«CICLO XXVII Anno 2015 The Constitutional Approach to poverty measurement. A multiple-deprivations framework for high-income countries. Dottorando: ...»
DOTTORATO DI RICERCA IN
ECONOMIA DELL’AMBIENTE E DELLO SVILUPPO
Scuola Dottorale in Economia e Metodi Quantitativi
The Constitutional Approach to poverty measurement. A multiple-deprivations framework for
Dottorando: Francesca Tosi Relatore: Prof. Paolo Liberati Correlatore: Dr. Francesco Burchi Coordinatore: Prof. Luca Salvatici Ai miei genitori, a mia sorella Valeria.
Abstract Today there is widespread agreement on the urge of adopting conceptual frameworks and evaluation tools that allow to include a variety of aspects of human life in evaluating people’s living standards, even at the highest levels of policy making. While there is no consensus on which specific life domains should be taken into account, the arguments in favour of a broadening of the informational basis for poverty analysis are cogent: low consumption surely is at the heart of the concept of ‘poverty’ but a number of other domains – like poor human health, limited access to education and powerlessness – are systematically concerned by inadequate living standards. This study hypothesises that the application of a theoretical framework for the reconceptualization of multidimensional poverty as a distributive justice question can be used to (i) address the need for multidimensionality in poverty assessment while minimizing the degree of arbitrariness with which normative choices are often made; and (ii) explain changes in living standards in high-income countries and inform policy makers in a more effective way compared to a unidimensional poverty framework. In Chapter I, we thoroughly review theories and metrics in support of multidimensional poverty analysis, with a particular focus on counting methods and axiomatically derived poverty indices. We highlight the advantages of applying a methodology that enables us to study the distribution of multiple deprivations that simultaneously affect the individuals and identify a new family of poverty measures that can serve our purpose. In Chapter II, we study the possibility of using conceptual instruments offered by contemporary theories of social justice to address a multidimensional poverty question for high-income countries. We then develop a theoretical framework inspired by John Rawls’ theory of justice to reconceptualise multidimensional poverty analysis, which enables us to frame an ethically sound and publicly justified empirical assessments of people’s living conditions in constitutional democracies through the application of a Constitutional Approach to dimensions selection. In Chapter III, we contextualize the research in the European framework, reviewing the history and the state of the art of multidimensional poverty analysis for Europe and Italy and identifying weaknesses and possible room for improvement in poverty measurement approaches currently in use. Chapter IV concludes with an investigation on the joint distribution of multiple deprivations in Italy based on EU-SILC cross-sectional data from 2004 to 2013. Two main factors tell the story of deprivation in Italy: age and geography. Deprivations follow a clear pattern through the different stages of life: the youth are threatened by unemployment and economic insecurity, while the elderly report more deprivations in health conditions and educational attainment at once. Geographically, multidimensional poverty estimates confirm the existence territorial disparities already accounted for by official monetary poverty measures; nonetheless, they also highlight remarkable exceptions, confirming that well-informed multidimensional estimates are able to generate non-trivial results. The study of the phenomenology of multiple deprivations in Italy shows that poverty breadth is higher in the South, as well as poverty i intensity; conversely, inequality among the poor is widespread in the North, especially in north-western regions. During the past ten years, multidimensional poverty has steadily decreased until 2010 and it started increasing without a clear pattern afterwards. Deterioration in both decent work and health conditions has mainly contributed to this change of pattern, while education and living environment conditions have improved across time. Comparing multidimensional figures to official statistics on poverty allows us formulating possible explanations about what are the factors to which different indicators are more likely to be sensitive. We conclude suggesting that a generalization of this study to assess multidimensional poverty at European level can make an interesting area for further research.
I would like to thank first and foremost my main supervisor, Professor Paolo Liberati, for his time, patience and availability to guide me through the completion of this dissertation. A special thanks goes also to my supervisor Dr. Francesco Burchi, for dedicating me fruitful discussion, emails and skype calls whenever I needed to dissipate my doubts, and to Professor Pasquale De Muro, for initiating me to topics of crucial importance to a young economist, such as social justice theories.
