UNCOUNTED Measurement as power (and the political choices for post-2015) Alex Cobham OxFID, 22 February 2014
Pop quiz Which measure would you prefer? P = y/x G = 58.1y 119.5x + 41.86
Overview: 50 ways to be uncounted Blatant uncounting Hidden by design Choice of reporting Manipulation of data and targets Patterns of being uncounted Choices for Post-2015 Choice of development priorities Choice of indicators, e.g. the Palma Modest proposals
Blatant 1: Hidden by design E.g. illicit flows Individuals and businesses Jurisdictions Bilateral data? E.g. Banking - to BIS not public Trade, investment only highly aggregated public
Blatant 2: Choice of reporting E.g. political suppression The Black Book of Sudan: Who holds power?
Blatant 2: Choice of reporting 140 120 Development expenditure per capita, 1996-2000 (North=100) Female IMR, 1993 (North=100) 100 80 60 40 20 0 North Central, ex Kh East Darfur Kordofan
Blatant 2: Choice of reporting E.g. political suppression The Black Book of Sudan: Who holds power? Data provision: to IMF not citizens (then; now)
Blatant 2: Choice of reporting The power to choose what is (counted and) reported can exacerbate inequality
Blatant 3: Manipulation of data 1990 revisionism of targets (Thomas Pogge) Language Baseline year 1996 Baseline, millions 1,656 World Food Summit Target 2015 reduction 50.0% Target for 2015, millions 828 Required annual reduction 3.58% Halve number MDG-1 as adopted Halve fraction of world population 2000 1,665 40.4% 993 3.39% MDG-1 as revised Halve fraction of dev g population 1990 1,813 27.0% 1,324 1.25%
Blatant 3: Manipulation The power to revise results can render targets meaningless The power to revise targets can eradicate ambition
Patterns of uncounted Uncounted, through power Top end of distributions (countries, people); Criminality Uncounted, through lack of power Bottom end of distributions; Marginality
Technical Criteria: Choice of indicators Pigou-Dalton transfer principle: rules out counter-intuitive responses to transfers Technical accuracy Political accountability Income scale independence: should not respond to proportional changes to all (Cost?) Dalton s principle of population: measure shouldn t respond to merging of identical populations An example: Inequality Anonymity or symmetry: measure is independent of any non-income characteristic of individuals Policy Atkinson axiom: Are the value judgments of using this indicator sufficiently explicit? Policy-signal axiom: Is it clear what signal given to policymakers on direction of change? Clarity axiom: Is it clear to a non-technical audience what has changed? Policy-response axiom: Is the policy response sufficiently clear? Decomposability: overall ineq is related consistently to ineq among sub-groups A horizontal axiom: Can measure also capture horizontal inequality?
Accountable counting? The Palma Ratio of national income shares: top 10% to bottom 40% Rests on Gabriel Palma finding: stability of middle deciles (5-9) Basically, it seems that a schoolteacher, a junior or midlevel civil servant, a young professional (other than economics graduates working in financial markets), a skilled worker, middle-manager or a taxi driver who owns his or her own car, all tend to earn the same income across the world as long as their incomes are normalized by income per capita of respective country.
Middle stability across countries 35% 30% Highest 10% Lowest 40% Middle 50% 25% 20% 15% 10% 5% 0% 1990 2010 Combined
Middle stability across time UK Venezuela Peru Paraguay Panama Mexico Honduras El Salvador Ecuador Dominican Rep. Costa Rica Colombia Chile Brazil Argentina Top 10% Middle 50% Bottom 40% 0% 2% 4% 6% 8% 10% 12% 14%
Stability across income stages 100% 80% 60% Top 10% Bottom 40% Middle 50% 40% 20% 0% -20% Argentina Brazil Mexico Peru
Most and least equal income distributions, 2010 Country Palma Gini Zambia 4.77 57.49 Colombia 4.52 55.91 Paraguay 3.73 52.42 Panama 3.63 51.92 Rwanda 3.22 50.82 Serbia 1.08 29.62 Montenegro 1.02 28.58 Belarus 0.98 27.7 Ukraine 0.89 25.62 Romania 0.80 24.24
Palma vs Gini Decile Income shares (%) 1 6.25 4.17 3.13 2.50 2.08 1.79 1.56 1.39 1.25 1.14 2 6.25 4.17 3.13 2.50 2.08 1.79 1.56 1.39 1.25 1.14 3 6.25 4.17 3.13 2.50 2.08 1.79 1.56 1.39 1.25 1.14 4 6.25 4.17 3.13 2.50 2.08 1.79 1.56 1.39 1.25 1.14 5 10 10 10 10 10 10 10 10 10 10 6 10 10 10 10 10 10 10 10 10 10 7 10 10 10 10 10 10 10 10 10 10 8 10 10 10 10 10 10 10 10 10 10 9 10 10 10 10 10 10 10 10 10 10 10 25.00 33.33 37.50 40.00 41.67 42.86 43.75 44.44 45.00 45.45 Palma 1 2 3 4 5 6 7 8 9 10 Gini 0.23 0.35 0.41 0.45 0.48 0.49 0.51 0.52 0.53 0.53
Palma vs Gini If Palma so closely related to Gini; and Palma ignores half the income distribution; why would you possibly want to use it? 1. Policy axioms 2. Gini flaws P = y/x G = 58.1y 119.5x + 41.86
Policy Atkinson axiom: Are the value judgments of using this indicator sufficiently explicit? Policy-signal axiom: Is it clear what signal given to policymakers on direction of change? Clarity axiom: Is it clear to a non-technical audience what has changed? Policy-response axiom: Is the policy response sufficiently clear? A horizontal axiom: Can measure also capture horizontal inequality?
Pop quiz Which measure would you prefer? P = y/x G = 58.1y 119.5x + 41.86
Choice of development aspects Millennium Development Positive Goals: Decisions by Technocrats Pragmatism Inertia? Post-2015 HLP Negative MDGs Gender Inequality Disaggregation & Illicit flows Economic inequality
Choice of development aspects High-Level Panel: We recognized that every country is wrestling with how to address income inequality, but felt that national policy in each country, not global goalsetting, must provide the answer. History also shows that countries tend to have cycles in their income inequality as conventionally measured; and countries differ widely both in their view of what levels of income inequality are acceptable and in the strategies they adopt to reduce it.
Choice of development aspects Thought experiment: We recognized that every country is wrestling with how to address gender violence, but felt that national policy in each country, not global goalsetting, must provide the answer. History also shows that countries tend to have cycles in their gender violence as conventionally measured; and countries differ widely both in their view of what levels of gender violence are acceptable and in the strategies they adopt to reduce it.
Choice of development aspects Alternative approach: Global process defines (global) political priorities; National (and local?) process sets targets i.e. if it s too political to impose targets, don t; instead, set targets through a national process. But global process can require a target be set.
Conclusions Being uncounted reflects power: the excess or the lack thereof. Decisions on statistics are fundamentally political, nowhere more so than development this cannot be ducked. Post-2015 offers great opportunities for progress (not least disaggregation, illicit flows); but not without recognising and confronting the (political) obstacles.
Inclusive politics (as far as possible) Indicators chosen for accountability Data to follow ambition (not v.v.) Funding now (NSOs <0.1% of aid)
Remember: counting is political counting is power0
Thank you cgdev.org/europe @AlexCobham