Social Exclusion - Short and Long Term Causes and Consequences-Pages-1-52

Social exclusion: short and long term causes and consequences
Author(s) Parodi, Giuliana ; Sciulli, Dario

Imprint Heidelberg : Springer, c2012
Extent 1 online resource (x, 264 p.)
Topic HM
Subject(s) Social isolation; Economic policy
Language English
ISBN 9783790827729, 379082772X
Permalink http://books.scholarsportal.info/viewdoc.html?id=500981
Pages 1 to 52
AIEL Series in Labour Economics
For further volumes:

http://www.springer.com/series/7370
Giuliana Parodi Dario Sciulli

Editors
Social Exclusion
Short and Long Term Causes
and Consequences
Prof. Giuliana Parodi Asst. Prof. Dario Sciulli
Dipartimento di Metodi Quantitativi Dipartimento di Metodi Quantitativi
e Teoria Economica e Teoria Economica
Universit G. dAnnunzio Universit G. dAnnunzio
di Chieti-Pescara di Chieti-Pescara
Viale Pindaro 42 Viale Pindaro 42
65127 Pescara 65127 Pescara
Italy Italy
parodi@unich.it d.sciulli@unich.it
ISSN 1863-916X
ISBN 978-3-7908-2771-2 e-ISBN 978-3-7908-2772-9
DOI 10.1007/978-3-7908-2772-9
Springer Heidelberg Dordrecht London New York
Library of Congress Control Number: 2011936144
Springer-Verlag Berlin Heidelberg 2012

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcast-
ing, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this
publication or parts thereof is permitted only under the provisions of the German Copyright Law of
September 9, 1965, in its current version, and permission for use must always be obtained from
Springer. Violations are liable to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
Cover design: eStudio Calamar, Berlin/Figueres
Printed on acid-free paper
Physica-Verlag is a brand of Springer

Springer is part of Springer Science+Business Media (www.springer.com)
Acknowledgments
This book contains a collection of papers, most of which have been presented at
the thematic session on Social Exclusion of the XXV Annual Conference of the
Italian Association of Labour Economists (AIEL), held at the University
G. dAnnunzio of Chieti-Pescara in September 2010.
All the papers, before being published, have been submitted to a double blind
peer-review process. This has been possible thanks to the contributions of the
referees: all the papers have benefited from their valuable suggestions.
We are also indebted with various institutions that have provided us the
financial support that has made possible this publication. We would like to thank:
the Italian Association of Labour Economists, the Camera di Commercio, Indu-
stria, Artigianato e Agricoltura di Chieti, the Camera di Commercio, Industria,
Artigianato e Agricoltura di Pescara, the Dipartimento di Metodi Quantitativi
e Teoria Economica of the Universit G. dAnnunzio di Chieti-Pescara, the
Facolt di Scienze Manageriali of the Universit G. dAnnunzio di Chieti-
Pescara and the Universit G. dAnnunzio di Chieti-Pescara.
Referees
Luca Beltrametti, Universit di Genova, Italy
Roger A. Bowles, University of York, UK
Floro Ernesto Caroleo, Universit Parthenope di Napoli, Italy
Sergio Destefanis, Universit di Salerno, Italy
Donata Favaro, Universit di Padova, Italy
Renata Livraghi, Universit di Parma, Italy
Emanuele Millemaci, Universit di Messina, Italy
Catia Nicodemo, Universitat Autonoma de Barcelona, Spain
Carmen Pagliari, Universit G. dAnnunzio di Chieti-Pescara, Italy
Giuliana Parodi, Universit G. dAnnunzio di Chieti-Pescara, Italy
Matteo Picchio, Universiteit van Tilburg, The Netherlands
Cristina Salvioni, Universit G. dAnnunzio di Chieti-Pescara, Italy
Dario Sciulli, Universit G. dAnnunzio di Chieti-Pescara, Italy
v
vi Acknowledgments
Antonello Scorcu, Universit di Bologna, Italy

Marcello Signorelli, Universit di Perugia, Italy
Peter J. Sloane, University of Swansea and IZA Bonn, UK
Massimo Tamperi, Universit Politecnica delle Marche, Italy
Francesco Timpano, Universit Cattolica del Sacro Cuore - Sede di Piacenza -,
Italy
Ainura Uzagalieva, Universidade dos Aores, Portugal
Robert J. Waldmann, Universit di Roma Tor Vergata, Italy
Contents
1 Introduction and Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Giuliana Parodi and Dario Sciulli
Part I The Structural and Long Term Causes of Social Exclusion
2 Health, Lifestyle and Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Gianluigi Coppola
3 A Comparative Analysis of Literacy Rate in Contributing

to Social Exclusion Insights. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Edgardo Bucciarelli, Carmen Pagliari, Fabrizio Muratore
and Iacopo Odoardi
4 Education and Socioeconomic Mobility in

Post-Communist Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Alina Veraschagina
5 Analysing Industrial Accidents in European Countries

Using Data Envelopment Analysis . . . . . . . . . . . . . . . . . . . . . . . . 93
Eugenia Nissi and Agnese Rapposelli
Part II Groups Structurally Vulnerable to Social Exclusion
6 Social Exclusion and Offending . . . . . . . . . . . . . . . . . . . . . . . . . . 105

Roger Bowles
7 Disability and Social Exclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 127

Peter J. Sloane and Melanie K. Jones
vii
viii Contents
8 Civilian Disability Pensions as an Antipoverty Policy Instrument?

A Spatial Analysis of Italian Provinces, 20032005 . . . . . . . . . . . 149
Massimiliano Agovino and Giuliana Parodi
Part III Economic Crisis, Labour Markets and Social Exclusion
9 Measuring the Long Wave. Unemployment, Discouragement

and Semi-Employment in Italy, During and After the Crisis . . . . 171
Leonello Tronti and Riccardo Gatto
10 Poverty and Unemployment: The Cases of Italy and Spain . . . . . 199

Tindara Addabbo, Rosa Garca-Fernndez, Carmen Llorca-Rodrguez
and Anna Maccagnan
11 Labour Market Transitions During the Financial

Crisis in Italy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Marco Lilla and Stefano Staffolani
12 Labour-Market Dynamics and Unemployment: New Scenarios

in the Italian Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Carlo Lucarelli and Chiara Mussida
Contributors
Tindara Addabbo Dipartimento di Economia, Universit di Modena e Reggio-

Emilia, Modena, Italy, e-mail: tindara.addabbo@unimore.it
Massimiliano Agovino Dipartimento di Metodi Quantitativi e Teoria Economica,
Universit G. dAnnunzio di Chieti-Pescara, Viale Pindaro 42, 65127 Pescara,
Italy, e-mail: agovino.massimo@gmail.com
Roger Bowles York Criminal Justice Economics York, UK, e-mail: roger.
bowles@btopenworld.com
Edgardo Bucciarelli Dipartimento di Metodi Quantitativi e Teoria Economica,
Italy, e-mail: e.bucciarelli@unich.it
Gianluigi Coppola Dipartimento di Scienze Economiche e Statistiche, Universit
di Salerno Salerno, Italy, e-mail: glcoppola@unisa.it
Rosa Garca-Fernndez Mtodos Cuantitativos para la Economa y Empresa,
Universidad de Granada Granada, Spain, e-mail: rosamgf@ugr.es
Riccardo Gatto ISTAT Rome, Italy, e-mail: rigatto@istat.it
Melanie K. Jones School of Business and Economics, Swansea University
Swansea, UK, e-mail: M.K.Jones@swansea.ac.uk
Marco Lilla Dipartimento di Economia, Universit Politecnica delle Marche,
Ancona, Italy, e-mail: m.lilla@univpm.it
Carmen Llorca-Rodrguez Departamento de Economa Internacional y de
Espaa, Universidad de Granada, Granada, Spain, e-mail: cmllorca@ugr.es
Carlo Lucarelli Ufficio Formazione e Lavoro, ISTAT, Rome, Italy, e-mail:
calucare@istat.it
Anna Maccagnan Dipartimento di Economia, Universit di Modena e Reggio-
Emilia, Modena, Italy, e-mail: anna.maccagnan@unimore.it
ix
x Contributors
Fabrizio Muratore Scuola Superiore, Universit G. DAnnunzio di Chieti-

Pescara, Pescara, Italy, e-mail: fabrizio.muratore@gmail.com
Chiara Mussida Dipartimento di Scienze Economiche e Sociali, Universit
Cattolica del Sacro Cuore, Sede di Piacenza, Italy, e-mail: chiara.mussida@
unicatt.it
Eugenia Nissi Dipartimento di Metodi Quantitativi e Teoria Economica,
Italy, e-mail: nissi@dmqte.unich.it
Iacopo Odoardi Scuola Superiore, Universit G. DAnnunzio di Chieti-
Pescara, Pescara, Italy, e-mail: iacopo.odoardi@alice.it
Carmen Pagliari Dipartimento di Metodi Quantitativi e Teoria Economica,
Italy, e-mail: c.pagliari@unich.it
Giuliana Parodi Dipartimento di Metodi Quantitativi e Teoria Economica,
Italy, e-mail: parodi@unich.it
Agnese Rapposelli Dipartimento di Metodi Quantitativi e Teoria Economica,
Italy, e-mail: a.rapposelli@unich.it
Dario Sciulli Dipartimento di Metodi Quantitativi e Teoria Economica, Universit
G. dAnnunzio di Chieti-Pescara, Viale Pindaro 42, 65127 Pescara, Italy,
e-mail: d.sciulli@unich.it
Peter J. Sloane School of Business and Economics, Swansea University, Swansea,
UK, e-mail: p.j.sloane@swan.ac.uk
Stefano Staffolani Dipartimento di Economia, Universit Politecnica delle
Marche, Ancona, Italy, e-mail: s.staffolani@univpm.it
Leonello Tronti Universit di Roma Tre, Rome, Italy, e-mail: l.tronti@funzione-
pubblica.it
Alina Veraschagina Dipartimento di Economia e Diritto, Universit di Roma La
Sapienza, Rome, Italy, e-mail: alinaver@gmail.com
Chapter 1
Introduction and Overview
Giuliana Parodi and Dario Sciulli
1.1 The XXV AIEL Conference
The XXV Conference of AIEL (Associazione Italiana degli Economisti del Lavoro)
was held in the Universit dAnnunzio in Chieti-Pescara in September 2010. The
theme chosen for the thematic session was Social Exclusion, in line with the
choice made by the European Union to make 2010 the European year to fight
poverty and social exclusion. The European Commission defines social exclusion
for societies and for individuals: for societies: disintegration and fragmentation of
social relations and hence of social cohesion for individuals a progressive
process of marginalization leading to economic deprivation and various forms of
social and cultural disadvantage. From another perspective, the socially excluded
can include individuals who are unable to participate in the basic economic and social
activities of the society in which they live. In the Conference, the themes under
investigation have ranged from causes to consequences of social exclusion, focusing
on disability, gender gaps, youth, income distribution, discrimination, accidents at
work, and literacy rates (LR). Selected works that were presented at the Conference
have been invited to be part of this book, to which distinguished authors, not present
at the Conference, have also been invited to contribute.
The book is organised in three parts: the first and the second parts deal with
structural, long-term causes of social exclusion (i.e., health, education, disability,
G. Parodi (&) D. Sciulli

