Nothing New Under the Sun
At least since the London School of Economics was founded in 1895, and especially since it established a Department of Social Administration in 1912 (with the dual purpose of carrying out investigations of social conditions and training of social workers), there has been a long history in social policy of unashamedly values-driven concerns to find empirical and evidence-based solutions to poverty, inequality and the great social problems of the day.
Speak Now or Forever Hold Your Peace
So although ‘social investment’ is not a new idea, the previous government got into trouble with the social services sector with its ‘big data’ approach, and in particular, for its policy of ‘personal data for funding’. The new Minister for Social Development, Carmel Sepuloni, recently announced national consultations on the current government’s new approach, which looks like being re-branded as “Investing for Social Wellbeing”. These consultations, which we are now half way though, also aim to deal with the thorny issues of how to best protect and use personal information. This is the consultation, whose absence our sector rightly complained about under the previous government - so we need to be well represented at these hui and present strong and well-thought out positions.
The dates and places for the consultations are set out on https://sia.govt.nz/our-work/yoursay - unfortunately not all the venues are listed so don’t be afraid to email the Social Investment Agency for more details. In addition, on the same website you’ll find links to two surveys: Investing for Social Wellbeing and Data Protection and Use. There is a risk that the more concrete issue of Data Protection and Use, may dominate both in attracting NGOs survey responses and in the local hui discussions.
Decency and Data
Last year, Social Service Providers Aotearoa, NZ Council of Christian Social Services and many others made it clear the kind of principles that should define the collection of personal data from some of the most vulnerable people in the country:
(i) what data are collected should be collaboratively determined by funders, NGOs, consumer representatives and others involved, and be independently audited by the Privacy Commissioner
(ii) only the minimum data necessary for justifiable purposes should be collected, and any data that is not able to be analysed and fed back to those who provide it within six months or so should not be collected
(iii) generally any data collected should be ‘anonymised’ (identifying details removed) unless there is an exceptional need, and only with the people’s explicit permission, freely given (not under threat of not having funded services they need)
(iv) a neutral, trusted agency with clear data protection culture, policies, and practices, such as Statistics NZ should be the repository for personal data – and it needs to be adequately funded to ensure proper protections can be maintained.
Previous sledge-hammer attempts to collect personal data under threat of withdrawing funding for needed services assumed NGOs were at best reluctant and obstructive or at worst the enemy. NGOs share a deep concern for effectiveness and making sure they have the greatest impact with the limited resources available to them. For funders to act otherwise, risks only reinforcing the worst behavior.
Redefining Social Investment
But its worth also making sure we equally turn our mind to the more abstract issue of re-defining social investment, as this can have fundamental impacts on what is funded and how for many years to come.
The high level principles expressed in the Investing survey and proposed for the new approach, are pretty difficult to argue with: put people at the centre; make better use of a wide range of information; work in partnership and build trust; and provide clear goals and robust measures.
In responding to these unobjectionable, high-level principles, it will be important not to lose sight of what was lacking or at great risk in the previous government’s approach.
Social investment is great when it means investing more resources up-stream, preventing social ills from developing or nipping them in the bud early on; it’s not useful if it just means increasingly narrower targeting of help or resources for a smaller group of a stigmatised ‘most vulnerable’ (like Predictive Risk Modeling, or our own real-life, unreliable version of the movie “Minority Report”).
Social investment is great when it helps us identify how to improve well-being for more people and society as a whole; its not useful when it is just a means to identify where to save funding and make cut backs.
Social investment is great when it addresses social (systemic) causes of undesirable outcomes; it’s not helpful if it’s just a more sophisticated way of ‘blaming the victim’ and individualising what are fundamentally societal problems.
Social investment is great when it monitors impact at a community-wide or even national level to make informed judgments about policy approaches; it’s not so useful when it tries to measure impact on an isolated, local NGO-basis. (The Social Investment Agency’s work on the social impact of public housing is a good example of it being used well.)
Social investment is great when it hears and responds to the voices of those directly experiencing social problems, and those grass roots workers, whanau and friends closest to them (bottom-up); its not useful if it gives most weight to distant ‘experts’ and large scale empirical averages (top-down).
This also means social policy will need to become more comfortable dealing with ambiguity, diversity and even apparently contradictory tensions. The elusive search for certainty and simple answers in social policy is more likely to lead us to be ‘precisely wrong’ rather than ‘roughly right’ in our responses. (Meth testing homes comes to mind.)
Staying ‘roughly right’ rather than lapsing into ‘precisely wrong’
Far too often in Western-dominated cultures, we privilege numbers over words. Numbers are great at summarising, standardising and reducing large amounts of information into more manageable (but potentially over-simplified) chunks, while words are best when we need to understand the complexity, inter-connectedness, diversity and fine-grained nuances associated with real lives and communities. The ‘small data’ of personal stories and interactions are at least as, if not more important than the ‘big data’ of mass collections.
As mathematician, Cathy O’Neill (2016) points out in “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”, big data’s tools are not neutral; they can be just another form of ‘racial profiling’, reinforcing pre-existing inequalities and discrimination and a ‘toxic cocktail for democracy’, because, although the algorithms are often opaque and difficult to contest, their mathematical simplicity means they are highly scalable, thereby amplifying any inherent biases across increasingly larger populations.
Predictive Risk Modeling (PRM) fits like a glove the neo-liberal agenda of an ever-shrinking state, ever-tightening targeting aimed squarely at pathologising families and victim-blaming, wrapped up in actuarial mumbo-jumbo. Ironically the work on PRM on child abuse and neglect in Aotearoa New Zealand has (inadvertently) identified the need to prioritise tackling the social problem of poverty if we want to prevent child maltreatment across society. While the big picture is clear, the predictive value of PRM on an individual level (which is how it was proposed to be used) is pretty fast and loose – with up to 69% false positives (more than two-thirds of those targeted not actually having substantiated maltreatment), almost as many false negatives (just over two-thirds of those with substantiated maltreatment not having been targeted), and Māori children being inaccurately targeted at a higher rate than non-Māori children.