Measuring Freedom On the Operationalisation of the Capability

Measuring Freedom On the Operationalisation of the Capability Approach Sebastian Silva-Leander QEH, Oxford 28 May 2008

Introduction • Is it possible to move beyond functionings in the measurement of capabilities? – Can it operate the transition from axiomatic discussion to operational framework? – Deconstruct the evolving concept of “capability” into its different components. – Identify conceptual hurdles that have prevented the operationalisation of the capability approach. – Put it back together in a way that avoids some of the pitfalls.

Caveats • Synthesis and overview rather than a new theory • Stretches across several disciplines • Comprehensive and didactic rather than technical

Structure • Part I: What are we trying to Measure? • Mapping • Deconstruction • Reconstruction • Part II: How are we going to Measure it? • Autonomy • Agency • Techniques • Part II: Measuring it! • Poverty • Development

1. 1. Mapping: A. K. Sen • Origins of neoclassical economics virtually indistinguishable from moral philosophy (utility, maximisation, pareto-optimality) • Utilitarianism (and economics) as the culmination of the “naturalist” project in moral philosophy [1] • Predictability: is the condition for applying scientific methodology 2] • But: The elimination of the Free will means the elimination of morality : Hobbesian man, maximisation [3]

1. 1. Mapping: Choice Inputs and outputs are no longer equivalent

1. 1. Mapping: Literature

1. 2. Deconstruction: Justice vs. Freedom • Qualified priority of liberty avoids the extremes of moral intrusion and moral skepticism… • But leaves open the question of trade-offs between justice and freedom: • List: Nussbaum vs. Sen • Context-dependence: role of public reason and discussion. • Mismatch between the positive language of capability (i. e. choice) and the normative language of freedom • Conceptual overlap: Concepts of functionings and capabilities cut across agency freedom, well-being freedom, etc.

1. 2. Deconstruction: Reason vs. Preference • Things that people value, and have reason to value… • Adaptive preferences, misinformation, perversion… • Functionings vs. Capabilities: some functionings, e. g. health involve substantive freedoms. • Refined functionings • Bad health: NHS vs. drug addiction • Functional interdependence: some capabilities (e. g. rational choice) may be dependent on the prior acheivement of certain functionings (e. g. literacy): • Content Independence of choice // Paternalism • Dignity and the distinctness of Reason

1. 2. Deconstruction: Wellbeing vs. Agency • Role of public action? • Discussion on China vs. India (largely instrumental role of democracy) • Effective vs. Control Freedom • Malaria free environment: confuses power and freedom (Cohen). • The central issue for freedom is not coincidence with my preferences, but control, i. e. democracy (robust).

1. 3. Reconstruction: Re-mapping

1. 3. Reconstruction: Formalising • Opportunity – Against indirect utility: – Cardinal: • Preference – Relevance of Preferences: – Reverse:

1. 3. Reconstruction: Formalising • Right – Unreasonable Pref. : – Reasonable Pref. : • Agency – Autonomy as IC: – Process:

Structure • Part I: What are we trying to Measure? • Mapping • Deconstruction • Reconstruction • Part II: How are we going to Measure it? • Autonomy • Agency • Techniques • Part II: Measuring it! • Poverty • Development

2. 1. Autonomy: Impossibility • Potentials: capability “production function” is informationally demanding (environmental, social, personal, mental, etc. ) • Haverman and Bershadker’s (2001) “self-reliant poverty”, Burchardt (2006) time and income poverty • Dynamics, first-second choices (e. g. going to college) • Counterfactuals: “hypothetical situations which never occurred and might never occur” (Brandolini and D’Alessio 1998, p. 12). • Ignores the information contained in choice • Latent functionings (structural equation modelling) • HDI: Netherland vs. US (Dowrick, Dunlop and Quiggin 2003 cxix) • Spontaneity: The power to create novel options ex-nihilo • ‘Fasting’ is qualitatively different from ‘dieting’: it is a novel option that is not adequately captured by the “maximum” attainable option of eating. • The range of potential options is not finite (originality)

2. 1. Autonomy: Reason, Moral Worth • Autonomy as purposeful self-restriction “You must bind me very tight, standing me up against the step of the mast and lashed to the mast itself so that I cannot stir from the spot. And if I beg and command you to release me, you must tighten and add to my bonds. ” (Homer, The Odyssey, p. 161) • Opportunity is important because there is an agent who utilises it to pursue valuable objectives • Value is now an integral part of the assessment (this is Sen’s revolution) • Move away from the expansionist approach • Missionaries of Charity’s vow of poverty • Sustainable development or Swedish welfare system