Dipartimento di Metodi Quantitativi e Teoria Economica,
Universit dAnnunzio di Chieti-Pescara,
Pescara, Italy
e-mail: parodi@unich.it
D. Sciulli
e-mail: d.sciulli@unich.it
G. Parodi and D. Sciulli (eds.), Social Exclusion, AIEL Series in Labour Economics, 1
DOI: 10.1007/978-3-7908-2772-9_1, Springer-Verlag Berlin Heidelberg 2012
2 G. Parodi and D. Sciulli
crime); the third part deals with short-term causes of social exclusion and, in
particular, the effects of the most recent economic crisis. In most of the papers,
social exclusion is assessed in terms of unemployment, income, and education.
The first part deals with the long-term, structural causes of social exclusion and,
in particular, the problems associated with its possible effects on health (Coppola),
education (Bucciarelli, Pagliari, Muratore, Odoardi, and also Verashchagina), and
work conditions (Nissi and Rapposelli). The second part deals with vulnerable
groups who are particularly at risk of social exclusion [i.e., former convicted
persons (Bowles) and disabled persons (Sloane and Jones, Agovino and Parodi)].
Finally, the third part discusses problems connected with the mechanisms of the
labour market, one of the main direct or indirect causes of social exclusion, with
special emphasis on the effects of the most recent economic crisis on social
exclusion, including unemployment (Tronti and Gatto), poverty (Addabbo, Garca-
Fernndez, Llorca-Rodrguez and Maccagnan), and transitions within the labour
market (Lilla and Staffolani and also Lucarelli and Mussida).
In what follows, we briefly discuss the approaches used to investigate social
exclusion; we also revise the main effects of the most recent economic crisis on the
labour market and policy measures put (or not put) into effect to overcome its
effects. We then present a brief summary of the contributions that appear in the
three parts of the book. Finally, we draw some conclusions.
1.2 Main Approaches to the Discussion of Social Exclusion
Social exclusion is a term that has jumped to the forefront of the literature since it
was first used in France in the 1970s. Since then, it has been widely adopted as a topic
of economic policy, both by governments and by researchers; for example, in 1997,
the British government set up a Social Exclusion Unit (SEU), and the London School
of Economics set up the Research Centre for the analysis on Social Exclusion
(CASE); in 2007, the Social Inclusion Unit was set up in Australia; within the
European Union, the definition of social exclusion has been sharpened over time
(Nolan and Whelan 2010).
Social exclusion is interpreted both as a static situation of deprivation and as a
process leading to it. Multidimensional poverty and social exclusion are closely
related concepts, but they do not overlap. Both of these concepts refer to Sens
intuition of capability failures, but in the multidimensional approach to poverty,
failure is viewed in terms of shortfalls from a threshold for each function, and in the
social exclusion approach, failure is interpreted as an inability to participate and,
therefore, identifies functioning failures. These two different approaches refer to
quantitative and qualitative aspects, respectively, of social exclusion (Chakravarty
and Zoli 2009).
Initially, the threshold was defined in terms of income. Social exclusion came
very close to the definition of poverty; subsequently, it was defined in terms of
observed and unobserved heterogeneity. At the individual level, shortfalls in the
1 Introduction and Overview 3
exclusion or deprivation score of a person depend on the number of characteristics

from which that individual is excluded. With respect to heterogeneity, the litera-
ture differs both with respect to the variables considered relevant hindrances to
take part in societal life, and with respect to the weights to be attributed to them.
There is a consensus on some of the variables to be included in the definition, such
as housing, health and education. Other authors include financial difficulties, basic
necessities, durables, social contacts, and dissatisfaction. Authors choices vary in
terms of whether the effect among these variables is multiplicative or additive.
If multiplicative, being excluded in just one dimension is sufficient to create social
exclusion. Moreover, some dimensions may be considered more relevant than
others, so the same or different weights can be attributed to shortcomings in each
characteristic. Gender effects of social exclusion have been investigated, even
though only few significant differences have been found between genders.
Furthermore, few significant differences have been found between cohorts with
respect to social exclusion; therefore, the findings about the factors affecting social
exclusion appear to be robust.
With respect to time, social exclusion shows a dynamic or longitudinal
dimension: duration of the time spent in deprivation is considered important in
defining social exclusion, even though the literature differs in deciding how long
permanence in deprivation must be in order to allow for a definition of social
exclusion, or to lead to social exclusion. True state dependence, and therefore
persistence, has been investigated with respect to social exclusion. In fact, sto-
chastically socially excluded people appear to experience a shorter duration in
social exclusion compared with the structurally socially excluded (Adato et al.
2004). The inter-generational legacy of social exclusion has also been investigated,
overall and with respect to specific variables. Research has found that traditional
elements of transmission of disadvantage, such as class, play a minor role when
compared with other more specific variables. Moreover, there are indications that
some legacies of disadvantage perpetuate themselves across generations in a
segmented way, i.e., bad housing perpetuates itself, as do low educational
achievements, demographics and family behaviours, personality behaviours and
mental health (Nolan and Whelan 2010).
Societys social exclusion can be interpreted as the aggregate of individuals
social exclusion. Some axiomatic measures of social exclusion have been proposed
(Chakravarty and Zoli 2009), where societies overall social exclusion has been
derived from the individuals social exclusion. A countrys social exclusion is
determined by the weighted social exclusion of the various groups of excluded
people within it. In policy terms, the elimination of exclusion in one group would
reduce social exclusion in the country for Europe, see for instance Chackravarty
and DAmbrosio (2003).
The policy implications of these findings do not always follow a single trend:
short-term intervention(s) are needed if persistence appears as the main factor
explaining social exclusion, while long-term intervention(s) are needed to deal
with structural factors that affect social exclusion. In the European Commission
Report on Social Exclusion (2010), an account is given of EU structural
interventions and the financial intervention of individual member states. The

European Union has tackled the problem of social exclusion with various initia-
tives. In 2006, the Social Policy Agenda, 20062010, was established and renewed
by the Social Agenda presented in 2008. The structural European Funds provide
direct financial intervention, and in particular, the European Social Fund for the
20072013 period allocates 756 billion to the Member States. Furthermore, the
EU launched specific programs for categories of people particularly at risk of
social exclusion, such as the 20072013 PROGRESS (Programme for Employ-
ment and Social Solidarity), which is financed with 743 million; programmes for
youths; the HELIOS programme for disabled people; programmes for the elderly;
and programmes for gender equality. In addition, the European Union takes an
active role in raising awareness towards the problems of social exclusion and
poverty with diversified initiatives: it ran an investigation on peoples attitudes
towards social exclusion (e.g., 2007, Eurobarometer) and nominated 2010 as the
year to fight poverty and social exclusion, which entailed financing several
important initiatives in the Members States. At the level of individual countries,
the 2010 European Commission Report provides information on various policies
that address poverty and social exclusion, in connection with the labour market,
education, health, housing, and social networks. The labour market is the only area
where Member States have agreed on a common benchmark for 2010: to
encourage at least 25% of the long-term unemployed into training, re-training,
work practice, employment or another employability measure, combined,
where appropriate, with on-going job search assistance. In all other areas, Member
States pursue their own independent policies. In 2006, the EU27 on average
devoted 27% of their respective GNP to social protection; these transfers cover
both prevention policies (such as pensions and health) and alleviation policies
(such as unemployment benefits and housing). As an indicator of the structural
effect of these policies, we can look at the reduction in the at risk of poverty rate
after social transfers (excluding pensions), which is 27% for the entire EU 27, with
a particular strong effect on children under 18, for whom the reduction in the risk
of poverty rate rises to 39%. This aggregate figure conceals great heterogeneity
among countries; for instance, the risk reduction after social transfer is particularly
low in Mediterranean countries, such as Italy, Greece, and Spain.
1.3 Economic Crisis, Labour Markets and Social Exclusion
Economic conditions are an important factor in determining poverty and social

exclusion. During economic downturns, people are exposed to higher unemploy-
ment and, possibly, to fiscal austerity applied by governments in order to control
deficit and public debt levels. Moreover, without adequate policies, economic
downturns are likely to produce unbalanced outcomes that affect disadvantaged
people and that lead to higher inequality, poverty and social exclusion.
Leschke and Watt (2010), focusing on the role of the 20082009 economic
crisis on incomes, public budget and labour markets, have highlighted the impli-
cations of economic crisis regarding social effects. According to the
European Commission statistics (2010), more than 80 million Europeans,
including 20 million children, live at risk of poverty and social exclusion. The
economic crisis has made things worse by putting an estimated 5 million more
people out of work, potentially exposing many more families to poverty and social
exclusion.
European labour markets were strongly affected by the 20082009 economic
crisis; it caused a severe output decline that was quickly transmitted to the labour
market, producing a reduction in employment and working hours and, in many
countries, a strong increase of unemployment rates and underemployment. These
effects represent a challenge for labour markets and for the social policies of
European governments to mitigate the welfare loss of their citizens. Many
European countries entered the recession with low unemployment rates, partly
because of the effects of structural labour market reforms implemented since the
1980s. In any case, unemployment has increased strongly in the European Union
during the 20082009 economic downturn, reaching 9.6% in March 2010
(EUROSTAT 2010). However, the rise in unemployment has been quite hetero-
geneous across European countries. According to the OECD (2010) analysis, job
losses were particularly high in countries (e.g., Spain, Ireland, USA) where the
housing market contributed the most to the economic downturn. On the contrary,
unemployment has only slightly increased in countries where the economic crisis
was mainly explained by a decline in exports (e.g., Germany, the Netherlands,
Slovak Republic, and Japan). Moreover, job losses have been distributed hetero-
geneously among worker groups and industries. The economic crisis has more
strongly affected construction, mining and manufacturing sectors; young,
low-skilled and temporary workers have been the most affected at the individual
level by the economic downturn. Statistics indicate that youth unemployment is
two to three times greater than the average unemployment rate, and this effect
is especially prevalent among low-qualified workers, for which exit from the
labour market and/or the loss of earnings should be carefully avoided. Similar
considerations apply to immigrants.
Heterogeneous effects at worker level also reflect the structure of the economy
and labour market institutions. For example, Bentolila et al. (2010) have found that
Spain would have avoided 40% of new unemployment if it had adopted French
employment protection legislation. It follows that, while flexibility policies have
strongly contributed to increase employment in Europe until the economic
recession, flexible labour markets during economic crises are associated with
higher unemployment rates, especially for specific worker groups (e.g., youth
unemployment).
The 20082009 recessions has also affected the unemployment dynamics of
many European countries; it has caused an increase in unemployment inflows and
a reduction in unemployment outflows, contributing to an increase in the unem-
ployment pool and in the average duration of unemployment. In this case as well,
unemployment dynamics have affected various countries and various worker

groups differently.
European governments have adopted various measures with the aim of reducing
the social impact of the economic downturn. In this sense, even though the economic
recovery had begun in 2009, many European countries have continued to expand or
hold constant resources to finance unemployment benefits and re-employment
assistance. Unemployment benefit systems should be designed in such a way that the
job search effort is not negatively affected. Income support policies adopted at
European level have also included minimum-income benefits as well as income
support for low income workers. Active labour market program (ALMPs) spending
has scaled up when compared with passive spending in the last decade, but it seems
to be acyclical (OECD 2009) in absolute terms, with a small number of exceptions.
Some countries have also reinforced expenditures for public employment services.
Moreover, some specific measures, such as employment subsidies [including short-
term work (STW) schemes] and direct public job creation, have been expanded in
various European countries. For example, participation in STW schemes has
increased in many European countries, playing an important role in preserving
(permanent) jobs, through the encouragement of work sharing to reduce layoffs.
Furthermore, STW has reduced the loss of permanent employment by over 200,000
in Germany and by about 120,000 in Italy (OECD 2010). However, as STW just
preserves regular employment, it contributes to increased labour market segmen-
tation against temporary and part-time workers.
In this context, Italy has been extensively affected by the 20082009 economic
downturn, even though its situation appears to be better than many other European
countries. During the recession, the Italian GDP has fallen by 6%, but the
unemployment rate increased to just 8.6% in 2010 (ISTAT 2011), which is below
the EU average rate. This result has been achieved through the extensive use of
STW schemes, which have helped to mitigate the negative effect of the economic
crisis on the permanently employed. However, the economic recovery does not
seem particularly brilliant, and the effect of the crisis is likely to persist in the next
months. In Italy, the manufacturing sector has suffered most, mainly because of a
reduction in exports. Moreover, the rise in unemployment has not been homoge-
nous across workers groups. The male unemployment rate has risen more than the
female unemployment rate, primarily because of the reduction in labour market
participation of women. The youth unemployment rate has risen to 28.9%, opening
questions about the effectiveness of the structure of labour market policies in Italy.
1.4 Brief Summary of the Papers in the Book
We now briefly summarise the content of the book, presenting the summary of the
contributions according to the order in which they appear: (Sect. 1.4.1) deals with
general causes of social exclusion (i.e., health, education, and work conditions);
(Sect. 1.4.2) deals with some vulnerable groups who are most likely to be
associated with social exclusion (i.e., former criminals and disabled people); and
(Sect. 1.4.3) deals directly with the impact that the most recent crisis has had on
social exclusion through the labour market.
1.4.1 The Structural and Long-Term Causes

