/primary third level 

education to total to total 

left 

left 

education education 

% % % Ratio Rank Ratio Rank 

Cavan 2.0 1.0 1.4 1.42 2 0.71 27 

Donegal 5.3 2.5 3.5 1.53 1 0.71 27 

Leitrim 0.9 0.5 0.7 1.29 6 0.74 25 

Louth 3.0 2.2 2.6 1.16 11 0.84 19 

Monaghan 1.8 0.9 1.3 1.32 4 0.68 32 

Sligo 1.6 1.4 1.5 1.07 16 0.94 12 

Border 14.6 8.4 11.0 1.33 1 0.77 6 

Laois 1.7 1.1 1.5 1.16 11 0.71 27 

Longford 1.0 0.5 0.8 1.30 5 0.67 33 

Offaly 1.9 1.1 1.6 1.16 11 0.67 33 

Westmeath 1.9 1.5 1.8 1.06 18 0.86 15 

Midlands 6.5 4.2 5.7 1.15 3 0.74 8 

Galway County Borough 0.9 2.4 1.6 0.60 33 1.55 2 

Galway County 4.5 3.2 3.6 1.24 8 0.87 14 

Mayo 4.1 2.3 3.0 1.35 3 0.75 24 

Roscommon 1.8 1.0 1.4 1.27 7 0.72 26 

West 11.3 8.8 9.6 1.18 2 0.92 4 

Dublin County Borough 13.6 15.7 13.3 1.02 20 1.18 4 

Fingal 3.0 6.2 4.9 0.62 32 1.28 3 

South Dublin 4.8 6.3 6.0 0.81 30 1.05 7 

Dún Laoghaire-Rathdown 2.6 8.6 4.9 0.53 34 1.76 1 

Dublin 24.0 36.8 29.0 0.83 8 1.27 1 

Kildare 3.1 4.5 4.0 0.78 31 1.10 5 

Meath 2.9 3.3 3.4 0.87 29 0.99 9 

Wicklow 2.6 3.1 2.9 0.89 27 1.06 6 

Mid-East 8.7 10.9 10.3 0.84 7 1.05 2 

Clare 2.6 2.5 2.7 0.97 25 0.94 12 

Limerick County Borough 1.6 1.2 1.4 1.19 10 0.86 15 

Limerick County 2.8 2.9 3.0 0.95 26 0.98 11 

Tipperary N.R. 1.7 1.2 1.6 1.06 18 0.76 23 

Mid-West 8.7 7.8 8.6 1.01 5 0.91 5 

Carlow 1.3 0.9 1.2 1.10 14 0.79 22 

Kilkenny 2.1 1.7 2.1 1.01 23 0.84 19 

Tipperary S.R. 2.2 1.4 2.0 1.07 16 0.71 27 

Waterford County Borough 1.1 0.9 1.1 1.00 24 0.82 21 

Waterford County 1.5 1.3 1.5 1.02 20 0.86 15 

Wexford 3.7 2.1 3.0 1.23 9 0.69 31 

South-East 11.8 8.3 10.9 1.09 4 0.77 6 

Cork County Borough 3.2 3.1 3.2 1.02 20 0.99 9 

Cork County 7.2 8.6 8.2 0.88 28 1.04 8 

Kerry 3.9 3.0 3.5 1.10 14 0.85 18 

South-West 14.3 14.7 14.9 0.96 6 0.99 3 

Total 100.0 100.0 100.0 


159



Table A1.8: Percentage of persons in each social class 

County Professional/ Non- Skilled Semi- Unskilled 

managerial/ manual Manual skilled 

technical 

% Rank % Rank % Rank % Rank % Rank 

Cavan 30.6 32 21.5 5 25.1 3 15.1 8 7.7 17 

Donegal 30.1 34 18.4 33 24.3 4 16.9 3 10.3 1 

Leitrim 33.0 25 21.1 7 23.4 9 14.9 12 7.5 18 

Louth 33.0 25 19.6 20 23.2 10 16.1 4 8.1 13 

Monaghan 30.9 31 19.4 22 26.8 1 14.9 12 8.0 15 

Sligo 38.0 12 21.1 7 20.2 29 14.5 17 6.2 28 

Border 32.2 8 19.7 6 23.9 1 15.8 1 8.4 2 

Laois 33.0 25 20.1 15 22.8 13 15.6 7 8.4 10 

Longford 33.2 23 20.3 10 24.1 6 13.9 19 8.5 8 

Offaly 32.0 30 18.7 31 25.3 2 15.0 10 8.9 5 

Westmeath 36.5 17 21.6 4 21.0 25 13.4 21 7.5 18 

Midlands 33.9 7 20.2 2 23.1 2 14.5 3 8.3 3 

Galway County Borough 46.6 3 20.1 15 15.5 33 13.1 25 4.7 31 

Galway County 36.6 15 19.2 25 23.1 11 14.1 18 7.1 22 

Mayo 32.4 29 19.4 22 23.6 8 16.1 4 8.5 8 

Roscommon 34.5 18 22.6 1 22.8 13 13.0 27 7.0 23 

West 36.5 5 19.9 4 22.0 4 14.4 4 7.1 5 

Dublin County Borough 39.2 10 21.3 6 19.2 31 13.1 25 7.2 21 

Fingal 46.9 2 21.9 3 17.5 32 9.7 33 4.0 33 

South Dublin 39.3 9 22.5 2 22.0 19 11.7 31 4.5 32 

Dún Laoghaire-Rathdown 60.2 1 18.9 28 11.6 34 6.5 34 2.8 34 

Dublin 44.4 1 21.3 1 18.1 8 11.0 8 5.2 8 

Kildare 41.3 5 20.3 10 20.3 28 11.7 31 6.3 27 

Meath 39.5 7 19.2 25 22.8 13 11.8 30 6.7 25 

Wicklow 42.2 4 18.9 28 20.1 30 12.1 29 6.7 25 

Mid-East 41.0 2 19.5 7 21.1 6 11.8 7 6.6 7 

Clare 38.9 11 20.2 13 21.4 22 13.4 21 6.1 29 

Limerick County Borough 30.5 33 20.3 10 21.5 21 18.9 1 8.8 6 

Limerick County 39.4 8 19.7 19 20.4 27 13.5 20 7.0 23 

Tipperary N.R. 36.6 15 21.0 9 21.3 23 13.2 24 7.9 16 

Mid-West 37.4 4 20.2 2 21.0 7 14.2 5 7.2 4 

Carlow 33.1 24 18.6 32 24.3 4 14.8 15 9.2 4 

Kilkenny 37.7 13 19.2 25 23.1 11 12.5 28 7.5 18 

Tipperary S.R. 33.3 21 19.8 18 22.2 17 15.1 8 9.7 2 

Waterford County Borough 33.3 21 18.8 30 22.3 16 17.2 2 8.4 10 

Waterford County 37.5 14 17.5 34 21.7 20 15.0 10 8.3 12 

Wexford 32.5 28 19.3 24 24.1 6 14.7 16 9.3 3 

South-East 34.5 6 19.0 8 23.1 2 14.7 2 8.8 1 

Cork County Borough 34.1 20 19.9 17 21.3 23 16.0 6 8.7 7 

Cork County 40.3 6 19.6 20 20.8 26 13.4 21 5.9 30 

Kerry 34.5 18 20.2 13 22.2 17 14.9 12 8.1 13 

South-West 37.8 3.0 19.8 5.0 21.2 5.0 14.2 5.0 6.9 6.0 

Total 38.6 20.2 21.0 13.3 6.9 


160


Appendix 1 

Table A1.9: Distributions of persons in Professional and 

Unskilled Manual class categories across region 

County A B C D E 

Professional/ Unskilled % total Ratio Ratio 

Managerial/ Manual population Professional Unskilled 

Technical to total Manual 

to total 

Col. % Col. % Col. % Ratio Rank Ratio Rank 

Cavan 1.2 1.6 1.4 0.81 28 1.13 17 

Donegal 2.7 5.2 3.5 0.78 33 1.49 1 

Leitrim 0.6 0.7 0.7 0.84 24 1.07 20 