I would also like to acknowledge the support this thesis has received by the Department of Sustainable Economic and Social Development of the German Development Institute in Bonn, the Bremen International Graduate School of Social Sciences of the University of Bremen, the European Commission’s 7th Framework Programme InGRID – Inclusive Growth Research Infrastructure Diffusion (FP7/2013-2017 – grant agreement no. 312691), as well as from the coordinators and participants of PRIN 2009 “Measuring human development and capabilities in Italy: methodological and empirical issues” research project.
Finally, to the friends and colleagues that have accompanied me through these years, especially Ylenia, Silvia and Matteo, for supporting me as times were getting harder – and for staying friends nonetheless.
To my parents, Lucia and Roberto, and to my family. For trusting my choices and donating me their happiness, love and stability. I know my thirst for travel and discovery has sometimes challenged them and I couldn’t be more grateful to them for being as strong and supportive as they are. Thank you all.
AF Alkire-Foster AN Anonimity AROPE At-risk-of-poverty-and-social-exclusion rate BES Benessere Equo e Solidale (Equitable and Sustainable Well-being) BNA Basic Needs Approach CA Capability Approach CN Continuity CNEL Consiglio Nazionale dell’Economia e del Lavoro (National Council of Economy and Labour) CSPI Correlation-Sensitive Poverty Index CVI Conjoint Vulnerability Index ECHP European Community Household Panel Survey EQLS European Quality of Life Survey ESRI Economic and Social Research Institute ESS European Social Survey EU European Union EU-SILC European Union Statistics on Income and Living Conditions FC Focus FD Factor Decomposability FGT Foster-Greer-Thorbecke GE Generalized Entropy G-SOEP German Socio-Economic Panel HBS Household Budget Survey HDR Human Development Report HR Headcount Ratio HRP Household Reference Person ISPL International Standard Poverty Line ISTAT Istituto Nazionale di Statistica (National Statistics Institute) IT Information Theory IS Inequality Sensitivity ISPI Inequality-Sensitive Poverty Index MN Monotonicity MPI Multidimensional Poverty Index MRS Marginal Rate of Substitution MW Monotonic Welfare NDCI Non-decreasingness under Correlation Increasing Arrangement NDS Non-decreasingness in Subsistence Levels of Attributes NICI Non-increasingness under Correlation Increasing Arrangement NIPA Non-increasingness under Pareto-efficiency Association Increasing Switch NLPCA Non-Linear Principal Component Analysis vi NPG Non-Poverty Growth NPV Normalized Poverty Value NR Normalization OECD Organisation for Economic Co-operation and Development ORW Ordinal Rank Weights OMC Open Method for Coordination OPCS Office of Population, Censuses and Surveys PCA Principal Component Analysis PD Pigou-Dalton Transfer Principle PG Poverty Gap PP Principle of Population PPF Production Possibility Frontier RE Relative Equity SD Subgroup Decomposability SHARE Survey of Health, Ageing and Retirement in Europe SI Scale Invariance SLL Swedish Level of Living Survey SPG Squared Poverty Gap SWF Social Welfare Function TFA Totally Fuzzy Approach TFR Totally Fuzzy and Relative Approach TR Transfer UDB User Database UM Uniform Majorization UNCTAD United Nations Conference on Trade and Development UNDP United Nations Development Programme UNEP United Nations Environment Programme WAI Weak Association Increasing Switch
For a long time, in the economic literature poverty has been generally associated with lack of goods and services, often overlooking the fact that it is also a synthesis of political, economic factors and social rights and, above all, a human phenomenon. The income-centred approach has shown notwithstanding severe limitations, mainly due to the lack of evidence of perfect correlation between income poverty and unsatisfied basic needs, especially at household level (Lipton and Ravallion, 1995; Fukuda-Parr, 1999;
Thorbecke, 2008), and to the recognition of the presence of imperfect or non-existing markets (Tsui, 2002; Bourguignon and Chakravarty, 2003) such as those of public goods. Low consumption surely is at the heart of the concept of ‘poverty’ but a number of other domains – like poor human health, limited access to education and powerlessness – are systematically concerned by inadequate living standards.