of Social Exclusion
Coppolas model predicts that the effect of income growth on health can be
ambiguous, providing an explanation to the observed health poverty trap. This
model provides micro foundations to Solows model of economic growth by
linking growth to health via considering the effect of lifestyle on growth. This
model centres on the concept of the lifestyle returns to scale that expresses the
elasticity of health with respect to income: a bad lifestyle implies that an increase
in income increases the consumption of commodities that are pernicious to health.
At the aggregate level, the Growth Model presented uses the Solow Growth
Model with a constant saving rate, diminishing returns of capital and labour,
labour augmenting technology, and constant returns to scale. The model develops
a health multiplier, on the assumption that health is labour augmenting; therefore,
individual choices about health affect society as a whole via externalities. How-
ever, lifestyle can also be an important channel to transmit economic growth to
health, depending on technological progress. Unlike economic growth, the rate of
health growth may be positive, null or negative, depending on the sign of the
Lifestyle Returns to scale: if it is negative, economic growth negatively affects
health.
Bucciarelli, Pagliari, Muratore, and Odoardi address the important issue of
investigating the connection between social exclusion and LR by analysing data
across 30 very diversified countries, some of which are OECD members, col-
lected for the period between 2007 and 2009. The topic under investigation is
important, as the LR and variables related to social exclusion can be seen as
proxies of human and social capital, which are essential for long-term devel-
opment. Knowledge of the variables that affect education is essential for poli-
cymaking purposes.
The paper has a comprehensive description of the statistical methodology that is
used in the investigation (i.e., multivariate regression analysis, maximum likeli-
hood with VARIMAX rotation, factor transformation matrix, factor analysis and
multivariate regression models with maximum likelihood components, and cluster
analysis). The paper is divided into several sections. Initially, it analyses LR as a
function of three groups of variables: variables strictly related to the educational
situation, i.e., school enrolment at three different levels and children out of school;
public intervention in education, expressed both in terms of general public
expenditures and in terms of GDP; and macro variables, such as real growth rate
and long-term unemployment rate. Because of the heterogeneity of the data, links
among the variables are not very evident. But, after applying the various methods
of analysis, it is found that LR can be affected by three components with the nature
of social exclusion variables, i.e., a synthetic indicator of social exclusion, an
indicator regarding educational level, and the GDP real growth rate. Most of the
countries analysed, including almost all of the OECD ones, show similar levels of
macro variables related to education, and instead differ on social exclusion. This
result means that all countries with medium or high levels of income per capita
show good levels of enrolment and attendance at school but that not all of them
can deal effectively with social problems. Finally, a detailed discussion of policy
towards literacy in the USA and Italy is discussed.
Veraschagina addresses the question of the role of education on social mobility
in 12 post-communist countries of Central Europe and the former Soviet Union.
Three hypotheses are tested: during the socialist era, the relation between the
education level of parents and their children weakened; the current levels of
educational mobility in post-communist countries are higher than in their Western
counterparts; and the transition to market economy caused an increase in educa-
tional persistence. Despite heterogeneity among countries, educational persistence
appears to decline sharply until the 1950s; however, evidence from later years does
not confirm this trend. To test the effect of this transition, identification of the
break was necessary. The CUSUM and the Chow test were used for this purpose,
and they showed that possibly not a single break occurred throughout the countries
but that additional breaks in the data referring to educational persistence took
place, on average about 10 years before the transition itself occurred. Moreover,
lower educational persistence in post-communist versus mature countries is con-
firmed by the evidence only for the generation born in the 19501960s. The
question of the sustainability of educational mobility was also investigated: the
increase of private returns to education (PRE), which the data show after the
transition, is expected to be associated with higher educational mobility.
This result is not confirmed by the evidence. A possible explanation is the shift in
the remuneration system, heavily relying on in kind transfers in the pre-transition
period, towards a completely monetary system of remuneration after the transition.
Taking this aspect into account, it is possible that the PRE has actually not been
altered. Unfortunately, the data do not allow an assessment of in kind transfers, so
the explanation provided here is a convincing speculation.
Nissi and Rapposelli apply the nonparametric technique Data Envelopment
Analysis to investigate comparative efficiency in terms of low numbers of acci-
dents at work to EUROSTAT data collected for 15 European countries in 2005, in
the sectors of manufacturing, construction and distribution trades. These sectors
were chosen because Europe registers the greatest number of accidents. Under the
traditional hypothesis of variable returns to scale, and under the specific
assumption of an input oriented approach, the paper develops the idea of unde-
sirable outputs, which only recently has been introduced in the literature and which
has mainly been applied to environmental studies. The results of this analysis show
that countries differ from each other with respect to efficiency in a different way,
according to the sector considered. In the manufacturing sector, four countries are
efficient (i.e., Germany, Ireland, Luxemburg, and the United Kingdom). They
make a total of four efficient countries in the construction sector as well; in the
distribution trades sector, Belgium and France join the other countries on the
efficient frontier, and Germany returns to being efficient, making a total of seven
efficient countries. The overall conclusion suggests that in the construction and
manufacturing sectors, the efficiency of reducing accidents at work can improve
considerably, while in the distribution trade sector, nearly 50% of the countries
considered are already efficient. Ireland, Luxemburg, and the United Kingdom are
efficient in all sectors considered, and Ireland and the United Kingdom are fre-
quently cited as exemplary efficient Decision Making Units, as they have a low
number of accidents at work.
1.4.2 Groups Structurally Vulnerable to Social Exclusion
The fragile groups considered in this part of the book include former convicted and
disabled people. Bowles uses various cross-sectional and longitudinal British data
sets to explore the relationship between crime and social exclusion, concentrating
the investigation on the human capital approach model of offending. According to
this model, crime is the result of a very high discount rate, leading to underin-
vestment in oneself. High discount rates can be explained in terms of a lack of
planning skills and an incapacity to anticipate the future, with special reference to
an inability to assess the dynamic consequences of offending. The formation of
this kind of individual preferences may be explained by variables that are inter-
twined with the concept of social exclusion. This model is well suited to explain
the offending behaviour of young people, who actually represent a large majority
of offenders. The variables that the data show to be correlated with offending are
experience of disadvantage as children, poor attendance at and exclusion from
school, exclusion from mainstream education, leaving school by the minimum
school leaving age, pre-custody drug use, and other crime victimisation rates.
Furthermore, reconviction rates appear to be related to prospects of unemployment
after release, and to problems of accommodation; the probability of reconviction
falls with age. There is also evidence of a comparatively high incidence of mental
health issues among prisoners. Of course, with all these factors, one cannot say
whether they are contributory factors to crime, or consequences of it, or whether
they both have a common cause. The policy implications of this model go well
beyond those of the punitive model of crime, and recommend both prevention
policies aimed at raising aspirations and encouraging investment in human capital,
and resettlement policies. A problem connected with this model is that the
predicted benefits of intervention spill over a wide range of potential savings in
costs, which affect various government departments, and it may be difficult to take
all of them into account if, for the purpose of financing, the benefits have to be
assessed in terms of crime prevention only.
Sloane and Jones present a very thorough review of the literature on disability
that is likely to become a standard tool of reference in the field. They pay par-
ticular attention to disabled people and the labour market, commenting on a
variety of datasets, ranging from the European ECHP and the Australian HILDA to
the British BHPS. The review touches on most aspects related to disability:
variability of the percentage of people who declare themselves disabled, the
probability of employment and seriousness of disability, evidence on disability and
low income from work, both with cross-sectional and with longitudinal data, and a
discussion of additional expenditure requirements for disabled people. The defi-
nitional aspect of disability is thoroughly investigated, with special emphasis on
work limited disability, non-work-limited disability, and the non-disabled. Other
issues also thoroughly investigated are problems associated with the age of onset
of disability, with self-reporting disability, with the simultaneous relations
between health and labour supply and vice versa, and with state dependence.
A thorough review of the literature using a UK background is also provided. The
problem of possible discrimination against disabled persons is investigated, with
special attention on disability and productivity, the importance of education and
disability on employment and earnings, work limited and non-work-limited
disability, part-time employment and self-employment, gender, possible job
mismatches, and job satisfaction. A thorough review of policies related to
disability, including income support, employment quota systems, sheltered
employment, anti-discrimination legislation, and employment policies, is also
provided. Individually targeted employment policies are presented as the most
promising policies to help disabled individuals to retain and gain employment for
which they are most suited.
Agovino and Parodi investigate whether Italian civilian invalidity pensions are
used as a form of income support, analysing data at the level of individual
provinces. Their analysis proceeds in two steps: a panel analysis and a Gen-
eralized Method of Moments where the spatial variable is introduced. The first
step of the analysis aims at identifying possible differences in disabling health
conditions in various Italian provinces. For this purpose, information about the
discharges from hospitals of people suffering from possibly invalidating illnesses
is analysed. The results of this analysis show that there is no significant dif-
ference in this variable across provinces. Subsequently, the analysis is developed
on the link between civilian invalidity pensions and indicators of social malaise
such as poverty. The poverty rate appears always significant as a regressor of
civilian invalidity pensions, even when the spatial lag is introduced in order to
take into account similarities between provinces which may go well beyond
administrative boundaries, because of historical and socio-economic factors.
These findings reveal a degree of flexibility in the allocation of civilian invalidity
pensions, which appears to be inconsistent with the strict rules defined by the
law, and suggest discretionary interpretation according to the locally prevailing
socio-economic conditions. The unsuitability of civilian invalidity pensions as a
form of income support is stressed.
1.4.3 Effects of the Recent Crisis on the Labour Market
Papers dealing with the effects of the 20082009 economic downturn on labour
markets and its socio-economic consequences include some case-studies from Tronti
and Gatto, Addabbo, Garcia-Fernandez, Llorca-Rodriguez and Maccagnan (Add-
abbo et al. hereafter), Lilla and Staffolani and, finally, Lucarelli and Mussida. Spe-
cifically, Tronti and Gatto investigate the 20082009 recession in the context of the
Italian labour market, focusing especially on the problems that concern the unem-
ployment measurement and its consequences in understanding the phenomena.
Addabbo et al. study the impact of unemployment on the probability of being income
poor and on the difficulties of accessing medical and dental visits treatments in Italy
and Spain to draw lessons that can be extended to the economic crisis. Finally, both
Lilla and Staffolani and Lucarelli and Mussida, focus on unemployment dynamics in
Italy clarifying the role of the 20082009 recession on labour market transitions.
Tronti and Gatto stress the relevance of considering various unemployment
measures to avoid shortcomings in evaluating the socio-economic impact of the
20082009 recession. According to official statistics, according to the standards
defined by the International Labour Office, the unemployment rate has risen to
8.6% in 2010, as a consequence of the strong economic downturn. Nevertheless,
the actual impact of the recession may be greater if the meaning of unemployment
is extended to include the concept of the labour underutilisation rate. Specifically,
Tronti and Gatto show that, considering relevant phenomena such as the dis-
couragement effect, with specific STW schemes providing for working zero hours
(Cassa Integrazione Guadagni), labour hoarding and semi-employment (atypical
workers), the resulting labour underutilisation rate is almost double the standard
unemployment rate. This finding has relevant implications when we consider the
impact of the crisis on the socio-economic conditions of individuals and house-
holds. Moreover, it changes the debate about the adequacy of official definitions of
unemployment in providing a reliable measure of the phenomenon and its
implications (see also Brandolini et al. 2006).
Addabbo et al. analyse the effect of being unemployed on the probability of
being income poor and on the difficulties of accessing medical and dental visits
treatments in Italy and Spain using 2007 IT-SILC/ES-SILC data. The authors also
describe the different unemployment insurance systems and the effect of the
economic crisis on both labour markets. Overall, Addabbo et al. find that being
unemployed increases the probability of being income poor and, partially, the
difficulties of accessing to medical and dental treatments in both countries. In Italy,
the effect is stronger for the previously self-employed now unemployed while in
Spain it is stronger for the unemployed who were never employed before. These
findings are relevant in analysing the possible effects of the 20082009 economic
downturn on households well being, considering the structure of welfare systems
with specific attention to the unemployment insurance systems.
Unemployment dynamics in Italy during the last few years have been studied in
two papers. The evidence from both of the papers is quite consistent, even though
some differences emerge, probably because of some differences in model speci-

fications and because of the data sets used. The results seem to reveal a strong
duality in the Italian economy, as the recession has had a strong impact on dis-
advantaged categories. Overall, young workers and immigrants have suffered more
than other groups during the jobs crisis because recent flexibility policies have
banished them from temporary employment. This finding raises questions about
the role of labour market policies and supports in Italy.
Specifically, Lilla and Staffolani use the ISTAT Labour Force Survey micro-data
to evaluate the consequences of the 20082009 economic crisis on the Italian labour
market by analysing the determinants of individuals transitions between occupa-
tional states. They use two different strategies to identify the determinants of tran-
sitions: first, they consider the perceived employment condition, available at the
individual level for the previous year; and, second, they use the ecological infer-
ence approach, i.e., they get quarterly transition probabilities for the official
employment condition. The estimation results confirm that during the crisis young
workers, men, immigrants, and individuals in the north of Italy faced the major risks
of becoming unemployed, while women, middle-aged and northern individuals had
major troubles in finding a job. With respect to the family types that were analysed
in the paper, the most affected individuals belong to couples without children and
to single parents without children. They experienced severe reductions in
employment stability along with an increase in the transitions to unemployment.
Finally, Lucarelli and Mussida provide a picture of the labour market transitions
in Italy for the 20042009 period, with specific attention on the effects of the
20082009 economic downturns. The authors find that particular characteristics
reduce the negative impact of the economic downturn. The individual character-
istics that increase the chances of remaining in employment emerge as being male,
no longer belonging to the young age group, and holding a high educational
qualification. Moreover, they find that full-time workers have a lower probability
of exiting employment. The same effect emerges for occupations requiring high
qualifications. With regard to outflows from inactivity, women appear less likely to
exit the nonlabour force. This finding indicates a strong discouragement effect for
women, especially at the beginning of their working career. Holding low quali-
fications and unfavourable labour market conditions inevitably exacerbate the
situation. All of these findings suggest a greater negative impact for disadvantaged
labour market groups, which leads to a widening of labour market duality. The
authors discuss changes in government labour policy promoted to fight the neg-
ative effects of the economic crisis on labour market outcomes.
1.5 Conclusions
The contributions to the book have emphasised the circular aspects of social
exclusion. Factors that generate social exclusion are themselves an effect of social
exclusion. Groups that are traditionally considered fragile, such as disabled people
and formerly convicted people, are particularly vulnerable to social exclusion.