2. 2. Agency: Assessing Autonomy • Value: Contextual dependence • • (1) the expertise-based method, (2) the empirical or statistical method (3) the participatory method and (4) the rights-based method, • Subjectivity: Inter-individual comparability • Adaptation: Idiosyncratic, time, social, value (survey techniques, SWB literature) • Intentionality: is not externally observable

2. 2. Agency: Public Action • Institutionalisation: laws, institutions render ephemeral acts permanent • Political Freedom: translates individual will into collective will • Juridical law can be considered as self-imposed (ie. Autonomous) • Process: intentionality not necessary • Consent • Process of inter-rational validation • procedural superset.

Structure • Part I: What are we trying to Measure? • Mapping • Deconstruction • Reconstruction • Part II: How are we going to Measure it? • Autonomy • Agency • Techniques • Part II: Measuring it! • Poverty • Development

3. 1. Poverty: multi-spatial • 6 dimensions, 2 evaluative spaces: • dimensions flexible (survey specific, info. value), evaluative spaces fixed (functionings, agency) • Intra-dimension corr. > inter-dimension • Functionings coor. with income; agency corr with “power to change” • Formal test: compare agency/functioning poverty profiles (ordered probit/logit) • • • Generalised Hausman: Ho “no difference in coefficients” Across spaces (income/change): rejected 1% Within spaces (multidim): rejected, but often of the same sign Within dim. (agency/funct. ): except mobility and employment Across dim. (robustness): accept except education/mobility

3. 1. Poverty: Adaptive Preferences • FGT more sensitive to adaptive pref.

3. 1. Poverty: Stochastic Dominance

3. 2. Development: as Freedom? • Opportunity: GDP • Aggregate quantity of opportunity (may need to be adjusted, e. g. Israel, Maldives) • GDP may be endogenous to collective decision (e. g. Scandinavian countries). • Process: Political Freedom • Not a dimension • GDP may be endogenous to collective decision (e. g. Scandinavian countries). • Objective: MDGs • Pros: rights based, weighed, • Cons: computationally heavy • Measuring effort: social spending, relative achievement.

3. 2. Development: Single Indicator • Comparisons of income distributions • Intuitive interpretation: probability that someones income in A > in B • Pros: valid across time/countries, conceptual counterpart of income, but not correlated with income, • Cons: not universally accepted, collective decision to sacrifice welfare of some memebers of the community (US social consensus) • But: empirically capture information contained in HDI/MDGs. • Truncation by political freedom • If fully democratic, the full distribution counts: even the income of Bill Gates is theoretically at the disposal of the community, if they decided to redistribute it for other purposes. • If undemocratic, the incomes of the rulers are not accessible for the masses.

3. 2. Development: Ranking gdpcens concordance 69% 68% size of correction (small) (-0. 3) (-6. 4) concordance (large) 73% 60% size of correction (large (-21. 2) Correlations gdpprob (-29. 8) Truncated Income Dist. GDP 0. 9363* gini -0. 5189* polfree -0. 5683* hdi -0. 8809* Top 10 Luxemburg Bottom 10 Losers (comp/GDP) Winners (comp/HDI) Losers (comp/HDI) Solomon Islands Colombia Botswana Kazakhstan China South Africa China Chad Benin Sao Tome and Principe Guinea Namibia Lebanon Denmark Mozambique Guyana Turkmenistan Djibouti Tajikistan Japan Guinea-Bissau Cape Verde Lebanon Solomon Islands Vietnam Norway Burundi Samoa Angola Gabon Turkmenistan Finland Ethiopia Bangladesh Equatorial Guinea Benin Libya Germany Burkina Faso Latvia Gabon Mauritius Cuba Canada Niger Jamaica Brazil St. Lucia Qatar Iceland Sierra Leone Saudi Arabia Oman Barbados Austria Belgium Mali Central African Republic Winners (comp/ GDP)

“But how do those, Socrates, who are trained to the art of ruling which you seem to me to consider as happiness, differ from those who undergo hardships from necessity, since they will have (though it be with their own consent) to endure hunger, and thirst, and cold, and want of sleep, and suffer all other inconveniences of the same kind? For I, for my own part, do not know what difference it makes to a man who is scourged on the same skin, whether it be voluntary or involuntary, or, in short, to one who suffers with the same body in all such points, whether he suffer with his consent or against it, except that folly is to be attributed to him who endures troubles voluntarily. ” (Xenophon Memorabilia, 17 -18, p. 390)