Louth 2.3 3.1 2.6 0.87 21 1.19 10 

Monaghan 1.1 1.6 1.3 0.82 27 1.19 10 

Sligo 1.4 1.3 1.5 0.96 13 0.88 30 

Border 9.2 13.5 11.0 0.84 8 1.23 2 

Laois 1.3 1.9 1.5 0.88 20 1.26 6 

Longford 0.6 0.9 0.8 0.81 28 1.16 14 

Offaly 1.3 2.1 1.6 0.81 28 1.27 5 

Westmeath 1.7 2.0 1.8 0.94 16 1.08 19 

Midlands 5.0 6.9 5.8 0.87 7 1.19 3 

Galway County Borough 1.8 1.0 1.7 1.05 7 0.59 33 

Galway County 3.5 3.7 3.7 0.95 15 1.02 24 

Mayo 2.5 3.7 3.0 0.84 24 1.24 8 

Roscommon 1.2 1.4 1.4 0.91 18 1.03 21 

West 9.0 9.9 9.7 0.93 5 1.02 5 

Dublin County Borough 11.7 12.0 12.7 0.93 17 0.95 26 

Fingal 6.4 3.0 5.0 1.27 2 0.60 32 

South Dublin 6.3 4.1 6.1 1.04 8 0.67 31 

Dún Laoghaire-Rathdown 7.9 2.1 4.9 1.62 1 0.43 34 

Dublin 32.4 21.2 28.7 1.13 1 0.74 8 

Kildare 4.6 4.0 4.2 1.11 4 0.95 26 

Meath 3.7 3.5 3.4 1.08 6 1.03 21 

Wicklow 3.3 2.9 2.9 1.12 3 0.99 25 

Mid-East 11.6 10.4 10.5 1.10 2 0.99 7 

Clare 2.7 2.4 2.6 1.03 9 0.90 28 

Limerick County Borough 1.0 1.7 1.4 0.74 34 1.21 9 

Limerick County 3.2 3.2 3.1 1.03 9 1.03 21 

Tipperary N.R. 1.5 1.8 1.6 0.96 13 1.17 12 

Mid-West 8.4 9.1 8.7 0.97 4 1.04 4 

Carlow 1.0 1.5 1.2 0.84 24 1.31 4 

Kilkenny 2.1 2.3 2.1 1.00 12 1.13 17 

Tipperary S.R. 1.7 2.9 2.0 0.86 23 1.41 2 

Waterford County Borough 0.9 1.3 1.1 0.81 28 1.14 16 

Waterford County 1.5 1.8 1.5 1.01 11 1.25 7 

Wexford 2.6 4.2 3.0 0.87 21 1.40 3 

South-East 9.8 14.0 10.8 0.90 6 1.29 1 

Cork County Borough 2.6 3.7 3.1 0.81 28 1.16 14 

Cork County 9.1 7.4 8.3 1.09 5 0.90 28 

Kerry 3.0 4.0 3.4 0.89 19 1.17 12 

South-West 14.6 15.0 14.8 0.99 3 1.02 5 

Total 100.0 100.0 100.0 


161



A1.2 Risk of Poverty by Planning Region from the LII 

Survey 

In Chapter 4 we reported poverty and deprivation results for 

regional authorities and local authorities. Here, drawing on the 

series of Living in Ireland Surveys, we report trends over time 

for planning regions, in which counties are grouped as 

follows. 

• East: Dublin, Kildare, Meath, Wicklow 

• South-West: Cork, Kerry 

• South-East: Carlow, Kilkenny, South Tipperary, Waterford, 

Wexford 

• North-East: Cavan, Louth, Monaghan 

• Mid-West: Clare, Limerick, North Tipperary 

• Midlands: Laois, Longford, Offaly, Roscommon, 

Westmeath 

• West: Galway, Mayo 

• North-West: Leitrim, Sligo 

• Donegal: Donegal. 

Risk of Poverty by Planning Region 

In Table A1.10 we show risk of poverty by planning region 

for 50 per cent and 60 per cent of mean income and for 

the consistent poverty line. At the 50 per cent line, with the 

exception of the East and the Mid-West, and to a lesser 

extent the South-East, there has been a sharp increase 

in the level of poverty between 1994 and 2000. The lowest 

rate by far of 18 per cent is observed in the East, while the 

highest rate of 37 per cent is observed in Donegal and the 

North-West. The range of regional variation is modest, with 

162


Appendix 1 

Table A1.10: Risk of poverty by planning region, 1987, 1994 and 2000 

Planning Region 50% income 60% income 60% line + 

line risk line risk deprivation risk 

1987 1994 2000 1987 1994 2000 1987 1994 2000 

% % % 

East 9.6 16.0 18.1 19.3 29.1 22.2 13.6 13.6 3.8 

South-West 17.7 17.9 32.4 31.0 34.4 40.0 16.3 13.0 7.5 

South-East 20.8 23.0 26.9 33.3 40.0 35.4 16.9 18.3 7.7 

North-East 24.6 15.9 34.1 35.7 38.2 46.2 20.6 12.8 11.6 

Mid-West 20.5 22.0 25.9 32.9 39.2 35.2 19.7 13.4 10.4 

Midlands 21.7 21.4 31.4 41.1 39.7 40.4 19.7 13.1 3.8 

West 19.6 20.2 30.2 31.0 35.5 39.0 13.8 7.3 5.6 

North-West & Donegal 27.3 24.7 36.7 42.6 43.5 46.5 22.9 23.5 8.9 

Total 16.9 18.9 25.8 29.0 34.7 32.9 16.4 14.9 6.2 


163



Table A1.11: Incidence of poverty by planning region, 1987, 1994 and 2000 

Planning Region 50% income 60% income 60% income Percentage of 

line incidence line incidence line + basic households 

deprivation incidence 

1987 1994 2000 1987 1994 2000 1987 1994 2000 1987 1994 2000 

% % % % 

East 20.2 32.2 27.6 23.7 31.9 26.6 29.5 39.9 24.3 35.6 38.1 39.8 

South-West 16.1 14.8 16.8 16.5 15.4 16.3 15.3 13.7 16.3 15.4 15.5 13.6 

South-East 11.0 12.7 11.7 10.2 12.0 12.1 9.2 13.0 14.1 8.9 10.4 9.9 

North-East 8.6 4.5 7.5 7.3 5.9 8.0 7.4 4.7 10.8 5.9 5.4 5.8 

Mid-West 13.2 10.2 7.7 12.4 9.9 8.2 13.0 7.8 13.1 10.9 8.9 7.7 

Midlands 8.6 8.3 10.0 9.5 8.4 10.2 7.9 6.3 5.0 6.7 7.3 8.5 

West 11.0 7.9 8.8 10.2 7.6 8.9 8.0 3.6 6.8 9.5 7.4 7.7 

North-West & Donegal 11.3 9.2 9.9 10.3 8.8 9.8 9.8 11.0 9.6 9.0 6.9 7.1 

Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100 100.0 100.0 100.0 


164


Appendix 1 

five of the eight regions being found between 29 and 37 per 

cent. 