Articles on the 20082009 economic downturn make clear that regarding social
exclusion, the crisis has affected a much wider range of people, including youths,
low-skilled workers and immigrants, in terms of income, job loss, job finding, even
participation. The result is a widening of socio-economic duality. The relevance of
definitions of socio-economic outcomes for measurement and evaluation of the
crisis have also been stressed, and their consequences for interventions of poli-
cymakers in fighting socio-economic effects of the economic downturn.
It has been said above that social exclusion is both an indicator of discomfort
for the people who suffer it and a threat to social cohesion, so policy interventions
to address social exclusion are required, on both grounds. However, both the
situation of specific disadvantaged groups and the asymmetric effect of the crisis
open questions about the effectiveness of the welfare system in addressing and
mitigating social exclusion. Given these premises, a revision of the welfare
structure should be considered, taking into account, for example, the changes in
labour market institutions introduced in the last few decades.
Finally, it seems desirable that the European Union should impose to member
countries parameters to be respected, and not only targets and guidelines, in terms
of social indicators, so that countries too remote, for instance, from the Lisbon
(original and revised) targets may find themselves under threat of sanctions.
References
Adato M, Carter MR, May J (2004) Sense in sociability? social exclusion and persistent poverty
in South Africa, University of Wisconsin-Madison, Department of Economics and Agricul-
tural Economics, Staff paper, No. 477
Bentolila S, Cahuc P, Dolado JJ, Le Barbanchon T (2010) Two tier labour markets in the great
recession: France vs. Spain. Paper presented at the XXV simposio de la asociacin espanola
de economia, Madrid, December 2010
Brandolini A, Cipollone P, Viviano E (2006) Does the ILO definition capture all unemployment?
J Eur Econ Assoc 4(1):153179
Chakravarty SR, DAmbrosio C (2003) The measurement of social exclusion, DIW, German
Institute for economic research, Berlin, Discussion papers 364
Chakravarty SR, Zoli C (2009) Social exclusion orderings, Working paper series, University of
Verona
European Commission (2010) Combating poverty and social exclusion: a statistical portrait of the
European union 2010. Eurostat statistical books, http://ec.europa.eu/social
EUROSTAT (2010) Official statistics. http://ec.europa.eu/eurostat
ISTAT (2011) www.istat.it
Leschke J, Watt A (2010) How do institutions affect the labour market adjustment to the
economic crisis in different EU countries, ET UI working paper no. 2010-04, European trade
union institute
Nolan B, Whelan CT (2010) Using non monetary deprivation indicators to analyze poverty and
social exclusion: lessons from Europe? J Policy Anal Manage 29(2):305325
OECD (2009) Employment outlook: tackling the jobs crisis. Paris
OECD (2010) Employment outlook: moving beyond the jobs crisis, Paris
Part I
The Structural and Long Term
Causes of Social Exclusion
Chapter 2
Health, Lifestyle and Growth
Gianluigi Coppola
Abstract In this article, I attempt to explain why lifestyle may have a positive
impact on economic growth. First, I consider the ways in which health affects a
consumers utility, and I then define a Health Production Function for which health
is the output and consumer good is the input. In this approach, the Lifestyle Return
to Scale (LRS) parameter is defined. The first result is that an increase in a
consumers personal income may have a positive or a negative effect on health.
That is, health may be a normal or an inferior good, depending on the Lifestyle
Return to Scale value. According to this result, I compute a health multiplier and
then modify the Solow Growth Model in which health is labour-augmenting. The
result is a model in which the Lifestyle Return to Scale positively affects per capita
income and per capita income growth.
Keywords Health Lifestyles Growth

JEL Classification I10 O40
2.1 Introduction
At the macro level, stylized facts indicate substantial differences in per capita
income and health status among countries and regions. This may imply that low
per capita incomes negatively affect health and vice versa.
G. Coppola (&)
Dipartimento di Scienze Economiche e Statistiche,
Universita Degli Studi Di Salerno and C.E.L.P.E, Salerno, Italy
e-mail: glcoppola@unisa.it
18 G. Coppola
In the last 20 years, the literature on economic growth has focused primarily on
the role of human capital accumulation, while health has occupied only a marginal
role in economic analyses. Second, if the income and health differences among
countries are significant, those among regions are even stronger and are essential
to economic growth.
The literature on health and health economic growth has not considered the
impacts that lifestyle has on Economic Growth. The aim of this theoretical paper is
to develop the relationship between Health and Growth by accounting for con-
sumer lifestyles.
Contoyannis and Joness hypothesis (2004) introduced a micro-model of con-
sumer choice to define lifestyles more accurately and to explain the effects of
consumers choices on Health. At the macro level, Weil (2005) asked if the forces
driving differences among regions were primarily derived from health or income.
In this context, I provide an answer by computing a health multiplier (Sect. 2.2)
and developing a simple modified Solow growth model in which health is labour-
augmenting (Sect. 2.3). Thus, this model includes the relationship among income,
lifestyle and health status first obtained at the micro level.
The first important result is that an increase in a consumers personal income
may have a positive or a negative effect on health if the consumer has a good or
a bad lifestyle, respectively. At the macro level another result of the model is
that lifestyle may be crucial for growth: a good lifestyle can generate a positive
impact on economic growth, while a bad lifestyle may also negatively affect
growth. The model also explains why health improvement positively affects
income, while increasing income may have a lower effect on health (Weil 2005).
2.2 Some Empirical Evidence
The empirical evidence can be divided into two categories: Long-Run stylized
facts and facts related to governments measures against the unhealthy habits of
the citizens.
The long-run stylized facts show increases in per capita income and life
expectancy. From 1820 to 2001, the per capita World GDP grew from $667 to
$5.709 (Maddison 2003). In the same period, world life expectancy at birth
increased from 28.5 years in 1820 to 65.2 years in 1990.
The differences among regions, in terms of both per capita GDP and life
expectancy at birth, have also increased (Riley 2005). In 1998, the per capita GDP
of the United States was 20 times that of Africa, while in 1820, it was only 3 times
larger.
In the period from 1800 to 1820, the differences among regions, in terms of life
expectancy at birth, were relatively low compared to the period from 1990 to 2001.
In the first period, the difference between the highest and the lowest regional life
expectancies was only 9.2 years. This gap widened to 26.3 years in the period
from 1990 to 2001 (Riley 2005).
2 Health, Lifestyle and Growth 19
These disparities may be explained by differences in Public Health, Medical

diagnostics and care, diet and, more generally, lifestyle.
As regards the second aspect, the governments measures, it exists a lot of
evidence of the importance that peoples lifestyle assumes.
For example, in England, many local governments offer incentives to encourage
consumers to lead healthier lifestyles. In Dundee, smokers are offered 12.50 a
week by the NHS if carbon monoxide testing shows that they have quit. In Essex,
pregnant women can claim a 20 food voucher from the NHS when they stop
smoking for one week, 40 after four weeks and another 40 after one year.
Brighton offers children 15 to stop smoking for 28 days. Overweight patients in
Kent are offered incentives for losing weight. In the US and other countries,
incentives are offered for weight loss, compliance with diabetes treatments, or
regularly testing negative for sexually transmitted diseases.1
Moreover, in Japan, a national law against obesity came into effect in 2008.
Under this law companies and local governments must measure the waistlines of
Japanese people between the ages of 40 and 74 during annual checkups.
In Italy, phrases warning consumers of the damage caused by smoking are
printed on cigarette packs, which are sold by a state monopoly. On some packages,
the following sentences appear: Smoking while pregnant harms your baby;
smoking kills.
These examples demonstrate government concern with lifestyles and the
implementation of policies to change consumption habits. In so doing, many
governments heavily influence individuals choices.
What determines this intrusive government interest? By reasoning backward,
government interest produces a conflict between individual and social choices that
originates in two factors: the rising cost of health care in many Western nations
and the assumption that bad lifestyles may negatively affect labour productivity.
Zargosky (2005) demonstrated a large negative association between Body Mass
Index (BMI) and White females net worth, a smaller negative association for
Black women and White males and no relationship for Black males in the US.
He also found that individuals who lose small amounts of weight experience small
changes in their net worth, but those who lose large amounts of weight have
improved financial positions.
For households in Sierra Leone, Strauss (1986) found a highly significant
relationship between caloric intake and labour productivity, providing solid sup-
port for the nutrition-productivity hypothesis. The marginal effect on productivity
decreases drastically as calorie consumption increases but remains positive at
moderately high levels of intake. One result from this situation is a decrease in the
effective price of food that is more significant for households that consume fewer
calories.
In general, it can be argued that health has a positive effect on the labour
productivity of individuals. Thus, lifestyle choices, such as smoking and drinking,
1
Financial Times Cash incentives seen as helping nations health, 11 April 2009.
20 G. Coppola
of individuals are of interest to society and, also to firms because they affect labour
productivity.
Ultimately, lifestyle choices generate externalities, a term that indicates pos-
sible conflicts that have not been resolved by the market. Externalities affect labour
productivity and healthcare costs above and beyond the level for which companies
are responsible. This relatively new concept of externalities is explained by Sassi
and Hurst (2008):
Lifestyle choices, as many other forms of consumption, may produce external
effects. There are immediate externalities that derive directly from acts of lifestyle con-
sumption, such as passive smoking, violent and disorderly behaviour associated with
alcohol abuse, or traffic accidents resulting from reckless driving. There are also deferred
externalities, which are generated through the link between lifestyle choices and chronic
diseases. Once chronic diseases emerge, and in some cases even before they emerge (e.g.,
when important risk factors such as hypertension or obesity begin to manifest themselves),
the individuals affected will become less productive, possibly entirely unproductive, they
will make a more intensive use of medical and social services, which may be publicly
funded, they may require care by members of the family and friends. Conversely, a
reduced life expectancy may mean a less prolonged use of publicly funded medical and
social services at the end of life, as well as reduced pension payments, which are not
themselves externalities, but would translate into a less onerous fiscal burden and therefore
less distortion in the way the economy works. All of these phenomena involve external-
ities (negative the former, positive the latter) on society at large, family and friends, which
can be attributed at least to some extent to the lifestyle choices originally made by the
individual. The extent to which externalities can be associated with lifestyle choices
depends, of course, on the strength of the link between lifestyles and disease, i.e., by the
increase in the risk of developing a chronic disease associated with adopting a particular
lifestyle.
These arguments are not new. John Stuart Mill (1859) wrote the book On
Liberty to fight against laws that would limit individual freedom. In the nineteenth
century in Great Britain, social degradation phenomena such as alcoholism were
prevalent. Several social movements asked the government to implement prohi-
bitionist measures to halt these phenomena. Mill contrasted these ideas,
arguing that
what happens inside a persons body or mind is that persons private business, not the
business of society and certainly not the business of the government over himself,
over his own body and mind, the individual is sovereign.
2.3 A Micro-Model
In the following section, I define lifestyle and develop a micro-funded model that
explains the relationship between health and income and the effects of income on
health.
First, let us suppose that an economy produces 3 goods: 2 commodities for

consumption, x and z, and Capital, K. The saving rate, s, is exogenous and
constant.
According to the Grossman model (1972), health capital and the demand for
health have been widely modelled in the economic literature.
Contoyannis and Jones (2004) developed a static model of lifestyle and health
production. In this model, (1) income is assumed to be endogenous, but there is no
direct influence of lifestyle or health on wages; (2) health affects consumer utility,
unlike Grossmans dynamic model (1972) in which health is considered to be a
stock that produces flows of pecuniary and non-pecuniary benefits as a result of
investments; (3) health is a result of a production function in which the inputs are a
vector of goods, a vector of exogenous influences on health and a vector of
unobservable influences on health; and (4) the money budget constraint and the
time constraint close the model. As a result of these conditions, the following are
maximised: consumer utility, using a Lagrangian function, the Marshallian
demand for goods and the level of consumer Health.
Contoyannis and Jones (2004) set the Health Production Function (HPF)
equal to
H hC; XU ; UH 2:1
where H is a measure of individual health; C is a vector of M goods; XU is a vector
of exogenous variables that influence health; and UH is a vector of unobservable
influence on health.
In this article, the model of Contoyannis and Jones (2004) is augmented to
produce a model with 2 equations: (1) the consumer utility function and (2) the
health production function.
2.3.1 The Consumer Utility Function
The consumer utility function is assumed to be a Cobb Douglas function in which

health, h, is an input. For this reason, health affects the consumer utility function
(among others, Kip Viscussi and Evans 1990). The other 2 inputs are the com-
modities x and z. The utility function can be written as
U h; x; z ha xb zd 2:2
Where a; b and dare the elasticities of h, x and z, respectively.

