The hidden dangers of targeting

Transcription

1 The hidden dangers of targeting Elena Granaglia (University of Roma Tre) Michele Raitano (Sapienza University of Rome) «Poverty in Europe and how to fight it» Sapienza University, Faculty of Economics May 26th, 2017

2 Different forms of targeting Three main forms of targeting (directing resources to those deemed most in need of them) - tagging (Akerlof) = resources are targeted on the basis of individual characteristics positively correlated to need (and negatively to earning ability) - categorical transfers: ie disability - means testing = resources are targeted to the poor - selective transfers, where selectivity could both aim at including (the poor/the disadvantaged) or excluding (the rich/the more advantaged) - targeting within universal transfers: additional resources are targeted to the disadvantaged within universal transfers - progressive/proportional universalism 2

3 Different forms of targeting Furthermore, an indirect form of targeting (to deal with the shortcomings informational failures - in the main forms of targeting) - self-targeting (self-revelation) - ie restrictions the choice of consumption or even ordeal mechanisms imposing mere deadweight cost such as waiting, complicated selection procedures ( Nichols and Zeckhauser,1982, Besley and Coate, 1998) so to deter the not needy from entering into the program 3

4 Focus of the contribution Targeting as means testing (selectivity) - pro-poor targeting (inclusive) - even though broader implications One of the several dangers of means testing - the unavoidable arbitrariness in the definition of the beneficiaries/ non beneficiaries - more specifically: the arbitrariness of - the poverty line - the means utilized in means testing Empirical estimates of this danger and some policy implications 4

5 Before that. A (very) brief survey of some other dangers of means testing, certainly recognized in the academic literature, but quite underrated in the actual design of means tested programs More precisely, at least four other main dangers of means testing 5

6 The many risks of means testing (1) The creation of the two nations ( us vs. them ) (well known topic, yet more attention to plurality of the implications) - the undervaluation of the role of the market and of policy failures in creating poverty: on the contrary, poverty as their responsibility - with respect to the role of the market, - the trends in labour demand (lack and low wage) - the increase in market inequalities and the correlation between high inequality and high poverty 6

7 The many risks of means testing (2) - with respect to policy failures, failures in ensuring - a fairer market distribution (in the fight against the many economic rents produced in today s markets) - equality of opportunity against the persistent/even increasing influence of the social lottery - increasing in so far as high income/wealth inequality could make it harder to ensure equality of opportunity - Tocqueville on the role of some equality of conditions - topic which needs more work, but some new interesting data - ie on increasing economic inequality, increasing social and territorial segmentation and lower equality of opportunity (Chetty et al. 2016, on the importance of neighborhood) 7

8 The many risks of means testing (3) Notice: if so, utilizing Scanlon s/roemer s terms, not only absence of accountability for the poor responsibility in a strong sense but even absence of attributability) On the contrary, in so far as policy failures, it is our responsibility 8

9 The many risks of means testing (4) - the risk of giving insufficiently to them, the poor - the Korpi- Palme Paradox - some criticism (Kenworthy, 2011, Madama, Natili, 2016, Marx et al. 2013) - yet, the role played by Denmark and Italy and the differences/discrimination among the poor 9

10 The many risks of means testing (5) - the overall denial of common citizenship as signalled - by the introduction in several minimum income schemes in the EU of self selection procedures treating differently (harshly) the poor (J. Daniel Blake) - even leaving aside self selection objectives, by the imposition of one-sided obligations and paternalistic requirements - ie the obsession with work conditionality when work becomes lacking - by the diffusion of a double morality : the cheater undeserving poor vs. the excess burden of labor taxation - in both cases, same reaction to incentives - the word itself of targeting: «the analogy of a target does not at all suggest that the recipient is an active person, functioning on her own, acting and doing things. The image is one of a passive receiver rather than of an active agent» (Sen, 1992) 10

11 The many risks of means testing (6) In brief, the risk that welfare policy has to do with «what should be done about the poor and the maladjusted» rather than «what constitutes general fairness in respect to the relation between citizens and the state» (Rothstein, 1998) 11

12 The many risks of means testing (7) The additional risks of - false negatives - up to 50% of the eligible population (Bargain et al. 2012)? - lags in the coverage due to the selection process - the further problems when work increasingly becomes a portfolio of activities (Atkinson) - discriminations against the second wage earners and of overall inequities within the family - the family as the reference unit (the role of the effective marginal rates and of who receive the transfers) - the costs for women and young people - fewer problems for the young with the household as the unit 12

13 The unavoidable arbitrariness in individual rankings: an empirical analysis An undervalued risk: the arbitrariness in the definition itself of the beneficiaries/non beneficiaries according to how we compute income or economic wellbeing Additional issues even if we were able exactly to measure full income, but: A. Which reference unit? And, in case, which equivalence scale? B. Which means to consider? Income, wealth? The case of the Italian ISEE that jointly considers income and wealth. C. How to include non monetary items that affect wellbeing? The case of the house of residence. Do aggregate indexes change? Do relative positions change? 13