At the 60 per cent line, reductions in the level of poverty 

between 1994 and 2000 were observed for the East, with 

the figure going from 29 to 22 per cent, the Mid-West (39 to 

35 per cent) and the South-East (40 to 35 per cent). 

Elsewhere slight increases were observed. The lowest rate 

of 22 per cent was again observed in the East and the 

highest rates of 46 to 47 per cent were found in the North- 

East and Donegal and in the North-West. In 2000, seven of 

the eight regions were found in the range running from 35 to 

47 per cent. 

For the combined income and deprivation measure, a sharp 

decline in poverty was found in all regions between 1994 

and 2000, with the exception of the North-East, where the 

decline was more modest. The largest absolute decline was 

in Donegal and the North-West, where it went from 24 to 9 

per cent, but similar proportionate increases were observed 

in a number of areas. The lowest poverty rate in 2000 of 

less than 4 per cent was observed in the East and the 

Midlands, followed by the West, where the figure was 6 per 

cent. It then rises to 8 to 9 per cent for the South-East and 

the South-West and the North-West and Donegal. It then 

reaches almost 12 per cent for the North-East. 

A1.3 Incidence of Poverty by Planning Region from 

the LII Survey 

In Table A1.11 we set out incidence of poverty by planning 

region. At the 50 per cent line changes in the distribution of 

poverty across planning region over time were extremely 

165



modest. The most notable change was a decline from 32 to 

28 per cent of poor households located in the East. 

The picture is very similar for the 60 per cent line, where the 

largest change between 1994 and 2000 is a reduction in the 

percentage for the East from 32 to 27 per cent. The changes 

for the consistent poverty line are somewhat more 

substantial. The number located in the East (including Dublin) 

increased significantly, from 30 to 40 per cent between 1987 

and 1994, and declined very significantly between 1994 and 

2000 to 24 per cent. Corresponding increases between 1994 

and 2000 are observed in the North-East, where the figure 

goes from 5 to 11 per cent, and in the West, where it goes 

from 4 to 7 per cent. 

166


Appendix 2 

Methodology 

A2.1 The Surveys 

As well as drawing on data from the 2002 Census of 

Population, this report makes extensive use of data from 

two surveys: the Living in Ireland Surveys (LII) and the 

National Survey of Housing Quality (NSHQ). The key 

features of each of these surveys are outlined below. 

The Living in Ireland Surveys 

The Living in Ireland Surveys form the Irish component of 

the European Community Household Panel (ECHP), an 

EU-wide project co-ordinated by Eurostat to conduct 

harmonised longitudinal surveys dealing with household 

income and labour situations in the member states. The aim 

of the ECHP is to produce a fully harmonised data set 

providing information on the social situation, financial 

circumstances and living standards of a panel of 

households to be followed over several years. The fact that 

the same set of households is interviewed each year means 

that it is possible to study changes in the characteristics 

and circumstances of particular households or individuals 

over time. 

167



The first wave of the LII was conducted in 1994, and the 

same individuals and households were followed each year. 

The wave conducted in 2001, therefore, was the eighth wave 

of the survey. In 2000, the Irish sample of individuals and 

households followed from Wave 1 was supplemented by the 

addition of 1,500 new households to the total. This was done 

in order to increase the overall sample size, which had 

declined due to attrition since 1994. A larger sample size 

ensures that the precision of estimates of key figures, such as 

the poverty rate and average equivalised household income, 

remained at a high level. It also allows a greater 

disaggregation of the data so that the situation of policyrelevant 

sub-groups, such as the unemployed or older adults, 

can be examined. These additional households, as well as 

the original sample, were followed in 2001. 

Survey Structure 

The ECHP involves a household questionnaire which is 

completed by the ‘reference person’ or person responsible for 

the accommodation, and an individual questionnaire which is 

completed by each adult (age 16 or over) in the household. 

The main items of information collected on the questionnaires 

were household size and composition, tenure, standard of 

living (things the household has or can afford to do), financial 

strain, economic activity, training and education, general 

health and very detailed information on income from 

employment, self-employment, personal and occupational 

pensions, social welfare, education and training-related 

allowances and grants, property (interests, dividends, rental 

income) and other sources. Income was measured in great 

detail and income information was collected from each adult 

in the household. 

168


Appendix 2 

The questionnaires were administered in a face-to-face 

interview by the ESRI’s team of interviewers. In farm 

households, a farm questionnaire was also completed to 

collect information to be used in conjunction with Teagasc’s 

National Farm Survey to estimate the income flow (family 

farm income) of farm households. This approach was 

necessary because the nature of farm income, being a 

combination of market profit or loss, grants and subsidies, 

makes it difficult for respondents to provide the figure directly. 

The Sample 

The sample of households was originally selected for the 

1994 wave of the survey. The objective of the sample design 

was to obtain a representative sample of private households 

in Ireland. Those living in institutions such as hospitals, 

nursing homes, convents, monasteries and prisons are 

excluded from the target population, in line with the 

harmonised guidelines set down by Eurostat and standard 

practice adopted in surveys of this kind (such as the 

Household Budget Survey conducted by the Central 

Statistics Office 1 ). Among those effectively excluded from the 

target population are a number of small groups that face a 

relatively high risk of poverty, such as the homeless and 

Travellers. To do justice to the particular circumstances of 

groups such as these would require a different research 

methodology. 

1 Collective households, however, were included. These are private 

households containing five or more unrelated persons with a looser 

budget-sharing relationship than in the standard private household. 

The main examples are boarding or lodging houses and army 

barracks. An individual living in a collective household is treated as a 

one-person ‘sub-household’. 

169



The sample was selected using the ESRI’s RANSAM system, 

which was developed at the institute and has been 

successfully used for selecting random samples from the 

electoral register for over two decades. RANSAM allows one 

to pre-stratify the sampling frame according to any 

combination of census variables. In selecting the sample for 

the Living in Ireland Survey, the following strata were used: 

• Province: Four categories – Dublin, rest of Leinster, 

Munster, Ulster/Connaght. 

• Urban/rural: Two categories – district electoral divisions 

(DEDs) with more than 50 per cent of their population in 

towns with a population of 1,500 or more versus the rest. 

• Unemployment: Two categories – DEDs with an 

unemployment rate of 16 per cent or more versus the rest. 

The target sample selected using the ESRI’s RANSAM 

procedure was a sample of persons, not of households. Since 

the probability of selection is greater for households with a 

larger number of registered voters, this means that the 

resulting sample will tend to over-represent larger 

households. This was taken into account in reweighting the 

sample for analysis. 

The total number of households successfully interviewed in 

1994 was 4,048, representing 57 per cent of the valid sample. 

This response rate is as one would expect in an intensive and 

demanding survey of this nature, and is comparable to the 

response rates achieved in the Household Budget Surveys. A 

total of 14,585 persons were members of the completed 

households. Of these, 10,418 were eligible for personal 

interview, i.e. born in 1997 or earlier, and 9,904 eligible 

respondents completed the full individual questionnaire (964 

on a proxy basis). 

170


Appendix 2 

Reinterviewing Households and Tracing Individuals Who 

Move 

The sample from the Wave 1 (1994) Living in Ireland Survey 

was followed in subsequent years and reinterviewed. The 

follow-up rules for the survey meant that new households 

might be included in each wave where a sample person from 

Wave 1 moved to another household. All individuals in the 

Wave 1 sample were to be followed in subsequent waves and 

household and individual interviews were to be conducted, as 

long as the person still lived in a private or collective 

household within the EU. 

Sample Supplementation in 2000 

Even with a relatively high year-on-year response rate, there 

was a substantial loss of respondents over time. Of the 

original sample individuals who were still ‘in scope’ in 2000, 

5,530 individuals still in scope 2 (40 per cent) were in 

completed households. 