Equation a 0 may be considered as the self-assigned weight of a consumer to
his/her own health. If a 0; health is not important to the consumer. On the
contrary, if a [ 0; then health is important.
b; d \
[ 0: If b\0 or d\0; x or z is not a good but a bad for the consumer
(Varian, 1992).
22 G. Coppola
This is a static model. There is no rational addiction, but positive elasticity

values indicate that a consumer knows the commoditys ophelimity.
2.3.2 The Health Production Function
Contoyannis and Jones (2004) assume that the utility of a consumer depends on a
set of goods C, Health H, XU ; a vector of observable exogenous influences on
U and lU , a vector of unobservable influences or U, defined by
U uC; H; XU ; lU 2:3
Furthermore, HPF depends on a set of goods C, a vector of observable exog-
enous influences on H, XH ; and a vector of observable exogenous influences on H,
lH : The vector H is defined by
H hC; XH ; lH 2:4
It is assumed that the consumption of a commodity can improve, worsen or
have no effect on the health of a consumer. Consumption can worsen health in the
case of smoking, alcohol and drugs.
For simplicity, let us assume that every commodity can only better or
worsen the health of a consumer. Further, no commodities can positively impact
health in small quantities and negatively impact health in stronger doses.2 It is also
assumed that x improves health, while z worsens health. The commodity x may be
defined as the virtuous, or sustainable, good, and z may be defined as the harmful
good.
Health also depends on the initial level of health h0 ; public health W; time
t and a stochastic component e: The Health Production Function is
hx; z; h0 ; w; t; e xq zc h0 we/t ee 2:5

The function can be split into two parts: xq zc can be interpreted as a con-
sumers activity, and h0 we/t ee can be attributed to other factors. To simplify the
model, we use the relation X h0 we/t ee : The HPF is then
hx; z; h0 ; w; t; e Xxq zc 2:6
In this health production function, there is one input, x, with a positive mar-
ginal productivity. This assumption is compatible with the neoclassic production
function theory (Gravelle and Rees 1992). The term q c is equal to the elas-
ticity of scale and can be positive, negative or null. Let h q c: Each input is
2
The ancient Romans said In Medio stat Virtus. In the model that hypothesis doesnt matter
for each single good.
assumed to exhibit a decreasing return; thus, 0\q\1; 0\c\1; and, therefore,

1\h\1:
Sassi and Hurst (2008) related individual lifestyles to individual behaviours that
affect health. Contoyannis and Jones (2004) also defined lifestyle as a set of
behaviours which are considered to influence health and are generally considered
to involve a considerable amount of free choice.
If h [ 0; consumption increases positively affect health, while for h\0; con-
sumption increases negatively affect health. With h 0; consumer behaviour has
no effect on health. For this reason, the parameter h may be defined as the Lifestyle
Return to Scale (hereafter LRS).
Before maximising the consumer utility function, hx; z; h0 ; w; t; e Xxq zc is
substituted into U h; x; z ha xb zd to yield
U h; x; z Xxaq zac xb zd 2:7
or
U h; x; z Xxaqb zdac 2:8

The elasticity with respect to x becomes aq b; and the elasticity with respect
to z becomes d ac:
The commodity x (or z) will be consumed only if aq b [ 0 (or d ac [ 0;)
that is, if the relative elasticity is positive. The other properties of the utility
function are the same as before.
Hence, the choice to consume x (or z) depends on 3 parameters: (1) the elas-
ticity of utility with respect to x (or z) b (or d), that is, the weight conferred by the
consumer to the commodity x (or z); (2) a, the importance of health to the con-
sumer; and (3) the benefit (or damage) of x (or z) on health b (or c ).
For example, an individual will consume a medicinal (x) only if the positive
health impact aq is given a value greater than the elasticity with respect to x (b).
On the contrary, a consumer will not drink alcohol if alcohol is not preferred
d\0 or if alcohol is enjoyed d [ 0but he assumes the negative impacts on
health ac to be greater than the elasticity of alcohols utility d ac [ 0:
A consumer may decide to drink alcohol even if the dangerous health effects are
known. Further, consumers who are aware of the damage of smoking may con-
tinue to smoke.3 Following this approach, alcohol consumption depends also on
factors other than the level of a consumers education.
Generally, by including health in the consumer utility function, the consump-
tion of commodities that benefit health increases, while the consumption of goods
that cause damage decreases because a [ 0.
3
See for example Berger and Leigh (1989) and Kenkel (1991) for the relationship between
schooling and health. See also Avitabile (2009) for the relationship between health and
information.
24 G. Coppola
2.3.3 The Utility Maximization Problem: The Optimal

Choice of x, z and h
Let X 1: The consumers budget constraint is px x pz z cy; where px ; pz are

the prices of the goods; y is the per capita income used for consumption, y YL ;
c is the average propensity to consume 0\c\1; and L is the population. The
consumer maximizes utility when maxx;z xaqb zdac such that px x pz z y.4
Recall that aq b [ 0 and d ac [ 0:
Optimally solving the Lagrangian maxx;z L Ux; z kpx x pz z y; where
k is the Langrage Multiplier, the quantities of commodities consumed are
(Mas-Colell et al. 1995)
aq b cy
x 2:9
b d a q c px
d ac cy
z 2:10
b d a q c pz
The weight of health, a; increases the consumption of virtuous goods and
reduces the consumption of harmful goods. Optimally, the health level is

aq b cy q d ac cy c
h 2:11
b d a q c px b d aq c pz
or
q c

aq b d ac pz c
h cyqc 2:12
b d a q c px q
b d a q c
q
aqb
Equation 2.11 is the health demand function, where bda qc and
c
dac
bdaqc are the shares of good x and good z, respectively, weighted for
their relative health elasticities.
The level of health and the price of virtuous good are negatively correlated.
If the price of good x increases (or decreases), it worsens (or improves) the level of
health. Conversely, health improves (or worsens) if the price of z increases
(or decreases).
The health elasticity with respect to income is q c h; the parameter LRS.
Unlike the parameters that can have only one sign, this parameter may be positive
or negative. If q c 0; income growth does not affect the level of health.
If q c\0; income negatively affects health. If q c [ 0; income positively
affects health.
4
This approach may be considered as a generalization of Wagstaffs model (1986). See
Appendix for details.
Thus, income growth does not always positively affect health. The sign and the
degree to which income affects health depend on the parameter h:
A proxy or Index of a consumers Lifestyle (LI) may be given by the weighted
average of the quantity of commodities consumed for the consumers health
elasticity. This variable follows the relationship 1\LI\1 and is given by

aq b d ac
LI q c 2:13
b d a q c b d a q c
Therefore, LRS h is a crucial variable in the model because it indicates the
attitude of a consumer, based on preferences and opportunities, toward leading a
particular lifestyle. Thus, h q c partially and indirectly reflects consumer
preferences because the health production function contains only those com-
modities that consumers prefer or can purchase.5
2.4 Comparative Static: The Health Multiplier
In the previous section, the effects of income on health were described. Assuming
the existence of a representative agent, Eq. 2.11 can be rewritten as
h
Y
ht 2:14
L
q c c
aqb dac pZ
where m bda qc bdaqc px q
c
A production function with a constant return to scale and for which both
technology and health are labour augmenting is assumed. This may be a Cobb
Douglas Production Function (i.e. Weil 2005; Sala-i-Martin 2005).
Y K a AhL1a 2:15
or
5
Three issues should be highlighted here. First, in this simple consumer model, choices are
made between two commodities. In reality, a commodity may be not consumed for three reasons,
the first two of which were outlined previously: (1) the consumer does not like a commodity;
(2) even if a commodity is liked, the health damage caused by the commodity may be greater than
the commoditys utility, preventing consumption of the commodity; and (3) the relative price of a
commodity may be greater than income, preventing consumption of the commodity. In the first
two cases, the commodity is not consumed as a result of free choice. In the second case, this
choice may be difficult. In the third case, price and income limits restrict access to the
commodity. In this article, we consider only the case in which individuals consume both
commodities.
26 G. Coppola
a
Y K
Ah1a 2:16
L L
From the system given by Eqs. 2.14 and 2.15, the impacts of a health shock
Dm and an income shock DA on health and income can be quantified.
Solving this system yields the effects in terms of elasticity. The results are
reported in Scheme 1 for 0\h\1:
Scheme 1
On health On income
1\h\1
Health shock d log h 1 d log Y 1 a

d log m 1 h1 a d log m 1 h1 a
Income shock d log h h1 a d log Y 1 a

d log A 1 h1 a d log A 1 h1 a
1
In terms of elasticity, the health multiplier is equal to 1h1a for health and
1a
1h1a for income.
For 0\h\1; both multipliers are positive and greater than one. If a lifestyle is
positive, a health shock will more strongly affect health and the growth of labour
productivity.
The effect of a health shock on income depends positively on both LRS and
income elasticity, with respect to labour or labour income share.
h1a
The effect of an income shock on income is equal to 1h 1a ; while the effect
1a
on health is equal to 1h 1a : In this case, for 0\h\1; both effects are greater
than zero.
The question of whether a health shock more strongly affects income or an
income shock more strongly affects health (Weil 2004) can now be answered: both
technological and health shocks produce the same effect on income. This effect is
1a 1
equal to 1h 1a ; which is greater than 1 a because 1h1a [ 1:
A health shock has an impact on health equal to 1h11a ; which is greater than
1a
an income shock 1h1a because 0\h\1: A health shock has an impact on
1a h1a
income equal to 1h1a ; which is greater than
an income shock on health 1h 1a :
The effect that a reduction of the price of commodity x has on health can be
quantified. The multipliers for income and health are
d log h d log h d log t 1
q 2:17
d log px d log t d log px 1 h 1 a
d log Y d log Y d log t 1 a

q 2:18
d log px d log t d log px 1 h 1 a
The effect of a price shock for x on health is equal to 1h11a q: It depends
1
both on the multiplier 1h1a and negatively on the parameter q; which is the
1a
health elasticity of x. The impact on income is 1h 1a q; which is the product of
the impact of the price of x on health, 1h11a q; and the impact of health on
income, 1 a. This impact has a negative sign because a reduction in the price
of x positively affects health and income.
Scheme 2 lists these health multipliers.
Scheme 2
On health On income
Health shock dh 1 dY 1 a Y
Yh
dm 1 h1 a dm 1 h1 a t
Income shock dh h1 a h dY 1 a Y

dA 1 h1 a A dA 1 h1 a A
2.5 A Growth Model with Health
In the comparative static framework, the level of population is constant. In this

section, I present a Growth Model that includes health as an input factor and also
considers lifestyle.
Interactions between health and economic growth are complex (Morand 2005).
Listing all of these interactions is beyond the scope of this paper. One of the positive
effects of economic growth on health is the possibility of consuming higher
quantities of better quality goods. Another consequence of economic growth is that
technological progress positively impacts medical technology and care.
Therefore, health may have a positive impact on Growth through several
channels. First, health may positively affect labour productivity (Marshall 2006).
In Chapter V, titled The Health and Strength of the Population, in Book 4 of the
Principles of Economy, Marshall wrote that
strength, physical, mental and moral, are the basis of industrial efficiency, on which the
production of material wealth depends; while conversely the chief importance of material
wealth lies in the fact that, when wisely used, it increases the health and strength, physical,
mental and moral, of the human race.
For Streeten (1994), one of the reasons for promoting human development is
that a well-nourished, healthy, educated, skilled and alert labour force is the most
important productive asset.
28 G. Coppola
In the Neoclassic Economic Growth literature, many models consider health as

a factor of growth. Lpez-Casasnovas et al. (2005) and Rivera and Currais (1999a)
used a conditional convergence regression in which the growth of per capita
income is a function of steady-state determinants. Assuming that health is an
important determinant of an enhanced labour force, they showed that health affects
income growth both positively and significantly. In another article (Rivera and
Currais 1999b), investment in health significantly explained variations in the
output of human capital, even in countries with presumably high levels of health.
Heshmati (2001) extended the model of Mankiw et al. (1992) by incorporating
health. The results show that Health Care Expenditures positively affect economic
growth and the speed of convergence.
For Morand (2005), increasing longevity may incentivise agents to increase
investments in capital and human capital and thereby reinforce economic growth.
One of the main goals of this article is to consider the effects of individual
lifestyles on economic growth. Let us now consider a Solow Growth Model
(Solow 1956) with a constant saving rate (s), diminishing returns of capital
0\a\1 and labour, Labour-augmenting technology and constant returns to
scale. We assume a Cobb Douglas production function:
Y t K ta AtLt1a 2:19
where K(t), A(t), L(t) are capital, level of technology, and labour, respectively.
Let us assume all the hypotheses of the Solows Growth Model.
Technological progress and the population growth rate are exogenous and
d ln At d ln Lt
constant: g; n:
dt dt
Assuming that health is a labour-augmenting factor (Weil 2005; Sala-i-Martin
2005), the production function becomes
Y t K ta AthtLt1a 2:20
With
h
Y
h
h ty t h tY h Lh ; 2:21
L
this becomes
1a
Y t K ta AttY th Lt1h 2:22
or
a
1a
Y t K t1h1a AttLt1h
1h1a
2:23
Equation 2.23 can be rewritten as

a
1 1
1h1a
1h1a
Y t K t1h1a At1h t1h Lt 2:24
a1h1a
Equation 2.24 shows a constant return to scale because 1h1a 1.
a 1h1a
Substituting a1 1h1a and a2 1h1a 1 a1 yields
1 1
1a1
Y t K ta1 At1h t1h Lt 2:25
This is a Solows Model, and the new technological rate is

