From 6-8 July, the SAP Institute for Digital Government is participating in the 23rd European Social Services Conference in Lisbon. This event is conducted by the European Social Network (ESN). The conference is organised annually by the ESN, an independent learning and development forum for local public social, health, education and employment services in Europe and associate member organisations. SAP is a proud sponsor of this important event. See more at http://conference.esn-eu.org/theme-lisbon.
As social protection funding sits front and centre within the current EU financial crisis, there are ongoing questions about achieving effective social outcomes from what is spent and how this investment can make a positive contribution to economic activity. Too often social protection funding is seen as an unproductive cost to the economy rather than as an investment in human capital. Even within the traditional social insurance based models where the principles of acquired rights and solidarity are enshrined in law, there is increasing demand for the social protection system to contribute to active ageing and/or labour market participation wherever possible.
To achieve this, there is growing interest in the use of predictive analytics, especially in those areas of the social protection system where ‘an ounce on prevention can save a pound on the cure’. The IT industry is aggressively promoting the potential of leveraging digital data. Analytics tools and cognitive computing is delivering clever and even self-learning algorithms to predict things such as a person’s risk of long term unemployment, the risk of serious harm for children from child abuse reports ,the risk of compliance failure or fraud from within a business process, the likelihood of return to work after a workplace accident and even the likelihood of better life outcomes from targeted interventions for people with a disability.
The Government of New Zealand is a pioneer in this field with their work on predicting the lifetime cost to the state for particular cohorts from a life on welfare payments. For several years now the government has been socially investing in targeted interventions and support services to alter the predicted pathways, for example providing support to teenage mothers to help them finish their year 12 schooling so their likelihood of entering the labour market once their child can go into care or start school are vastly improved. This results in a win-win situation with better social and economic outcomes for individuals and reduced lifetime welfare liabilities for the state. The new National Disability Insurance Scheme in Australia is following a similar model to help people with a disability improve their potential for labour market engagement through targeted social investment at the individual level.
During the ESN conference in Lisbon, I am leading a workshop titled ”Managing risk in a predictive manner to enhance decision making based on real-time analysis of data”. Following is the workshop synopsis: The social protection industry aims to reduce poverty and vulnerability by promoting efficient labour markets, diminishing people’s exposure to unemployment, exclusion, sickness, disability and old age, and enhancing their capacity to manage these risks. Many social protection agencies are promoting a “digital first” service delivery model, through which citizen needs can be addressed in real-time and people can be empowered to help themselves using online and mobile technologies. This workshop will examine the case for real-time analytics within the context of the social protection industry and will discover the balance point for ensuring technology enhances the professional judgment of case workers rather than obstructing it.
The workshop attendees will be asked to consider a new form of moral hazard arising from these predictive models. Moral hazard, an economic term from the insurance industry, finds its way to the social protection industry in a similar manner – ie. people undertake more risky behaviour on the basis they are insured, over and above what they would normally take. For example a worker becomes less diligent on safety issues on a work site because he knows he is covered by labour accident insurance if something untoward should happen. Essentially risk is transferred to someone else (the social fund) thereby adversely modifying the behaviour of the insured person.
Using predictive models in social protection brings an increased risk of behavioural change amongst case workers as they begin to (over) rely on the accuracy of the machines and withdraw from exercising their professional judgment. They are at risk of regarding the machine based predictive models as an insurance like safety net for their professional judgment. Decision making risk is transferred from the human case worker to the machine. This raises the question when a case worker withdraws from exercising their professional judgment and relies on the predictive models for decision making, where does accountability lie when things go wrong – for example a false negative from the predictive model on the likelihood of long term unemployment risk and the case worker does not exercise professional judgment leading to the person not being offered appropriate interventions at the right time.
While the proponents of predictive models will point to the evidentiary base to support the accuracy of the models, the question arises to the role of professional judgment within a people orientated social protection system seeking to achieve better social and economic outcomes. While predictive analytics is largely a scientific based domain of mathematicians and data scientists, social protection is a function of human behaviour and relationships.
While there can be little doubt these predictive models can make vast improvements in efficiency and effectiveness to social investment decision making at the individual level, how they become integrated into the human dimension of the social protection system will determine their ultimate value. As the world goes evermore digital, predictive analytics will be a powerful tool for optimising social protection expenditure in terms of better outcomes – the New Zealanders have proven it. At the same time it is prudent to explore new risks, such as moral hazard, arising from this approach. The European Social Network annual conference is the ideal forum for these discussions to take place.