The main reason for household non-response was refusal 

(ranging from 9 per cent of the eligible sample in Wave 2 to 5 

per cent in Wave 5). Among the newly generated households, 

difficulties in obtaining forwarding addresses for those who 

moved also contributed to the non-response rate. 

Detailed checks on the pattern of attrition between waves 

showed that there was no evidence of a disproportionate loss 

of households from the upper or lower ends of the income 

distribution of the kind that would tend to bias estimates of 

average household incomes or poverty measures. 

2 Of the original 14,585 individuals, a total of 400 had died by 2000 and 

324 had moved to an institution or outside the EU. This left 13,861 

individuals still ‘in scope’ by 2000. 

171



However, the reduced sample size was addressed by 

supplementing the sample in Wave 7. The new sample was 

selected using the same procedure as for the first wave of the 

survey in 1994, using the ESRI’s RANSAM programme based 

on the electoral register. The household response rate 

reached 57 per cent for the 2,661 new sample households 

contacted by interviewers. This is the same as the rate 

achieved in Wave 1 and is in line with the typical response 

rate in other surveys of a demanding nature, such as the 

Household Budget Survey. 

The sample supplementation exercise, together with the 

follow-up of continuing households, resulted in a completed 

sample in 2000 of 11,450 individuals in 3,467 households. 

Individual interviews were conducted with 8,056 respondents, 

representing 93 per cent of those who were eligible (born in 

1983 or earlier). 

Sample Weights for the LII Data 

The purpose of sample weighting is to compensate for any 

biases in the distribution of characteristics in the completed 

survey sample compared to the population of interest, 

whether such biases occur because of sampling error, from 

the nature of the sampling frame used, differential response 

rates or attrition. 

Whatever the source of the discrepancy between the sample 

and population distributions, we would like to adjust the 

distributional characteristics of the sample in terms of factors 

such as age, sex, economic status and so on to match that 

of the population. In a cross-sectional survey, or in the first 

wave of a panel survey, the only way to check the 

distributional characteristics of the sample is to compare 

172


Appendix 2 

sample characteristics to external population figures from 

sources such as the census, the Labour Force Survey, official 

statistics on the number of social welfare recipients from the 

Department of Social Welfare and so on. In waves following 

the first wave of a panel survey, we can also compare the 

characteristics of the individuals and households successfully 

followed to those of the individuals and households in a 

previous wave of the survey. In constructing the weights for 

the Living in Ireland Survey in Waves 2 and subsequently, 

both of these methods were used. 

The household weights were developed in a number of steps: 

• The first step involved adjusting the continuing sample for 

attrition. 

• The second step was to calibrate the sample totals 

against population totals from external sources. 

The external sources of information used were the Quarterly 

National Household Survey (QNHS), the Department of 

Social, Community and Family Affairs statistics on social 

welfare recipiency levels and figures from Teagasc on the 

total number of farms by farm size. The result of the 

weighting procedure was to ensure as close a match as 

possible between the sample and the population in terms of 

the distribution of the characteristics shown in Table A2.1. 

Apart from incorporating weights to control for attrition from 

previous waves and the availability of new technology for 

constructing weights, the logic and general strategy in 

developing the weights for Waves 2 to 7 was very similar to 

that used in Wave 1. The resulting match between the 

weighted sample characteristics and the population 

173



characteristics for the 2001 data was highly satisfactory, 

confirming that the weights are effective in adjusting the 

achieved sample to population characteristics. 

Table A2.1: External population characteristics used in the 

construction of household weights for the Living in Ireland 

Surveys 

Household characteristics: 

Household size (total size, number over 18 and number over 65) 

Location (Dublin, other county borough, rural) 

Number of persons at work (0, 1 and 2 or more) 

Head age (under 25, age 25 and over) 

Number of farms in each of six size categories 

Individual characteristics: 

Number of males and females by 10 age categories 

Number of males and females age 15+ by 11 age/marital status 

categories 

Number of recipients of 12 major social welfare payments 

Number of males and females by seven economic status categories (at 

work (ILO), unemployed (ILO), unemployed (not ILO), student, home 

duties, retired, other) 

Number of males and females age 20–64 by level of education (four 

categories) 

The National Survey of Housing Quality (NSHQ) 

The Economic and Social Research Institute (ESRI) was 

commissioned by the Department of the Environment to carry 

out the Irish National Survey of Housing Quality (NSHQ) in 

2001/2002. The purpose of the survey was to record very 

174


Appendix 2 

detailed information on the condition of the national housing 

stock. Similar surveys were undertaken in 1981 and 1991, but 

were conducted by the local authorities themselves. This is 

the first time that the methodology, administration and 

protocols for the survey have been completely centralised, 

ensuring a harmonised set of data across local authority 

areas. 

The survey collected detailed information on the condition of 

the dwelling; on residents’ satisfaction with aspects of their 

dwelling, such as costs, heating system and water supply; on 

problems in the area where the dwelling is located; and on 

problems with the affordability of the dwelling itself, with 

heating the dwelling or with home appliances and furnishings. 

The Sample 

One of the requirements of the survey was to provide a 

database to the Department with a large enough sample to 

yield separate breakdowns at local authority level. The NSHQ 

completed sample size was over 40,000 households 

throughout the country. This is an extremely large sample by 

the standards of other sample surveys which have been 

previously carried out in Ireland. The sample of addresses 

was selected using the ESRI’s RANSAM programme, which 

uses a multi-stage randomised design based on the electoral 

register. 

The Survey 

The survey is based on a questionnaire interview of a 

household respondent. In each household, the person 

responsible for the accommodation (the owner, purchaser or 

tenant) was interviewed. A pilot test of the questionnaire was 

175



conducted in August 2001 and the main survey went into the 

field in September. The fieldwork for the main survey 

extended from September 2001 to summer 2002. The 

questionnaire had an average completion time of 30 minutes. 

The overall response rate was 75 per cent. By the standards 

of statistical probability surveys currently undertaken in 

Ireland, these response levels are extremely high. The higher 

than usual response level in the NSHQ can be attributed to 

intensive interviewer training and sustained call-backs on the 

part of the interviewers. 

Sample Weights 

As noted above, sample weights are constructed to ensure 

that the sample is representative of the population along a 

number of key dimensions, such as region, household size, 

labour force participation, age of dwelling and so on. These 

weights adjust the sample for any lack of overall 

representativeness arising from sample design, the sampling 

frame available and patterns of non-response. The sample 

design for the NSHQ would have over-represented rural areas 

because of the requirement, noted above, for a sample of 

sufficient size to provide local authority-level tables. This 

meant that smaller local authority areas were overrepresented 

in the sample compared to their populations. The 

sampling frame, based on the electoral register, also tends to 

over-represent households with a larger number of persons 

over age 18. Differences in response rates are typically found 

between urban and rural areas, with higher response rates in 

the latter. 

The sample weights were constructed by adjusting the 

sample proportions to population figures based on the most 

176


Appendix 2 

up-to-date information available. The population figures drew 

on data from reliable external sources, such as the 

preliminary figures from the 2002 census, from the Quarterly 

National Household Surveys and from the 1996 census with 

adjustments for population change. 

There were a number of steps involved in constructing the 

weights. The first involved constructing a weight to control for 

the fact that the sampling frame (based on the electoral 

registers) will tend to over-represent households with a larger 

number of adults. The weight was Wt1 = 1/A, where A is the 

number of adults age 18 or over in the household. 