1
d ln A2 t d ln At1h 1
dt
dt 1h g
At equilibrium, the Income growth rate and per capita income growth rate are

d ln Yt 1
gn 2:26
dt 1h

d ln Yt
Lt 1
g 2:27
dt 1h
The income level is
a
Y t 1 1 s 1h1a
A0v1h e1hgt g 2:28
Lt 1h n
The Health growth and Health level are

d ln ht d ln Yt
Lt 1
h h g 2:29
dt dt 1h
1hah1a
1 h h s
h t t A 0 e
1h 1h 1hgt 2:30
g
1h n
The first result is that even if the Solow model with health remains an exog-
enous growth model, the parameter LRS positively affects per capita income
growth and the level of income per capita at the steady state.
For example, if h 0:5; the income growth rate is equal to 2g; while a negative
LRS h\0 results in a per capita income growth rate that is less than that of
technological progress. Thus, a good lifestyle can improve economic growth,
while a bad lifestyle can slow growth.
The second result is that health increases more slowly than per capita income.
Unlike economic growth, the rate of health growth may be positive, null or neg-
ative, depending on the sign of the LRS parameters. For h\0; economic growth
negatively affects health, which worsens.
30 G. Coppola
Fig. 2.1 The effect of LRS h

on the growth rate of per
capita income with a constant
technological growth rate
(g = 2%)
The second scenario may be the case of a health-poverty trap. Such a trap
was noted in Russia; male life expectancy, which can be considered a health
indicator, plummeted by 7 years from 1989 to 1994 because of high levels of
alcohol consumption (UNDP 2010).
The results of the model can be explained differently. Per capita income growth
depends on the product of the technical progress parameter and LRS, both of
which are exogenous.
Technical progress is considered to be manna from heaven. In fact, the aim
of the endogenous growth theory is to identify those factors and mechanisms that
could be controlled by the government to ensure higher and more durable eco-
nomic growth.
In the model presented in this article, another exogenous parameter, the Life-
style Return to Scale, h, impacts economic growth. In the introduction, we pre-
sented several cases in which governments have attempted to control lifestyles.
This type of governmental behaviour has several possible explanations.
Equation 2.27 can be re-written as
g
h 1 Yt 2:31
d ln Lt
dt
Suppose that the government has established a target for its economic growth
rate, denoted by ^y ; that can be controlled by technological progress g. If ^y [ g;
the economic growth rate fixed by the government is greater than technical pro-
gress, and the government can strive for higher growth by controlling or trying to
modify lifestyles, which are one of the channels that transmit the effects of eco-
nomic growth to health. Conversely, if technological progress is high, lifestyles
may not be important because technological progress can ensure a high level of
economic growth.
Further, health growth is equal to economic growth multiplied the parameter h.
As shown in Fig. 2.1, with a fixed a technological rate g, the effects of economic
growth on health improvement depend on the value of h. Lower values of h, result
in weaker links between economic production and health improvements. This
relationship could become negative for h \ 0. Hence, economic growth is a
necessary but not sufficient condition for improving health. If a governments

priority is to improve health rather than economic growth, then LFS must be
positive. In this scenario, the conflict between public and private interests can
become stronger.
2.6 Conclusions
In this article, I attempted to formalize what Jean Anthelme Brillat-Savarin, the

author of Philosologie of taste Brillat-Savarin 2004, wrote two hundred years ago:
(1) animals feed themselves; men eat, but only wise men know the art of eating,
and (2) the destiny of nations depends on the manner in which they are fed.
The crucial hypotheses of the model are that (1) individuals can rationally
choose to consume goods that negatively affect health, (2) individuals are
co-producers of their health and (3) health positively affects labour productivity.
First, I developed a consumer micro-model with health and two goods, both of
which are positively correlated to the Consumers Utility. Health is the output of a
consumers production function with the two commodities as inputs. The first
commodity has a positive impact on health, while the second one has a negative
impact.
The result is that the elasticity of consumer health with respect to income,
referred to as Lifestyle Return to Scale and denoted by the parameter h is equal to
the algebraic sum of the health elasticity with respect to commodities. It may be
positive, negative or neutral. In opposition to healths role as a normal good, as
reported by Wagstaff (1986), health can also be an inferior good in this model.
Second, I computed health multipliers. The impacts of a health shock on health
and income depend on labour share and are higher if the Lifestyle Return to Scale
is positive.
Third, the micro-behaviour function was introduced in the Solow growth model
in which the return to scale is constant. The most important results are that (1)
Lifestyle Return to Scale affects economic growth (the growth of income per
capita is higher than the technical progress if LRS is positive), and (2) health
improvement depends on the parameter LRS, h, it is lower than economic growth,
and it may be negative, even if economic growth is positive. In fact, the existence
of a health poverty trap in which economic growth diminishes health can be
demonstrated.
In conclusion, lifestyle is another aspect of society that governments can
attempt to control or regulate.For this reason, the Aristotelian concept of inter-
mediates is useful in the definition of virtue. In Nicomachean Ethics (Aristotle
2009), Aristotle explains that
drink or food which is above or below a certain amount destroys the health, while that
which is proportionate both produces and increases and preserves it (Nicomachean Ethics,
Book 2, Chap. 3).
32 G. Coppola
The absence of this notion of intermediate in individual behaviours negatively

affects society, and possible responses may be social exclusion and government
interventions in the private sphere, as condemned by Mill (1859), the intensities of
which may vary until an ethical Hegelian state is established.
Acknowledgements I would like to thank seminar participants at the University of Salerno and
the University of Pescara. I am grateful for many helpful comments, received during the course of
those presentations. Thanks also to Adalgiso Amendola, Alberto Bennardo, Dimitrios Christelis,
Floro Ernesto Caroleo, Marcello DAmato, Sergio Destefanis, Fernanda Mazzotta, Niall OHi-
gghins, Carmen Pagliari and Giuliana Parodi for their very useful suggestions. The usual dis-
claimer applies.
A.1 Appendix
A.1.1 A Generalization of Wagstaffs Model
Starting with Michael Grossmans Model (1972) and Wagstaff (1986) developed a
one-period model of demand for health. The four hypotheses of the model include
the following: (1) an individuals health is determined by the consumption of
health inputs h x xq ; (2) preferences are non-lexicographic: individuals desire
health but not above everything else; (3) individuals also consume other com-
modities that have a positive cost for consumers, so U uh; z with dU dU
dh ; dz [ 0
2 2
and d Udhh;z ; d Udzh;z \0; and (4) consumers have limited economic resources or
budget constraints: px x pz z Y; where px and pz are the prices of commodities
x and z, respectively, and Y is the income.
Assuming a Cobb Douglas Utility function and a Health production function
h x xq ; the Wagstaff Model can be formulated with the following formulas:
U h; z ha zd 2:32
h x x q 2:33
px x pz z Y 2:34
where 0\a\1 and 0\d\1 are the utility elasticities with respect to x and z,
respectively, and 0\q\1 is the elasticity of h with respect to x.
This is a special case of the Consumers model (Sect. 2) with b 0: The
commodity x is not in the Consumers utility function with c 0; thus, z does
not affect health.
The solutions can be obtained from two different methods. The first was pro-
posed by Wagstaff:
max U h; z ha zd s:t: px hq pz z Y 2:35

h;z
In this case the Budget Constraint is not linear. The consumer chooses between
health and z. The second possible solution is
max U x; z xqa zd s:t: px x pz z Y 2:36

x;z
The consumer chooses the quantities of x and z that maximize utility.

Both methods yield the same solutions:
aq Y
x 2:37
d aq px
d Y
z 2:38
d aq pz

aq Y q
h 2:39
d aq px
The main differences include the following: (1) in the Wagstaff model, Health
can only be a normal good because dh dy [ 0 (conversely, in the model proposed in
this paper, Health may also be an inferior good), and (2) this result depends on the
lifestyle of the consumer.
References
Aristotle (2009) Nicomachean Ethics (trans: Ross WD). World Library Classical Books
Avitabile C (2009) Does conditionality matter for adults health? Evidence from a randomized
experiment. CSEF Working Paper 222
Berger MC, Leigh JP (1989) Schooling, self-selection, and health. J Human Res 24(3):433455
(Summer, 1989)
Brillat Savarin JA (2004) The phisiology of taste. Kessinger Publishing, US
Contoyannis P, Jones AM (2004) Socio-economic status, health and lifestyle. J Health Econ
23:965995
Gravelle H, Rees R (1992) Microeconomics, 2nd edn. Longman Group UK Limited, England
Grossman M (1972) On the concept of health capital and the demand for health. J Political Econ
80(2):223255
Heshmati A (2001) On the causality between GDP and health care expenditure in augmented
Solow growth model. SEE/EFI Working paper series in Economics and Finance 423
Kenkel D (1991) Health behaviour, health knowledge and schooling. J Political Econ 99:287305
Kip Viscussi W, Evans WN (1990) Utility function that depend on health status. Am Econ Rev
80(3):353374
Lpez-Casasnovas et al (2005) Health and economic growth: findings and policy implications
In: Lpez-Casasnovas G, Rivera B, Currais L (eds) Cambridge MIT press, Cambridge
Maddison A (2003) The world economy: historical Statistics. Paris Development Centre of the
Organisation for Economic Co operation and Development, Paris
Mankiw NG, Romer D, Weil N (1992) Contribution to the empirics of economic growth.
Quarterly Journal of Economics 107:407437
34 G. Coppola
Marshall A (2006) Principles of Economics Cosmio Inc. original published (1890) by Prometheus
Book
Mas-Colell A, Whinston M, Green J (1995) Microeconomic theory. Oxford University Press,
Oxford
Mill JS (1859) On liberty. Republished by Agora Publication inc (2003)
Morand OF (2005) Economic growth, health, and longevity in the very long term: facts and
mechanisms. In: Lpez-Casasnovas G, Rivera B, Currais L (eds) Health and economic
growth: findings and policy implications. Cambridge MIT press, Cambridge
Riley JC (2005) Estimates of regional and global life expectancy 18002001. Population Dev Rev
31(3):537543
Rivera B, Currais L (1999a) Economic growth and health: direct impact or reverse causation?
Appl Econ Lett 6(11):761764
Rivera B, Currais L (1999b) Income variation and health expenditure: evidence for OECD
countries. Rev Dev Econ 3(3):258267
Sala-i-Martin X (2005) On the healthpoverty trap. In: Lpez-Casasnovas G, Rivera B, Currais L
(eds) Health and economic growth: findings and policy implications. Cambridge MIT press,
Cambridge
Sassi F, Hurst J (2008) The prevention of lifestyle related chronic diseases: an economic
framework. OECD Publishing 32
Solow R (1956) A contribution to the theory of economic growth. Q J Econ 70:6594 (February)
Strauss J (1986) Does better nutrition raise farm productivity? J Political Econ 94(2):297320
Streeten P (1994) Human development: mean and ends. Am Econ Rev 84(2), In: Papers and
Proceedings of the 106th annual meeting of the American Economic Association, May 1994,
pp 232237
UNDP (2010) Human development report. Second printing Palgrave Macmillian, London
Varian H (1992) Microeconomic analysis, 3rd edn. Norton, New York
Wagstaff A (1986) The demand for health: a simplified Grossman model. Bull Econ Res 38:1
Weil DN (2004) Economic growth. Addison-Wesley, Reading
Weil DN (2005) Accounting for the effect of health on economic growth. NBER Working Paper
w1145
Zargosky J (2005) Health and wealth: the late-20th century obesity epidemic in the US. Econ
Human Biol 3(2):296313
Chapter 3
A Comparative Analysis of Literacy
Rate in Contributing to Social
Exclusion Insights
Edgardo Bucciarelli, Carmen Pagliari, Fabrizio Muratore

and Iacopo Odoardi
Abstract Our contribution aims to analyze the relationship between the phe-
nomena of social exclusion and literacy levels, and to consider the significant
implication of this relationship on economic growth. The goal, which is reached by
analyzing cross-country data, is firstly to describe the situation of social exclusion
with the use of specific socio-economic variables, and secondly to compare the
levels of education and training for each considered country. These two phe-
nomena are mutually influenced, as a low level of literacy in affecting the
employment status precludes the possibility to enter and operate freely in society,
while poverty and persistent social exclusion of a person or family make difficult
to address appropriate educational and training paths. Therefore, our study has
rejoined two issues which influence almost all decisions adopted by policy makers,
especially in the Western world. The opening issue is the level of education, which
should constitute the human capital of a country, through appropriate investment,
and the second one is the relational condition of social dynamics, which highlight
the so-called social capital. Together these two types of intangible capitals con-
stitute a strong support for the long-term development of a country. Our quanti-
tative analysis is also addressed to detect differences and peculiarities among the
E. Bucciarelli (&) C. Pagliari

Dipartimento di Metodi Quantitativi e Teoria Economica,
Universit dAnnunzio di Chieti-Pescara, Pescara, Italy
e-mail: e.bucciarelli@unich.it
C. Pagliari
e-mail: c.pagliari@unich.it
F. Muratore I. Odoardi
Scuola Superiore, Universit dAnnunzio di Chieti-Pescara, Pescara, Italy
e-mail: fabrizio.muratore@gmail.com
I. Odoardi
e-mail: iacopo.odoardi@alice.it
36 E. Bucciarelli et al.
different national realities, with the ultimate purpose to recognize which socio-
economic variables affect more directly the processes of education.
Keywords Social exclusion