14 A. What changes when using different equivalence scales? Coefficients to homogenize household differences. When the number of components increases family income should increase by a lower proportion for keeping constant living standards => there are households returns to scale. Several scales and methodologies, but no one seems preferable. Econometric scales e.g. estimated on family consumption following some theoretical models or pragmatic scale. Comparisons on EU-SILC 2014 of 6 EU countries (DE, ES, FR, IT, SE, UK) using 3 scales: modified OECD, square root and Carbonaro (estimated for Italy assuming that 2 households of different size have the same living standard when they spend the same share for foods). The higher the scale, the lower returns of scale and the poorer the larger households (who then are more likely to receive transfers). 14

15 A. Equivalent components by household size Oecd modified (median values) Square root Carbonaro 15

16 A. Changes in Gini of disposable income DE ES FR IT SE UK Oecd modified Square root Carbonaro 16

17 A. Changes in poverty headcount 24.0% 22.0% 20.0% 18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% DE ES FR IT SE UK Oecd modified Square root Carbonaro 17

18 A. Exit out of poverty 26.0% 24.0% Share of poor according to the OECD scales escaping from poverty according to other scales 23.8% 22.0% 20.0% 18.0% 16.0% 14.0% 12.0% 11.3% 12.1% 11.4% 10.0% 8.0% 6.0% 6.1% 7.1% 5.9% 6.4% 5.9% 4.0% 2.0% 2.9% 2.9% 2.1% 0.0% DE ES FR IT SE UK Square root Carbonaro 18

19 A. Entry into poverty 4.0% Share of non-poor according to the OECD scales dropping into poverty according to other scales 3.0% 2.6% 2.4% 2.6% 2.1% 2.0% 1.7% 1.8% 1.8% 1.7% 1.3% 1.4% 1.4% 1.4% 1.0% 0.0% DE ES FR IT SE UK Square root Carbonaro 19

20 A. Reranking of individuals from OEDC to square root scale (Italy) 0 Square root OECD scale Oecd Square root Total Total

21 A. Reranking of individuals from OEDC to Carbonaro scale (Italy) 0 Carbonaro OECD scale Oecd Carbonaro Total Total

22 B. The impact on reranking of the ISEE reform in Italy 22

23 B. The ISEE reranking for those who can be entitled to the social card 23

24 C. Including the value of home ownership in income measures Empirically it is very difficult to measure all income components, e.g.: capital gains, unpaid work, in kind benefits, imputed rents from housing. Suggestion to add imputed rents to disposable income (Canberra Group 2011), but how to estimate IR? Rental equivalence method; Capital user costs; self-assessment. Limits: equivalence between IR and monetary income; estimates robustness. An alternative approach: the out of pocket approach being household a «basic need»=> subtracting expenses (rents, interests on mortgage and ordinary housing costs) from disp. inc. Estimates for 6 EU countries using EU-SILC

25 C. Changes in Gini of disposable income DE ES FR IT SE UK Disp. inc. Disp. inc. plus imputed rents Disp. inc. minus housing costs 25

26 C. Changes in poverty headcount DE ES FR IT SE UK Disp. inc. Disp. inc. plus imputed rents Disp. inc. minus housing costs 26

27 C. Exit out of poverty 28.0% 26.0% 24.0% Share of poor according to disposable income escaping from poverty according to the IR or the out of pocket approach 24.5% 26.2% 22.0% 20.0% 18.0% 16.0% 14.8% 18.7% 16.5% 17.5% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 2.4% 2.1% 5.0% 4.1% 1.8% 5.9% 0.0% DE ES FR IT SE UK Disp. inc. plus imputed rents Disp. inc. minus housing costs 27

28 C. Entry into poverty 12.0% 11.0% Share of non-poor according to disposable income dropping into poverty according to the IR or the out of pocket approach 11.6% 10.0% 9.7% 9.0% 8.0% 7.6% 7.7% 7.0% 6.0% 5.0% 4.8% 5.5% 5.2% 4.0% 3.3% 3.7% 3.0% 2.7% 2.7% 2.0% 1.6% 1.0% 0.0% DE ES FR IT SE UK Disp. inc. plus imputed rents Disp. inc. minus housing costs 28

29 C. Reranking from IR to O.o.P.: Italy 0 Out of pocket approach Imputed rent approach Imputed rent approach Out of pocket approach Total Total

30 Implications The inevitability of horizontal inequities The risks of re-ranking The risk of war among the poor and the weakening of social cohesion 30

31 Policy conclusions (1) Given all these dangers - the role of universal/quasi-universal transfers - ie a categorical transfer for all (almost all) children (Atkinson) - the role of targeting within universal services - a limited residual role for means testing 31

32 Policy conclusions (2) Certainly, the unavoidable weight of value judgments (absence of the perfect policy), yet against the widespread and somewhat blind support given to means testing, at least, recognition of - the many dangers of means testing - even in terms of target efficiency - the importance not only of monitoring social policies, but of transparency in the monitoring processes - seemingly minor methodological aspects (ie what does an index exactly measure?) are crucial to highlight the distributive effects (winners and losers) 32