The second weight grossed the number of sample cases in 

each local authority area up to the total number of private 

households in that local authority area, using preliminary 

figures provided by the Central Statistics Office based on the 

2002 census. Wt2 = (Wt1*PL )/SL, where PL refers to the total 

number of households in the local authority area and SL 

refers to the number of sample households in that local 

authority area. 

The next stage involved what is normally referred to as 

calibration (see, for example, Deville and Särndal 1992) – the 

second weight (Wt2) was adjusted so as to match the sample 

distribution of a given set of characteristics to the population 

distribution of these characteristics derived from external 

sources. The Gross programme was used to gross this 

second weight to local authority- and region-level totals for a 

set of control variables. 3 177 

3 This programme, developed by Johanna Gomulka, uses a minimum 

distance algorithm to adjust an initial weight (in this case Wt2) so that 

the distribution of cases in the sample matches a set of control totals.



The region-level totals were obtained from the Central 

Statistics Office, which provided special tabulations from 

the QNHS (second quarter, 2001). The local authority-level 

totals were obtained from the 1996 census (household size, 

number of persons at work) and the Department of the 

Environment Housing Statistics (number of local authority 

rented dwellings, new dwellings built after 1991). The countylevel 

figures from the 1996 census were updated to 2002 

figures using region-level information from the QNHS and 

preliminary county-level population and household totals from 

the 2002 census. 

At the time of constructing the weights, only the total males, 

total females and an estimate of the total number of 

households was available from the 2002 census. These 

figures were used to adjust the total number of males, 

females and households for each local authority to the figures 

for 2002. 

Table A2.2 shows the population checks that were included 

and the level (county or region level). For some variables, 

recent information was only available at the level of planning 

region (from the QNHS). 

A2.2 Imputation of Missing Information on NSHQ 

Missing information on key background variables in the 

NSHQ 2001/2002 was imputed based on other data on the 

household. This was done to ensure that all figures in a table 

were based on the same set of cases. Imputation was also 

conducted for the variables used to construct the weights. 

The variables where the level of missing information 

exceeded 5 per cent are shown in Table A2.3. The figure also 

178


Appendix 2 

Table A2.2: Population checks for sample weighting 

Population checks at county level: 

• Household size (number of persons age 18 or over). From 1996 

census adjusted to 2002 figures using QNHS 2001 at region level and 

preliminary figures from 2002 census of number of males, females 

and households by local authority area. 

• Number of persons in household at work (three categories: none, one, 

two or more). From 1996 census adjusted to 2002 figures using 

QNHS at region level for second quarter 2001 and preliminary figures 

from 2002 census. 

• Number of local authority rented dwellings (from Department of the 

Environment Housing Statistics, September 2001). 

• Age of dwelling (from the 1991 census, updated using figures from the 

Department of the Environment Housing Statistics 2002 on new 

dwellings built since then 4 ). 

Population checks at level of planning region: 

• Household size (six categories, persons of all ages; from QNHS 2001). 

• Household type (five categories; from QNHS 2001). 

• Tenure (owner occupied, renter, other tenure; from QNHS 2001). 

• Age by sex (10 age groups; from QNHS 2001). 

• Occupation of oldest person, if at work (ISCO88, five categories; from 

QNHS 2001). 

• Education by sex (three categories of education; from QNHS 2001). 

• Economic status by sex (at work, unemployed, home duties, retired, 

student, other; from QNHS 2001). 

4 It was assumed that 0.6 per cent per annum of the 1991 housing stock 

was lost through demolition by 2002 – a total of 64,471 dwellings. It 

was further assumed that older dwellings would be lost at a greater 

rate: 70 per cent from the pre-1919 stock; 20 per cent from the 

1919–40 stock and 10 per cent from the 1941–60 stock. Of dwellings 

built since 1991, it was assumed that 1 per cent of the total built from 

1991–96 were for holiday use, rising to 1.5 per cent of the total built 

after 1996. 

179



shows the percentage of values imputed and the basis on 

which imputation was conducted. 

The imputation involved matching the household where the 

data was missing to a similar household with non-missing 

data using a set of related characteristics (typically county, 

cluster, tenure, household size and other variables that are 

predictive of the variable to be imputed). The imputed value 

was taken from the household with the closest match in 

terms of these characteristics. This approach is preferable to 

imputing an average value since it preserves the variation of 

the variable being imputed. 

Information on household members (age, sex, education, 

economic status and occupation of the oldest person) was 

needed for all households to weight the sample. 

A2.3 Income Correction Factor 

Measures of household income based on a single item, as is 

done in the NSHQ, will tend to understate income compared 

to the figure obtained if all household members are asked 

about their income from different sources. We know this from 

the Living in Ireland (LII) Surveys. The LII Surveys make use of 

both a single-item measure on the household questionnaire 

and a detailed set of questions on each income source posed 

to all adults in the household. The single-item measure 

understates total household income by 19 per cent (of the full 

measure) on average (or 24 per cent of the single-item 

measure). Table A2.4 shows that the degree to which the 

single-item measure understates total income is greater for 

households with a large number of income sources (typically 

associated with a larger number of adults) and households 

180


Appendix 2 

Table A2.3: Level of missing information on key variables and 

imputation procedure 

Variable % imputed Variables used to impute the value 

Sex of household 1.4 Household size, local authority area, 

member 

cluster, sex of spouse (where applicable) 

Age of household 5.8 Age of spouse/parent/child (where 

member (all members) 

applicable), economic status (retired), 

household size, local authority area, 

cluster 

Highest level of 19.8 Occupation, age, sex, household size, 

education achieved 

housing tenure, local authority area, 

by each household 

cluster 

member 

Economic status 3.7 Age, sex, household size, local 

of household 

authority area, cluster 

member 

Occupational group 3.3 Housing tenure, age, education, Local 

of oldest person in 


the household, if 

at work 

Household income 12.3 Social class of householder, number of 

persons at work, number of adults, local 


Housing tenure 2.2 Household size, local authority area, 

cluster 

Household type 1.3 Household size, age of reference person, 

age of other persons, local authority area, 

cluster 

Size of place 2.3 Local authority area, cluster 

Age of dwelling 1.7 Local authority area, cluster 

Floor area 75.0 No imputation 

Presence of wall 18.0 No imputation 

insulation 

Source: National Survey of Housing Quality, 2001/2002 

181



where the main source of income is from self-employment or 

agriculture. The difference between the two measures is 

smallest for one-adult or two-adult households relying on 

pension or social welfare income. 

The understatement is particularly marked where there is 

income from work and where there are a large number of 

adults in the household. A regression-based model was 

developed to correct for this understatement using variables 

that are measured on both the NSHQ and the LII Survey from 

2000. The model was developed using the LII Survey and 

then the coefficients for the model were used to ‘correct’ the 

income measure on the NSHQ. The single-item measure of 

income in the LII Survey recorded income as a continuous 

amount, or into 10 categories if an exact amount could not 

be provided. Since the NSHQ used a categorical variable, the 

LII incomes were recoded into a categorical format before 

running the model. This would enable us to simulate the 

relationship between the continuous distribution of income 

based on aggregating information collected in detail from all 

adults in the household and a categorical measure recorded 

by the householder. 

