Transcript Statistics Users Forum November 2009 Geoff Mulgan, Director The

Statistics Users Forum
November 2009
Geoff Mulgan, Director
The Young Foundation
Young Foundation – combines empirical research, local
collaborations, creation of new organisations in
education, health, democracy - and now developing
global field of social innovation
Metrics help governments and societies to see, hear
and make judgements – part of the cognitive toolkit of
society.
19th century governments measured their success by
their military prowess
20th century governments by GDP growth
21st century governments will primarily judge their
success in terms of progress in well-being and human
flourishing.
Long-term evolutionary trends…
• from taxable things (buildings, animals, people) for
government to measure society, to measures for society to
judge itself and what government is doing (such as school
exam results, or corruption)
• from physical objects (steel, wheat etc) through aggregate
concepts (like GDP) to intangibles (such as innovation or
creative industries)
• from single measures to indices (poverty, HDI, Indices of Civic
Health)
• from activities to outputs, to true outcomes (QALYs and DALYs,
carbon reductions etc)
• from material to subjective measures (fear of crime, patient
satisfaction &c)
Yet many of the most important problems not
caught well by the numbers: emissions, isolation,
waste, drugs, obesity - and subjective perceptions
...
Is what matters measured?: perceptions of ‘what is
important’ across 27 EU countries
The relationship between
democracy and well-being
All measures evolve in shifting relationship
with the policy tools used by governments…
• formal quantitative targets in strategic plans (US states,
UK government) – which helps to sharpen up policy,
strategy and accountability for results
• open coordination methods of the kinds used by the EU
• performance management methods used by higher tiers
of government on lower ones
• metrics used to support quasi-markets (eg enabling
incentives for outcomes)
• metrics as part of deeper conversations with the public
about priorities and policies (eg Oregon Benchmarks and
equivalents).
Lessons : many advantages from measurement,
transparency &c but:
• Metrics risk gaming, distortion, inflexibility etc and
diminishing impact. Hence the need for constant
refreshment.
• Metrics risk being technocratic (eg measuring
everything in services except what the public care
about)
• Metrics risk wrong levels of granularity – when
increasingly the most valuable information comes
from disaggregation rather than aggregation
GDP as example
• 1930s-1940s: concept and measurement of GDP
brought counterintuitive systemic thinking to
interactions of consumption, production,
investment and savings
• 21st century requires another step in systemic
thinking seeing economies as dependent on other
systems: ecology, family, public goods
• GDP adapted to better reflect true value of finance,
household activity, transaction costs &c
UK distribution of income –
average up, stretched at the top ,
sagging at the bottom
UK distribution of psychological well
being – most contented but long and
thickening tail of unhappiness,
loneliness and stress.
Source: BHPS 2007/0
Prosperity
Self-harm and
suicide in
children of
rich
Isolated homeowning older
person
Need
Material
wellbeing
Homeless
offenders
with
addictions
Victims of
trafficking
Need
AB
employed
good mix
of
protective
factors
DE
employed
good mix
of
protective
factors
Basic state pension
high social support High social
capital
NINA strong
immigrant
family
communitie
s
Prosperity
Psychological Wellbeing
Dissatisfied with high income
Area: ‘Aspiring households’ (2.6)
Area: ‘Asian communities’ (2.2)
Area: ‘Settled in the city’ (2.0)
Couple without dependent children (2.0)
Hardly talks to neighbours (1.8)
Poor/ v poor health (1.7)
Life is full of opportunities – disagrees (1.7)
Long term sick and disabled (1.5)
Separated from spouse (1.4)
Dissatisfied with low income
Unemployed (4.0)
Area: ‘Public housing’ (3.0)
Dislike neighbourhood (3.0)
Work is driving/travelling (2.4)
Used social worker (2.4)
Long term sick/ disabled (2.3)
Separated from spouse (2.0)
Caring for sick/disabled in household (1.9)
Lone parent (1.8)
Aged 16 to 20 (1.7)
Full time family carer (1.6)
Area – ‘Asian community’ (1.6)
Rarely involved sports (1.5)
Life is full of opportunities - disagrees (1.7)
Satisfied with high income
Most proportionately represented among
available variables
Have a degree (1.8)
Area :‘Thriving suburbs’ (1.8)‘
Area: ‘Aspiring households’ (1.5)
More likely to vote Conservative (1.5)
Caring for sick disabled in household – less likely
(0.5)
Satisfied with low income
Unemployed (1.8)
Full time family carer (1.6),
Single elderly (1.5)
No qualifications (1.5)
Retired (1.4)
Widowed (1.4)
Who is at the bottom?
4%
No one to comfort me if I’m
really upset
11%
No one outside household
to help if I’m depressed
2% lack all three types of
support
3%
No one appreciates me as a
person
Percentage responses to questions on emotional support 2007/08
Source: BHPS analysis
Less visible trends:
anxiety and depression
14%
12%
10%
8%
6%
4%
Anxiety, depression etc
Extrapolated trend
2%
0%
1985
1990
1995
2000
2005
2010
2015
2020
2025
Or raising new questions – like correlations
between happiness and blood pressure
•Headline numbers provide relatively little
information - greatest insight comes from
disaggregation, by place, class, gender, age etc –
seeing the surprises, the patterns
•GDP part of a system of causal relationships
which required aggregate numbers – well-being is
not
•Benefits from common architectures , platforms
and comparability, but not from replacing GDP
with another number
•Growing importance of subjective measures healthy and unavoidable, but lack of adequate ones
The UK national indicator set
198 national
indicators
Comprehensive area
assessments
At local level greatest insights come not from
headline figures but from relevant comparison plus
disaggregation by domain, by group or service &c
Young
What else is needed?
i) New tools for segmenting populations by
lifestyle, culture and attitude as well as
socioeconomic status
• Key to understanding culture and behavior change,
and the efficacy of many policy tools in critical
fields such as healthcare and the environment (eg
around obesity, recycling, learning).
• Yet the science of segmentation remains in its
infancy, and the field is dominated by commercial
companies not public statisticians.
ii) new tools for measuring the potential value of
public, social and service innovations at different
stages, a critical gap in current measurements
and vital for testing policies for well-being.
Iii) new means for citizens to mash, mine and
adapt data-sets (eg http://mashupaustralia.org/)
iv) New tools that assess current levels of wellbeing but also dynamics, resilience and future
readiness …
self
support
system
self
support
system
current state
strengths
vulnerability
Well-being,
income,
health
Self-efficacy,
assets,
resilience
Mental illness,
poverty, risk
factors
Daily care,
Family, social
support (time), networks,
provision
services
Isolation,
exclusion
GDP, job
creation,
rights,
governance
Poverty, weak
governance,
dangerous
environments
Government
capacity,
dynamic
economy,
future
readiness
• Enables assessment of resilience to shocks at
different levels – personal, community, system
- and analysis of policy tools to reinforce
strengths and address weaknesses
.
Emerging tools at root about democratizing
and humanizing data …
• making it fit better with lived experience and public
priorities (what matters to citizens);
• opening up a broader range of sources and
engagement (engagement of citizens);
• enriching data, not replacing one aggregate calculator
with another (making numbers useable by citizens)
•and turning metrics from a tool of power over society to
a tool of power with society.