Literacy rate Education Human capital

Maximum likelihood Hierarchical cluster
JEL Codes C82 I21 O50 O57 Y10
3.1 Introduction and Related Literature
The level of education for an individual represents, especially in modern Western

societies, a potential effective indicator of the level of working capacity, the so-
called productivity, but also implies other less observable, but equally significant
skills. These are the essential knowledge for living in society as dynamic players,
without incurring the risk of being excluded from the traditional and contemporary
activities of society. But today there are many limitations to the normal training of
an individual, and in some cases, to the groups of disadvantaged people, for
example, the problem of poverty which affects the most vulnerable groups such as
the children, the elderly, and the ethnic minorities. Children living in families at
social risk or below the poverty line see foreclosed the best channels of education
and training, and find obstacles in the first years of their training school. This
situation could affect childrens subsequent paths of training and prevent them
from achieving adequate levels of education. For adequate levels of education we
mean the level of training which is considered optimal from the perspective of
each culture and society in order to access to stable employment as well as to
participate in society. However, if households living below the poverty line are the
most at risk, we must remember those with only one parent, those who live in
neighbourhoods with little social organization, and even those belonging to ethnic
or religious minority. Moreover, the strong positive correlation between illiteracy
and social exclusion is underlined by the wide set of definitions of literacy which
have been given over time. An example is provided by the General Conference of
UNESCO 1978 which provides several definitions of literacy. The first one states
that [] a person is literate who can with understanding both read and write a
short simple statement on his everyday life, while the second one is more
complex: [] a person is functionally literate who can engage in all those
activities in which literacy is required for effective functioning of his group and
community and also for enabling him to continue to use reading, writing and
calculation for his own and the communitys development. The two definitions
range from the simple ability of writing and reading to the reasoning skills of
adequately using the knowledge acquired. We are not referring merely to the
increasing of more cognitive abilities which allow to reach higher levels of
understanding, but also to the manner in which those who have these skills can
3 A Comparative Analysis of Literacy 37
exploit them to coexist in the stable and balanced society they belong to. This
particularly includes the definition offered by the International Adult Literacy
Survey (see Statistics Canada and OECD 2000): The ability to understand and
employ printed information in daily activities, at home, at work and in the com-
munityto achieve ones goals, and to develop ones knowledge and potential.
The definition used by OECD (2010) adds that [] differences in levels of
literacy matter both economically and socially: literacy affects, inter alia, labour
flexibility and quality, employment, training opportunities, income from work and
wider participation in civic society. Furthermore, the IALS Final Report (Sta-
tistics Canada and OECD 2000) shows a number of relationships about literacy:
countries with higher literacy scores have higher labour force participation and
shorter work hours whereas countries with a high proportion of adults with low
prose skills have lower GDP per capita. There is a correlation between a low GDP
per capita of a country and the rising proportion of adults with low prose skills.
Public intervention may partially cope and withstand the mentioned problems
related to the lack of literacy, by trying to obtain a plentiful and robust level of
national human capital in the long run. This is now a fundamental part of many
economic studies concerning growth and development. But the complex phenom-
enon of educating and training individuals cannot be regarded only as a general
educational teaching base of general educational foundations. This means not only
prepared trainers, and adequate public and private investment, but also a rational
project in order to develop in all individuals a sense of common rules through
education (see, for example, Grossman and Kim 1997). These are the foundations
for building a solid social capital which is itself a determinant of socio-economic
development processes (see, among others, Gradstein and Justman 2002, who have
also included the education processes in the relationship between social capital and
economic growth).
People who cannot reach satisfactory levels of literacy either by choice or
compulsion are the weakest in society and go toward the phenomenon of social
exclusion. The problems caused by poor literacy can be observed through two joint
events. The first one concerns the limited prospects of finding safeguarded
employment opportunities, the second one considers the defective contribution
which people offer to the socio-economic system through their labour productivity.
The difficulty in finding a job involves a certain separation from society. This
separation deprives individuals of their ability to fully exploit the possibilities
offered them by contemporary world. The individuals, though not in poverty, will
not have the opportunity to be engaged in specified conducts which are charac-
teristics of each society, as constitutional elements of general needs. Therefore,
this path is degenerative, and leads to conditions which are close to relative
poverty which tends to become a phenomenon that is passed between generations.
In fact, poor parents cannot guarantee the optimal education for their children (see,
among others, Grossmann 2008; Galor and Moav 2004), and socially excluded
parents represent a source of social exclusion for their children. The education
level of a person can be found either through the years of school attended, or
through the qualifications acquired. In particular, in this work, we observe the
influence which some typical variables, related to the processes of social exclu-
sion, can have on the average level of education of a country. It is known that the
economic conditions may affect the ability to achieve high average levels of
education, but also the strength of education and social capital constitute a lever of
development (see, among others, Temple and Johnson 1998). Gradstein and
Justman (2002) analyze the importance of a broad common cultural basis to start
those basic functions of effective interaction between individuals. The spread of
education, and, in general, of literacy may be a fundamental policy goal in trying
to establish virtuous processes from an economic standpoint. Essentially, the main
aim of our empirical study is to demonstrate how certain variables,1 which char-
acterize social exclusion, have obvious influences on the literacy rate. In order to
pursue this aim, we gather data concerning 30 countries and then we briefly
compare the effect of the above mentioned variables on the level of economic
development of the countries considered. In our particular case study, the literacy
rate represents the percentage of people with the ability to read and write without
specifying the level of education.
The framework of the paper is organized as follows. In Sect. 3.2, we start by
showing the methodology used. Then we proceed in Sect. 3.3 with a preliminary
empirical investigation on literacy rate, and on the expectancy variables.
In Sect. 3.4, the results of the multivariate regression model are presented. The
findings of the factor analysis with maximum likelihood method and VARIMAX
rotation are shown in Sect. 3.5. Furthermore, Sect. 3.6 focuses on multivari-
ate regression model for literacy rate with maximum likelihood components.
In Sect. 3.7, we represent and analyze the countries in a hierarchical cluster with
variables related to literacy. In Sect. 3.8 we focus our attention on specific
American literacy policies which are particularly significant for our study. We
present our conclusions in Sect. 3.9.
3.2 Methodology Applied for Empirical Analysis

on Literacy Rate for 30 World Countries
In this section we analyze literacy rate by using the least-squares method, the
factor analysis with maximum likelihood method and the hierarchical cluster. The
method of least squares is a standard approach to the approximate solution of over
determined systems (Moser 1996; Freund 2003), i.e. sets of equations in which
there are more equations than unknowns. Least squares means that the overall
solution minimizes the sum of the squares of the errors made in solving every
single equation. The most important application is in data fitting: the best fit in the
1
We have collected the cross-country dataset from UNESCO and World Bank. For the cross-
country analysis we provide, see also Levine and Renelt (1991, 1992); de Gregorio and
Lee (2002); Hoover and Perez (2004).
least-squares sense minimizes the sum of squared residuals, a residual being the
difference between an observed value and the value provided by a model (Bjrck
1996). We consider a linear regression model: hence, the model comprises a linear
combination of the parameters:
X
m
f xi ; b bj /j xi 3:1
j1
where the coefficient /j is functions of xi. Letting:
of xi ; b
Xij /j xi 3:2
obj
Then we can see in case (3.2) the least square estimate (or estimator, in the
context of a random sample), b is given by:
b X T X1 X T y 3:3
Consequently, we consider the response variable as a linear function of the
regressors:
yt b0 b1 x1 bk xk et 3:4
In our paper, we analyze literacy rate (LR) as a function of school enrolment
consisting in pre-primary (PRE), primary (PRI), secondary (SEC) and tertiary
(TER); GDP real growth rate (GDP); long term unemployment rate (LUR); public
spending on education as percentage of GDP (PSE) and public spending on
education as percentage of government expenditure (PGE); children out of school
(COS). In our analysis, we consider different levels of education: pre-primary,
primary, secondary, and tertiary education. These levels are generally found in
similar ways in many countries, and generally go from kindergarten to the so
called higher education. The analysis regard 30 countries2 from different conti-
nents, and we consider mostly countries of Organizations for Economic Co-
operation and Development (OECD). The model we provide is the following:
LR b0 b1 PRE b2 PRI b3 SEC b4 TER b5 GDP b6 LUR b7 PSE

b8 PGE b9 COS 3:5
Therefore, in the next study we apply factor analysis in order to reduce the high
number of variables of model (3.5). Factor analysis is a statistical method used to
2
The countries considered by this research are: Australia, Austria, Belgium, Bulgaria, Canada,
Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy,
Japan, Republic of Korea, Latvia, Luxembourg, Mexico, Netherlands, New Zealand, Norway,
Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Turkey, United Kingdom, United
States.
describe variability among observed variables in terms of a potentially lower

number of unobserved variables called factors (Mardia and Kent 1980). In other
words, it is possible, for example, that variations in three or four observed variables
mainly reflect the variations in a single unobserved variable, or in a reduced number
of unobserved variables. Factor analysis searches for such joint variations in
response to unobserved latent variables. The observed variables are modeled as
linear combinations of the potential factors, plus an error term. The information
gained about the interdependencies between observed variables can be used later to
reduce the set of variables in a dataset. Factor analysis is related to principal
component analysis (PCA) but it is not identical, because PCA performs a variance-
maximizing rotation of the variable space and it takes into account all variability in
the variables. In contrast, factor analysis estimates how much of the variability is
due to common factors (communality). The two methods become essentially
equivalent when the error terms in the factor analysis model (the variability not
explained by common factors, see below) can assume the constant variance (Shore
and Sheppard 1996). In fact, PCA involves a mathematical procedure that trans-
forms a number of possibly correlated variables into a smaller number of uncor-
related variables called principal components. The first principal component
accounts for as much as possible of the variability in the data, and each succeeding
component accounts for as much as possible of the remaining variability. PCA is
mathematically defined as an orthogonal linear transformation that transforms the
data to a new coordinate system such that the greatest variance by any projection of
the data comes to lie on the first coordinate (called the first principal component),
the second greatest variance on the second coordinate, and so on. PCA is theo-
retically the optimum transform for given data in least square terms (Boersma and
Weenink 2001). For a data matrix, XT, with zero empirical mean (the empirical
mean of the distribution has been subtracted from the data set), where each row
represents a different repetition of the experiment, and each column gives the
results from a particular problem, the PCA transformation is given by:
Y T X T W VRT 3:6
where the matrix R is an m-by-n diagonal matrix with non-negative real numbers
on the diagonal and WRT V is the singular value decomposition (svd) of X. In our
analysis we tried to apply principal component analysis but we do not find good
results. Therefore we use maximum likelihood method with VARIMAX rotation
considering an extraction based on eigenvalue with eigenvalue greater than 0.5.
Indeed, in this kind of choice we find a good response of components. Before
trying this method, we analyze literacy rate also considering an extraction base on
eigenvalue greater than one but we discover a low result. In fact maximum like-
lihood estimation (MLE) is a popular statistical method used for fitting a statistical
model to data, and providing estimates for the models parameters (Besset 2001).
The method of maximum likelihood corresponds to many well-known estimation
methods in statistics. The sample mean is then the maximum likelihood estimator
of the population mean, and the sample variance is a close approximation to the
maximum likelihood estimator of the population variance. For a fixed set of data
and underlying probability model, the method of maximum likelihood selects

values of the model parameters which maximize the likelihood function. Maxi-
mum likelihood estimation gives a unified approach to estimation, which is well-
defined in the case of the normal distribution and many other problems.
In applying MLE we suppose that there is a sample x1, x2, , xn of n inde-
pendent and identically distributed (i.i.d.) observations, coming from an unknown
distribution f0(). It is however known that the function f0 belongs to a certain
family of distributions ff jh; h 2 Hg; called the parametric model, so that f0
f jh0 : The value h0 is unknown and is referred to as the true value of the
parameter. It is desirable to find some ^
h (the estimator) which would be as close to
the true value h0 as possible. Both the observed variables xi and the parameter h
can be vectors. The variables xi may be non-i.i.d., in which case the formula below
for joint density will not separate into individual terms; however the general
principles would still apply. To use the method of maximum likelihood, one first
specifies the joint density function for all observations. For i.i.d. sample this joint
density function will be
f x1 ; x2 ; . . .; xn jh f x1 jh f x2 jh f xn jh 3:7
We may extend the domain of the density function so that the density is also a
function of the parameter h. Then, for a given sample of data with observed values
x1, x2, xn, the extended density can be considered a function of the parameter h.
This extended density is the likelihood function of the parameter:
Y
n
Lhjx1 ; . . .; xn f x1 ; x2 ; . . .; xn jh f xi jh 3:8
i1
However, in general, the likelihood function is not a probability density. In fact,

it does not need to be an additive function, thus it is not a probability measure. In
practice it is often more convenient to work with the logarithm of the likelihood
function, ln L; called the log-likelihood, or its scaled version, called the average
log-likelihood:
Xn
1
ln Lhjx1 ; . . .; xn ln f xi jh ^ ln L 3:9
i1
n
Indeed, ^ estimates the expected log-likelihood of a single observation in the

model. The method of maximum likelihood estimates h0 by finding a value of h
^
that maximizes hjx: This method of estimation is a maximum likelihood esti-
mator (MLE) of h0:
^ ^ 1 ; . . .xn
hmle arg max hjx 3:10
h2H
By applying maximum likelihood estimation, we can identify a point estimate

referred to each country considered in the analysis. Furthermore, for the appli-
cation of maximum likelihood we use VARIMAX rotation (Jakeman 2008). We
know that in quantitative methods a VARIMAX rotation is a change of coordinates
used in principal component analysis and factor analysis which maximizes the sum
of the variances of the squared loadings. That is, it seeks a basis that most eco-
nomically represents each individual, so that each country can be well described
by a linear combination of only a few basic functions:
0 !2 1
X k X P
c X k Xp
RVARIMAX arg max@ 4
KRij KRij A
2
3:11
R
j1 i1
p j1 i1
where c 1 for VARIMAX. Variance maximizing rotation is often used in sur-

veys to see how groups of countries measure the same phenomenon. In our case
how the 30 countries considered distance themselves.
Ultimately, we use hierarchical clustering, which is a method of cluster analysis
which seeks to build a hierarchy of clusters. Strategies for hierarchical clustering
generally fall into two types, but we use only agglomerative method that is a
bottom up approach, in fact each observation starts in its own cluster, and pairs of
clusters are merged as one moves up the hierarchy. In general, the merges and
splits are determined in a greedy manner. The results of hierarchical clustering are
usually presented in a dendrogram. A measure of dissimilarity between sets of
observations is required in order to decide which clusters should be combined (for
agglomerative) (Szkely and Rizzo 2005). In most methods of hierarchical clus-
tering, this is achieved by use of an appropriate metric (a measure of distance
between pairs of observations), and a linkage criteria which specifies the dissim-
ilarity of sets as a function of the pairwise distances of observations in the sets. We
chose Euclidean distance:
r
X
ka bk2 ai bi 2 3:12
i
We use the linkage criteria maximum or complete linkage clustering that

determines the distance between sets of observations as a function of the pairwise
distances between observations:
max fda; b : a 2 A; b 2 Bg 3:13
We apply the method of hierarchical cluster in order to identify any consistency
between the countries considered, and part of the same cluster as well as the
heterogeneity between clusters.
3.3 Preliminary Empirical Investigation on the Literacy