The coefficients from the model are shown in Table A2.5. The 

r-squared for the model is .644, indicating that the variables 

included in the model explain about 64 per cent of the 

variance in income. 5 

The model used the income category (coded as a 

dichotomous variable with a value of 1 for each category), the 

number of adults in the household, the number of children in 

the household and the number of persons at work. 6 Table 

5 The r-square for the model with the income categories alone is .54. 

182


Appendix 2 

Table A2.4: Mean weekly household income (in £) using full 

measure and single-item measure by number of adults and 

number of persons at work 

Single- Full Difference Difference Difference 

item measure as % of as % of 

measure single-item full 

measure measure 

Number over 18 

1 184 195 11 6% 5% 

2 422 495 73 17% 15% 

3 483 633 149 31% 24% 

4 or more 621 932 311 50% 33% 

Number at work 

0 170 169 –2 –1% –1% 

1 345 407 62 18% 15% 

2 or more 569 748 179 31% 24% 

Total 389 482 93 24% 19% 


A2.5 shows that incomes clearly bear a strong relationship to 

the income category. The category coefficients in Table A2.4 

are below the lower bound of the category itself because they 

are shown net of the effect of number of adults and number 

of persons at work. Each household will have at least one 

adult and at higher levels of income are likely to have at least 

one person at work. The number of adults and the number of 

6 A number of more complex models were tested, including variables 

such as tenure, region, education and age of householder and 

dichotomous variables for number of adults and number at work, but 

no improvement in the predictive power of the model was achieved. 

183



adults at work also have strong coefficients. The effect of 

additional children is much weaker and does not reach 

statistical significance. Nevertheless, it was included in the 

model because in a household survey such as the NSHQ, 

which does not have income as a central focus, it is likely 

that many householders did not include Child Benefit in their 

estimate of total income. 

Table A2.5: Model based on LII to correct for understatement 

of income when a single-item measure is used 

Variables Coefficient Std. error 

Constant –9.18 89.62 

Number of adults over 18 55.42 5.03 

Number of children under 18 2.73 3.45 

Number of adults at work 96.44 5.73 

Income £50–£99 50.41 90.43 

Income £100–£149 67.34 90.55 

Income £150–£199 81.62 90.06 

Income £200–£249 110.00 90.48 

Income £250–£299 173.06 90.28 

Income £300–£399 215.77 89.98 

Income £400–£499 275.46 90.16 

Income £500–£599 374.17 90.45 

Income £600–£999 521.26 90.30 

Income over £1,000 888.55 92.44 

Source: Living in Ireland Survey, 2000. Note: The omitted category for 

income is ‘under £50’. 

184


Appendix 2 

Table A2.6: Income category midpoints and coefficients applied 

to the survey of house quality 

Lower Upper Point Lower Upper Point Coeffibound 

bound estimate bound bound estimate cients 

£ £ £ € € € 

Four-category 0 190 132 0 241 167 0.54 

measure 191 360 267 243 457 339 0.63 

(2.4% of 361 570 454 458 724 576 0.61 

households) 571 1,000 787 725 1270 999 0.65 

Sixteen- 0 85 75 0 108 54 0.00 

category 86 110 98 109 140 124 0.63 

measure 111 150 131 141 190 166 0.54 

(85.3% of 151 190 171 192 241 216 0.47 

households) 191 220 206 243 279 261 0.49 

221 270 246 281 343 312 0.63 

271 320 296 344 406 375 0.63 

321 360 341 408 457 432 0.62 

361 400 381 458 508 483 0.62 

401 450 426 509 571 540 0.61 

451 500 476 573 635 604 0.61 

501 570 536 636 724 680 0.68 

571 650 611 725 825 775 0.68 

651 750 701 827 952 889 0.65 

751 950 851 954 1,206 1,080 0.65 

951 Open 1,100 1,208 Open 1,333 0.81 

In the NSHQ, there were 16 income categories, rather than 10 

as in the LII data, and the amounts were presented to the 

respondents either in Irish pounds or in euro, depending on 

respondent preference, since the survey spanned the period 

of the euro changeover. The midpoints of the NSHQ income 

185



categories were matched to the nearest category from the 

LII so that the appropriate correction could be applied to 

the income category. The coefficients used were obtained 

by dividing the midpoint of each income category (shown 

in Table A2.6) by the corresponding coefficient in the 

model. The coefficients applied to each category are 

shown in Table A2.6. 

Table A2.7 shows the mean ‘corrected’ income for each 

household income category. Overall, incomes are adjusted 

upwards by about 24 per cent (see Table A2.8). In general, 

incomes in the lower categories tend to be adjusted 

upwards to a greater extent than incomes in the higher 

categories. 

The final column of Table A2.7 shows the percentage by 

which the predicted income would have been understated 

if the midpoint of the categories based on the single item 

had been used instead of the ‘corrected’ income. The 

biggest change is to the lowest category (€0 to €108). 

For the lowest income category, taking the midpoint of the 

category as a point estimate would not have been a good 

choice in any case. The general shape of the income 

distribution, rising steeply towards the lower end, would 

indicate the choice of a point estimate towards the upper 

bound of this category rather in the middle of it. 

Table A2.8 shows that the difference between the singleitem 

measure and the ‘corrected’ income is minimal where 

there is only one adult in the household or where there is 

nobody at work in the household. The difference is much 

larger where there are several adults in the household (the 

average increase is 65 per cent where there are four or 

more adults in the household) and where there are adults at 

186


Appendix 2 

Table A2.7: Mean ‘corrected’ income for each original income 

category in the NSHQ 

Lower Upper Mean Implied 

bound (€) bound (€) ‘corrected’ ‘under- 

Income statement’ 

(€) 

Four-category 0 241 292 43 

measure 243 457 562 40 

(2.4% of households) 458 724 759 24 

725 1,270 1,142 13 

Sixteen-category 0 108 198 73 

measure (85.3% of 109 140 219 43 

households) 141 190 260 36 

192 241 333 35 

243 279 386 32 

281 343 491 37 

344 406 572 34 

408 457 633 32 

458 508 678 29 

509 571 723 25 

573 635 781 23 

636 724 899 24 

725 825 981 21 

827 952 1,050 15 

954 1,206 1,193 9 

1,208 Open 1,607 17 

work. The increase is 12 per cent where there is one person 

at work and 37 per cent, on average, where there are two or 

more people at work. 

187



Table A2.8: Average income before and after correction by 

number of adults and number at work in the NSHQ 

A B C D 

Household Household Difference Difference 

income income (B – A) % 

(uncorrected) (corrected) (C ÷ A) 

€ per week € per week 

Number of adults (18+) 

1 316 316 0 0 

2 559 643 84 15 

3 601 815 214 36 

4 or more 726 1195 470 65 


0 261 260 –2 –1 

1 499 558 59 12 

2 or more 719 983 263 37 

Total 519 641 122 24 

A2.4 Comparing Disparities in Relative Poverty Risk 

in the LII and NSHQ 

As noted in Chapter 3, the analysis of poverty risk using the 

NSHQ focuses on disparities in risk, i.e. on the ratio of the 

risk faced by a particular sub-group to the overall risk at a 

national level. The following tables compare the disparities in 

risk of income poverty found using the NSHQ to those using 

the LII for a comparable set of variables. 