During the last decade, the acknowledgment that poverty involves much more than just low income has inspired a great scientific literature production on methodologies to measure or analyse deprivation in multiple dimensions: since the pioneering works of Tsui (2002) and Chakravarty and Bourguignon (2003), a number of approaches were developed to contribute to the debate, e.g., the axiomatic approach (Chakravarty et al., 1998; Alkire and Foster, 2011a), the information theory approach (Deutsch and Silber, 2005; Maasoumi and Lugo, 2008) and the fuzzy set approach (Cheli and Lemmi, 1995; ChiapperoMartinetti, 1994; 2000). At policy level, the opportunity to use multidimensional poverty evaluation systems has been highly debated as well. Many countries have already officially adopted multidimensional measures for the evaluation of poverty and social exclusion: among others, Colombia, Mexico, Bhutan, El Salvador, Malaysia, Peru and Philippines. Not surprisingly, developing countries have been the driving force of this change of perspective. Following the prominent examples of the World Bank and the United Nations Development Programme, that respectively publish the Millennium Development Goals and the Multidimensional Poverty Index on a yearly basis, the way of conceptualizing deprivation in poor countries has radically changed and is now more and more focused on the monitoring of achievements in multiple aspects of life (Alkire and Santos, 2010; 2014). At OECD and European Union level, some attempts of reconceptualising poverty and social exclusion through the use of composite indices have recently been made as well – see for instance the At-risk-of-poverty-and-social-exclusion rate released yearly since 2009 by Eurostat – and there is a general growing interest expressed by Governments and the civil society in understanding poverty as a multifaceted phenomenon (Stiglitz et al., 2009). Nonetheless, still a few studies contribute to the production of sub-national deprivation indicators or country-specific multidimensional poverty measures for wealthy countries, where multidimensionality in living standards – especially in the Anglo-Saxon tradition – is still mainly associated with material deprivation only (Whelan et al., 2001; Layte et al., 2001; Whelan et al., 2002; Guio and Maquet, 2006; Coromaldi and Zoli, 2012).
While the advantages of applying a multidimensional approach to poverty measurement are widely 1 acknowledged, aggregate poverty measures are far from being universally welcomed (Fleurbaey, 2009;
Ravallion, 2011; Ferreira and Lugo, 2013). With this regard, the main disagreement in the literature lies in the way normative choices in support of the construction of poverty indices are made. The choice of dimensions, indicators, weights and cut-offs to be used to identify the poor requires in fact an evaluative process that we can expect to be diverse for each case, depending on the context, the scope of the analysis and the contingencies. Without an explicit reference to a sound theoretical ground, the risk the researcher encounters is to counterbalance the intrinsic strengths of a multidimensional framework – i.e., its inclusiveness and richness of information – with an excessive arbitrariness and a lack of theoretical justification.
Our research joins the scientific debate at this point and seeks to provide a clear answer to the following research questions: (i) how can we address the need for multidimensionality in poverty measurement while minimizing arbitrariness in normative choices and unambiguously informing empirical assessments? and (ii) can a multiple deprivation framework be used to explain changes in living standards in high-income countries and to inform policy makers more effectively than a unidimensional one?
This study is organized as follows. In Chapter I, we thoroughly review theories and metrics in support of multidimensional poverty analysis, with a particular focus on counting methods and axiomatically derived poverty indices. In Chapter II, we study the possibility of using conceptual instruments offered by contemporary theories of social justice to address a multidimensional poverty question for high-income countries. We then develop a theoretical framework, i.e., the Constitutional Approach to poverty measurement, to inform ethically sound and publicly justified empirical assessments in constitutional democracies. In Chapter III, we contextualize the research in the European framework, reviewing the history and the state of the art of multidimensional poverty analysis for Europe and Italy and identifying weaknesses and possible room for improvement in poverty measurement approaches currently in use.
Chapter IV concludes with an investigation on the joint distribution of multiple deprivations in Italy based on EU-SILC cross-sectional data from 2004 to 2013.