Rate and Expectancy Variables
The following analysis refers to the 30 countries some of which are OECD
members, over the period of time 20072009. The countries we chose to include in
our cross-country analysis are those that represent manifold and diversified
Table 3.1 Average, lowest and highest values for dependent and expectancy variables
Indicator Mean Lowest Highest
Literacy rate 98.25 88.70 (TR) 99.80 (LV)
GDP r.g.r. 0.003 -0.041 (LV) 0.045 (PL)
Pre-primary 0.891 0.159 (TR) 1.233 (ES)
Primary 1.058 0.549 (PT) 1.577 (ES)
Secondary 0.785 0.183 (NL) 1.479 (AU)
Tertiary 0.413 0.01 (NL) 0,968 (U.S.)
l.t.u.r. 1.760 0.045 (MX) 5.988 (SK)
p.e.p. GDP 0.051 0.019 (AT) 0.086 (JP)
p.e.p. GOV 0.132 0.034 (PT) 0.281 (LV)
c.o.s. 0.030 0.0002 (LU) 0.416 (U.S.)
Mean, maximum and minimum values for dependent and expectancy variables. In parentheses is
given the international symbol of the various countries
socio-economic contexts. However, the principal context studied is that of the

most advanced economies, which can be divided in numerous subgroups of
remarkable interest. Indeed, in the European case it is possible to observe the
substantial differences between the Scandinavian and the Mediterranean countries,
or the dissimilarities in the American States with Anglo-Saxon or Latin origin, in
theme of economy, education and society. In further investigations, we do not
exclude to consider and compare socio-economic contexts characterized by a very
dissimilar situation, like rich and poor countries, but it is essential to consider the
relevant problem of quality and updating of so numerous variables, as those used
in this paper. The first analysis performs the calculation of average values, lowest
and highest values reported for each considered variable. Variables that we ana-
lyze are: literacy rate,3 GDP real growth rate (GDP r.g.r.), pre-primary, primary,
secondary and tertiary schooling as percentage of total population by every
education levels of every country,4 long term unemployment rate (l.t.u.r.), public
spending as percentage of GDP (p.e.p. GDP) and public spending as percentage of
government expenditure (p.e.p. GOV), children out of school (c.o.s.) as percentage
of total children out of school. In the next page we show the results obtained from
descriptive analysis (Table 3.1).
Literacy rate has a mean value equal to 98.254%, this average is referred to the
30 countries considered in our analysis. The minimum is 88.7% and is represented
by Turkey (TR), the maximum value is represented by Latvia (LV) and is almost
100%. GDP real growth rate has its minimum value in Latvia, the same regards
other countries such as Hungary (HU) and Ireland (IE) which show negative
3
The literacy rate represents the percentage of people with the ability to read and write, and in
this case without specifying the level of education.
4
In our analysis we take into consideration different levels of education, such as pre-primary,
primary, secondary and tertiary education: these levels are generally found in similar ways in
many countries, and represent usually from pre-primary until the so called higher education
(tertiary).
values, while the highest refers to Poland (PL). The average GDP real growth rate
is near to zero, because in recent years there has been stagnation and recession of
the economy. The several education levels get the minimum value by Turkey,
Portugal (PT) and Netherlands (NL) in the four categories considered. Slovakia
(SK) shows the highest value reported for long unemployment rate. Furthermore,
Austria (AT) and Portugal are the countries with lowest usage rates of public
spending in education. Finally the variable related to children out of school shows
its maximum value in United States (US) and this percentage is always condi-
tioned by a high population abundance, and we note simultaneously a high degree
of school drop-outs. In the next section we begin the analysis of the literacy rate
applying multivariate regression method, maximum likelihood method and clus-
tering analysis to highlight the differences in education levels in recent years in the
countries considered.
3.4 Results from the Multivariate Regression Model
A first analysis of the variables related to education highlights a high variability in

the data due to different density of population of countries considered. For example,
United States has only in the pre-primary school more than 12 million of people,
whereas the U.S. population exceeds 300 million inhabitants in 2009 (U.S. Census
Bureau data). This indicates that data are based also on countrys population size.
The same is true for Mexico, Japan and France, which show a high abundance of
individuals enrolled in different educational levels. In contrast, countries which
show less members are represented by Iceland, Ireland, Latvia, Luxembourg and
Slovenia. This variability is also confirmed by symmetry and kurtosis indices
showing high amounts; this indicates an asymmetry of these distributions compared
to the normal one. We consider the parameter as normal when most of data are
distributed in average value, however in our results values are distributed more
even in areas of tail. Hence, literacy rate and children out of school reveal an
asymmetry of information, but it is only apparent as there are slight differences in
their percentages. Some countries have 99% of schooling, but many others are
below 90%. While long term unemployment rate, public spending and GDP real
growth rate report almost a normal distribution with low variability. By examining
interdependence between variables we notice strong groupings of countries in
relation to education levels and economic growth. Considering different levels of
education relate to each other, we note high direct relations. However, we note that
in most countries the literacy rate has percentages between 90 and 99% while the
long run unemployment rate shows a clear grouping of countries in the lowest
percentages of literacy rate. This indicates that with an increasing of the degree of
education levels the long-run unemployment rate decreases, as with a higher level
of education attained, individuals find more job opportunities.
Our aim here is to analyze literacy rate relationship with other variables pre-
sented in Sect. 3.3. For this purpose we use a multivariate regression analysis and the
mathematical model we propose is the (3.4), and in terms of variables is the (3.5).
Before applying the (3.5) we perform other types of investigations in order to
understand the influence of literacy rate compared to single macro groups5 under
analysis: we have first investigated literacy rate as a function only of education
variables, then literacy rate only with variables related to social exclusion.
Observing our results of statistical tests concerning the significance of the param-
eters and considering the best fitting to data, we choose to consider model as a
whole. Indeed, we use also education variables as percentages of total education in
different countries in order to identify the optimal model applied; resulting data are
not satisfactory for the significance of the parameters in any kind of linear combi-
nation we calculate. Thus, the combination of economic growth, education levels
and social exclusion variables leads to a better result than analytical models con-
sidered singly.
We analyze the presence of multicollinearity using the tolerance analysis that is
equal to:tol k R2 . When the values of this analysis are close to zero there is a
combination of other endogenous variables covered by our study. Moreover, we
use the variance inflation factor (VIF) that is the reciprocal of tolerance, in order
to achieve further confirmation of the result found with the tolerance analysis. The
VIF has a threshold value equal to 0.10 and we can check if there is collinearity
among independent variables considered. Tests for investigating endogeneity have
not been reported in this study since they have provided results in the standard and
not particularly interesting for our goals.
In addition, an important test we apply to detect the presence of autocorrelation
between independent variables is the Durbin-Watson statistics which is used in the
erratic component (et) of a regression analysis. The related formula is equal to:
PT 2
t2 et et1
d PT 2 3:14
t1 et
We are also interested in applying the heterogeneity test which concerns the
system of variables we have selected in order to describe the phenomenon of social
exclusion in terms of illiteracy as our data set is quite multifaceted (many vari-
ables, different from each other and variability between countries). For identifying
any heterogeneity we use the ANOVA test (analysis of variance), which locates
the dispersion of observed values compared with an expected value, so that we can
detect the possible homogeneity between the variables; the resulting variances
should be quite similar to each other otherwise it will confirm the hypothesis of
heterogeneity of the system. Finally, all the explanatory variables should be
uncorrelated with the residual terms (uncorrelated residual values), that is the
residue of an observation must be independent from the residual value resulted
from another observation. The results are shown in Tables 3.2 and 3.3.
5
For macro groups we mean groupings of the endogenous variables based on their similarities.

Social Exclusion - Short and Long Term Causes and Consequences-Pages-1-52

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Social Exclusion - Short and Long Term Causes and Consequences-Pages-1-52

Transféré par

Droits d'auteur :

Formats disponibles

Social exclusion: short and long term causes and consequences

Author(s) Parodi, Giuliana ; Sciulli, Dario

For further volumes:

Library of Congress Control Number: 2011936144

Springer-Verlag Berlin Heidelberg 2012

Cover design: eStudio Calamar, Berlin/Figueres

Printed on acid-free paper

Physica-Verlag is a brand of Springer

Antonello Scorcu, Universit di Bologna, Italy

1 Introduction and Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Part I The Structural and Long Term Causes of Social Exclusion

2 Health, Lifestyle and Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3 A Comparative Analysis of Literacy Rate in Contributing

4 Education and Socioeconomic Mobility in

5 Analysing Industrial Accidents in European Countries

Part II Groups Structurally Vulnerable to Social Exclusion

6 Social Exclusion and Offending . . . . . . . . . . . . . . . . . . . . . . . . . . 105

7 Disability and Social Exclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 127

8 Civilian Disability Pensions as an Antipoverty Policy Instrument?

Part III Economic Crisis, Labour Markets and Social Exclusion

9 Measuring the Long Wave. Unemployment, Discouragement

10 Poverty and Unemployment: The Cases of Italy and Spain . . . . . 199

11 Labour Market Transitions During the Financial

12 Labour-Market Dynamics and Unemployment: New Scenarios

Tindara Addabbo Dipartimento di Economia, Universit di Modena e Reggio-

Fabrizio Muratore Scuola Superiore, Universit G. DAnnunzio di Chieti-

Giuliana Parodi and Dario Sciulli

1.1 The XXV AIEL Conference

G. Parodi (&) D. Sciulli

1.2 Main Approaches to the Discussion of Social Exclusion

exclusion or deprivation score of a person depend on the number of characteristics

interventions and the financial intervention of individual member states. The

1.3 Economic Crisis, Labour Markets and Social Exclusion

Economic conditions are an important factor in determining poverty and social

unemployment dynamics have affected various countries and various worker

1.4 Brief Summary of the Papers in the Book

1.4.1 The Structural and Long-Term Causes

1.4.2 Groups Structurally Vulnerable to Social Exclusion

1.4.3 Effects of the Recent Crisis on the Labour Market

some differences emerge, probably because of some differences in model speci-

and formerly convicted people, are particularly vulnerable to social exclusion.

Keywords Health Lifestyles Growth

2.2 Some Empirical Evidence

These disparities may be explained by differences in Public Health, Medical

First, let us suppose that an economy produces 3 goods: 2 commodities for

2.3.1 The Consumer Utility Function

The consumer utility function is assumed to be a Cobb Douglas function in which

Where a; b and dare the elasticities of h, x and z, respectively.

This is a static model. There is no rational addiction, but positive elasticity

2.3.2 The Health Production Function

hx; z; h0 ; w; t; e xq zc h0 we/t ee 2:5

assumed to exhibit a decreasing return; thus, 0\q\1; 0\c\1; and, therefore,

U h; x; z Xxaq zac xb zd 2:7

U h; x; z Xxaqb zdac 2:8

2.3.3 The Utility Maximization Problem: The Optimal

Let X 1: The consumers budget constraint is px x pz z cy; where px ; pz are

2.4 Comparative Static: The Health Multiplier

d log Y d log Y d log t 1 a

2.5 A Growth Model with Health

In the comparative static framework, the level of population is constant. In this

In the Neoclassic Economic Growth literature, many models consider health as

Equation 2.23 can be rewritten as

This is a Solows Model, and the new technological rate is