As can be seen from Table A2.9, the figures from both 

sources are very similar for relative income poverty. However, 

there is some tendency for the relative risk associated with 

working or not working to be exaggerated in the NSHQ. This 

188


Appendix 2 

Table A2.9: Disparities in risk of income poverty by region, 

tenure, household size, economic status and social class from 

the LII Survey and the NSHQ 

Poverty risk, 50% Poverty risk, 50% 

mean, Scale A mean, Scale A over 

over household household 

LII NSHQ LII NSHQ 

Region Border 1.4 1.3 1.4 1.3 

Dublin 0.7 0.7 0.6 0.8 

Mid-East 0.8 0.8 0.8 0.8 

Midlands 1.1 1.2 1.0 1.2 

Mid-West 1.0 1.1 1.1 1.1 

South-East 1.1 1.2 1.1 1.1 

South-West 1.2 1.1 1.2 1.1 

West 1.3 1.2 1.4 1.1 

Tenure Owner 0.8 0.8 0.8 0.8 

LATP 1.2 1.4 1.3 1.4 

Rent free 1.1 1.2 1.1 1.1 

LA tenant 2.6 2.4 2.4 2.2 

Private tenant 0.7 0.8 0.7 0.8 

Household 1 1.9 1.8 1.6 1.8 

size (all ages) 2 1.0 1.3 1.1 1.0 

3 0.6 0.6 0.7 0.7 

4 0.6 0.4 0.6 0.5 

5 or more 0.7 0.7 0.8 0.9 

Economic Work 0.4 0.2 0.5 0.3 

status of Unemployed 2.5 3.9 2.3 3.3 

householder Retired 1.5 2.6 1.6 2.4 

Home duties 2.5 3.4 2.1 3.1 

Student 2.4 2.2 2.4 2.2 

Other 2.7 3.8 2.6 3.2 

Social High professional/ 

class of managerial 0.3 0.2 0.3 0.3 

householder Low professional/ 

managerial 0.5 0.4 0.5 0.5 

Other non-manual 0.7 0.7 0.7 0.8 

Skilled manual 0.9 0.8 1.0 0.8 

Semi-skilled manual 1.5 1.1 1.5 1.1 

Unskilled manual 2.0 1.5 1.9 1.5 

Unknown 2.4 2.3 2.2 2.1 

Number at 0 2.7 3.2 2.4 2.8 

work in 1 0.6 0.3 0.8 0.5 

household 2 0.2 0.0 0.2 0.1 

3 or more 0.1 0.0 0.2 0.0 

Total 1.0 1.0 1.0 1.0 

189



is likely to be due to the use of the work variable in the 

income correction equation described above. As we shall 

see later, the differences disappear when material 

deprivation, as well as income, is taken into consideration. 

A2.5 Comparing the Pattern of Disparities in 

Deprivation in the LII Survey and NSHQ 

Table A2.10 presents corresponding results for the 

modified consistent poverty (MCP) measure discussed in 

Chapter 3. Note that the measures use different sets of 

items for the ‘basic deprivation’ set so that the results from 

the NSHQ are for ‘modified’ consistent poverty, rather than 

for consistent poverty as measured in the LII Survey. 

Consistent poverty in the LII Survey is measured as being 

below 60 per cent of the mean equivalised household 

income (using Scale A) and lacking at least one item from 

the set of basic deprivation items (meat, new clothes, 

adequate heat, overcoat, shoes, roast, debt, do without 

meal, do without heat). MCP in the NSHQ is measured as 

being below 60 per cent of the mean equivalised 

household income (Scale A) and lacking at least one of the 

items in the basic set from the NSHQ (meat, new clothes, 

adequate heat, arrears, socialising, furniture and holiday). 

Given the differences between the measures of deprivation, 

it should not be surprising that the results diverge to a 

greater extent than for income poverty risk. Nevertheless, 

the main patterns still hold up well across both measures. 

The NSHQ measure shows somewhat less variation by 

regional authority, but the Border region remains most 

deprived and the Dublin region least deprived. The 

190


Appendix 2 

Table A2.10: Disparities in risk of (modified) consistent poverty 

(60% mean and basic) by region, tenure, household size, 

economic status and social class from the LII Survey and the 

NSHQ 

Consistent 

poverty LII 

Modified 

consistent 

poverty NSHQ 

Regional authority area 

Border 1.8 1.4 

Dublin 0.5 0.7 

Mid-East 1.0 0.8 

Midlands 0.6 1.2 

Mid-West 1.5 1.1 

South-East 1.1 1.1 

South-West 1.1 1.1 

West 1.0 1.1 

Tenure 

Owner 0.5 0.7 

LATP 0.8 1.5 

Rent free 2.0 1.1 

LA tenant 5.2 2.8 

Private tenant 1.6 0.9 

Household size 

1 1.8 1.8 

2 0.9 1.0 

3 0.9 0.7 

4 0.5 0.6 

5 0.8 0.9 

Economic status of householder 

Work 0.3 0.3 

Unemployed 5.7 4.1 

Retired 1.2 2.2 

Home duties 2.4 3.3 

Student – 2.6 

Other 2.4 4.0 

Social class of householder 

High professional/managerial 0.1 0.2 

Low professional/managerial 0.3 0.4 

Other non-manual 1.0 0.7 

Skilled manual 0.8 0.8 

Semi-skilled manual 2.1 1.1 

Unskilled manual 1.7 1.7 

Unknown 1.8 2.2 


0 2.7 2.9 

1 0.7 0.5 

2 0.1 0.1 

3 0.0 0.0 

Total 1.0 1.0 

191



Midlands appears more deprived than average on the NSHQ 

measure, but less deprived than average on the LII measure. 

The Mid-West appears 50 per cent more likely than average 

to face consistent poverty on the LII, but is only 10 per cent 

above average on the NSHQ. 

In terms of tenure, the local authority tenant purchasers show 

50 per cent more deprivation than average on the NSHQ 

measure, but 20 per cent less on the LII measure. On the 

other hand, private sector renters appear 60 per cent more 

deprived than average on the LII measure, but 20 per cent 

less deprived than average on the NSHQ. The relative 

disadvantage of local authority renters is much greater on the 

LII measure (5.2 times the average versus 2.8 times). 

The patterns using both measures are very similar across 

household size categories and for number of persons in the 

household at work and are broadly similar for social class of 

householder. 

Differences are also apparent for economic status of the 

householder. The extent of disadvantage experienced by the 

unemployed is greater on the LII, while the extent of 

disadvantage experienced by those retired, engaged on 

home duties or otherwise inactive (including people unable to 

work due to illness or disability) is greater on the NSHQ. 

192

xxGlossary/Refs 7/7/05 6:59 am Page 193 

Glossary 

Consistent poverty A measure of poverty that takes account 

of living standards as well as income. To be ‘consistently 

poor’, a household must fall below 60 per cent of the mean 

equivalised household income and lack basic necessities. 

Equivalised income An adjustment to total household 

income to take account of the household size and 

composition. It is the income per ‘adult equivalent’ in the 

household. It is calculated by giving a weight to each adult 

(age 15 and over) and each child (under age 15) in the 

household and dividing total household income by the sum of 

the weights for household members. 

Incidence of poverty This term refers to the proportion of all 

poor households who are found in a specific sub-group or 

region, e.g. the percentage of all poor households that are 

located in the Border-Midlands-West (BMW) region. 

LII The Living in Ireland Surveys – a panel survey of several 

thousand households that followed the same individuals over 

time between 1994 and 2000. It collected detailed information 

on incomes, economic activity, household composition and 

living standards. In 2000 the sample was supplemented by 

adding 1,500 new households. 

193



Nagelkerke R 2 A measure of how much of the variation in a 

dependent variable is explained by a set of independent 

variables. Like the R 2 in ordinary least squares regression 

(OLS), it can vary between 0 and 1. A value closer to 1 

indicates that more variation is ‘explained’ by the 

independent variables. 

NAPS National Anti-Poverty Strategy. 

NSHQ National Survey of Housing Quality conducted in 

2001/2002, with a sample size of over 40,000 households. 

Odds ratio A measure used to compare the risk of different 

groups. For instance, in 2000, 3.9 per cent of homeowners 

(without a mortgage) fell below the consistent poverty line, 

giving them a 0.04 (3.9/96.1) odds of being poor. On the other 

hand, 32.4 per cent of local authority tenants were found 

below the consistent poverty line, giving them a 0.48 

(32.4/67.6) odds of being poor. The difference in risk of being 

consistently poor between these two groups (local authority 

tenants and outright owners) can be indexed by the ratio of 

these odds (.48/.04), which gives an odds ratio of 12:1. Thus, 

local authority tenants had a 12 times greater risk of being 

poor than homeowners. 

Risk of poverty The percentage of persons or households in 

a group that fall below the poverty line, e.g. the percentage of 

households in the Border-Midlands-West (BMW) region that 

are poor. 

194


References 

Callan, T., Nolan, B. and Whelan, C.T. (1993), ‘Resources, Deprivation and 

the Measurement of Poverty’, Journal of Social Policy, vol. 22, no. 2, 

pp. 141–72. 

Callan, T., Nolan, B., Whelan, B.J., Hannan, D.F. and Creighton, S. (1989), 

Poverty, Income and Welfare, Dublin: Economic and Social Research 

Institute. 

Callan, T., Nolan, B., Whelan, B.J., Whelan, C.T. and Williams, J. (1996), 

Poverty in the 90s: Evidence from the 1994 Living in Ireland Survey, 

Dublin: Oak Tree Press. 

Deville, J.C. and Särndal, C.E. (1992), ‘Calibration Estimators in Survey 

Sampling’, Journal of the American Statistical Association, no. 87, 

pp. 376–82. 

Fahey, T. and Williams, J. (2000), ‘The Spatial Distribution of Disadvantage 

in Ireland’, in B. Nolan, P. O’Connell and C.T. Whelan (eds.), Bust to 

Boom: The Irish Experience of Growth and Inequality, Dublin: 

Institute of Public Administration. 

Friedrichs, J. (1998), ‘Do Poor Residents Make their Residents Poorer 

Context Effects of Poor Neighbourhoods on Residents’, in H.J. 

Andreß (ed.), Empirical Poverty Research in Comparative 

Perspective, Aldershot: Ashgate. 

Gomulka, J. (1992), ‘Grossing-up Revisited’, in R. Hancock and H. 

Sutherland (eds.), Microsimulation Models for Public Policy Analysis: 

New Frontiers, STICERD Occasional Paper 17, LSE. 

Gomulka, J. (1994), ‘Grossing Up: A Note on Calculating Household 

Weights from Family Composition Totals’, Microsimulation Unit 

Research Note MU/RN/4, University of Cambridge Department of 

Economics, March 1994. 

195



Haase, T., McKeown, K. and Rourke, S. (1996), Local Development 

Strategies for Disadvantaged Areas, 192-195: Evaluation of the 

Global Grant in Ireland, Dublin: Area Development Management 

(ADM) Ltd. 

Jargowski, P.A. (1996), Poverty and Place: Ghettos, Barrios and the 

American City, New York: Russell Sage Foundation. 

Jencks, C. and Mayer, S. (1990), ‘The Social Consequences of Growing 

Up in a Poor Neighbourhood’, in L. Lynn, Jr. and M. McGeary, 

Poverty in the United States, Washington, DC: National Academy 

Press. 

Layte, R., Nolan, B. and Whelan, C.T. (2000), ‘Cumulative Disadvantage 

and Polarisation’, in B. Nolan, P. O’Connell and C.T. Whelan (eds.), 

Bust to Boom: The Irish Experience of Growth and Inequality, Dublin: 

Institute of Public Administration. 

Layte, R., Nolan, B. and Whelan, C.T. (2001), ‘Reassessing Income and 

Deprivation Approaches to Poverty in Ireland’, Economic and Social 

Review, vol. 32, no. 3, pp. 239–61. 

Murray, C. (1984), Losing Ground: American Social Policy, 1950–1980, 

New York: Basic Books. 

National Economic and Social Council (1990), Strategies for the Nineties: 

Economic Stability and Structural Change: Report 89, Dublin: NESC. 

Nolan, B. and Whelan, C.T. (1996), Resources, Deprivation and Poverty, 

Oxford: Clarendon Press. 

Nolan, B. and Whelan, C.T. (2000), ‘“Urban Housing” and the Role of 

“Underclass” Processes: The Case of Ireland’, Journal of European 

Social Policy, vol. 10, no. 1, pp. 5–21. 

Nolan, B., Gannon, B., Layte, R., Watson, D., Whelan, C.T. and Williams, 

J. (2002), Monitoring Poverty Trends in Ireland: Results from the 

2000 Living in Ireland Survey, Dublin: Economic and Social Research 

Institute. 

Nolan, B., Whelan, C.T. and Williams, J. (1998), Where are Poor 

Households: The Spatial Distribution of Poverty in Ireland, Dublin: 

Oak Tree Press. 

Pringle, D., Cook, S., Poole, M.A. and Moore, A.J. (2000), Cross-border 

Deprivation Analysis, Dublin: Combat Poverty Agency. 

Pringle, D.G. (1999), ‘Something Old, Something New: Lessons to be 

Learnt from Previous Strategies of Positive Territorial Discrimination’, 

196


References 

in D.G. Pringle, J. Walsh and M. Hennessy (eds.), Poor People: Poor 

Places, Dublin: Oak Tree Press. 

Pringle, D.G and Walsh, J. (1999), ‘Poor People, Poor Places: Conclusion’ 

in D.G. Pringle, J. Walsh and M. Hennessy (eds.), Poor People: Poor 

Places, Dublin: Oak Tree Press. 

Sampson, R.J., Morenoff, J.D. and Gannin-Rowley, T. (2002), ‘Assessing 

“Neighbourhood Effects”: Social Processes and New Directions in 

Research’, Annual Review of Sociology, no. 28, pp. 443–78. 

Tunstall, R. and Lupton, R. (2003), ‘Is Targeting Deprived Areas an 

Effective Means to Reach Poor People An Assessment of One 

Rationale for Area-Based Gunning Programmes’, CASE paper 70, 

Centre for Analysis of Exclusion, London School of Economics. 

Walsh, J. (1999), ‘The Role of Area-based Programmes in Tackling 

Poverty’, in D.G. Pringle, J. Walsh and M. Hennessy, (eds.), Poor 

People: Poor Places, Dublin: Oak Tree Press. 

Watson, D. and Williams, J. (2003), Irish National Survey of Housing 

Quality 2001–2002, Dublin: Economic and Social Research 

Institute/Department of the Environment, Local Government and 

Heritage, ESRI Books and Monographs Series No. 173. 

Wilson (1987), The Truly Disadvantaged: The Inner City, The Underclass 

and Public Policy, Chicago, IL: Chicago University Press. 

197

Mapping poverty is a longstanding concern, both for researchers and 

policy makers. This study updates and extends previous research on 

the spatial distribution of poverty using recent national data sources: 

the Census of Population (2002), the Living in Ireland Survey (2000) 

and, for the first time, the National Survey of Housing Quality (2001). 

The research uses three poverty indicators (household income, 

material deprivation and socio-demographic variables) to measure the 

distribution of poverty at various spatial and administrative levels, 

including county and city councils for the first time. 

The study addresses three key aspects of the spatial distribution 

of poverty: 

■ It identifies patterns with regard to the concentration of poverty 

and how these have evolved over time. 

■ It assesses if these patterns are significant in terms of the overall 

incidence of poverty. 

■ It considers the processes underlying poverty clustering, 

distinguishing between factors that influence the location and the 

causes of poverty. 

The findings of this study increase our understanding of the location 

of poverty and highlight numerous implications for policy on 

combating poverty and social exclusion, including area programmes, 

social housing and local anti-poverty strategies. The study will be of 

relevance to anti-poverty organisations, regional and local government 

and teachers/students of social policy and geography. 

€ 20.00

Mapping Poverty - Combat Poverty Agency

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?