Social Impact Bonds (SIBs) are supposed to be about aligned win-win-wins for commissioner, provider and investor. In practice, however, connections between the business case, underlying logic and SIB structure can become disconnected. The more complex the SIB and its operating environment are, the more unpredictable tensions can get. This is even more true if the data foundation isn’t strong. Given the data focus of a SIB is often new to some stakeholders, this can prove a particular challenge in new projects, such as the one to support some 3000 social isolated and lonely over 50s in Worcestershire 2015-21.
Locally commissioned SIBs are more complex if they involve different commissioners (Worcestershire Local Authority and 3 CCGs in this case) and if they include a top-up contribution from both central government (Social Outcomes Fund) and The Fund. Top-up funding can stimulate commissioners to carry on engaging, but it can also distort understanding of what works – and what doesn’t – in a SIB.
Investors (MacMillan, Age UK, Callouste Gulbenkian, Care & Wellbeing and NESTA) helped keep this SIB together: they ploughed in cashflow to support delivery and learning-based course-correction. This readiness to bear risk helped build better connections between the SIB case, the service delivered, and the people using the service to access local communities.
Some assumptions in Reconnections’ business case proved over-optimistic, and so some correction was needed. The SIB’s architects, Social Finance (SF), had rightly set minimum-need thresholds for self-referral (to rule-out skewing payments towards easy wins).They hadn’t, however, reckoned on much more-intense-than-expected need among self-referrals, nor the low rate of GP referrals (also a challenge for other social prescribing projects). Nor had they fully factored-in community support, rural transport and accessibility costs.
SF designed payments for the outcome of a reduction in loneliness using an evidence-based R-UCLA scale. Practitioners worried that the three ‘how do you feel right now’ snapshots of a single client’s loneliness of the UCLA model produced a distorted picture of progression. This was mitigated by payment triggers using averaged score changes. These averages smoothed out snapshot fluctuations. The lack of a suitable counterfactual or comparator (to check result attribution to the intervention), however, remained a challenge. Without it, a big assumption was made that outcomes were attributable to the SIB’s interventions.
Reconnecting the system
Due to investor interests, close attention is paid to data in SIBs. Stakeholders respond to early data and Key Performance Indicator warnings, signalling cashflow and risk-management issues. In theory, this might lead the SIB to drive active ‘performance’ management than other commissioning arrangements. Given the need to keep cash flowing, management response to warnings can put pressure on delivery teams and can lead to hasty jumping to re-negotiate payment terms. This can cause Voluntary Community and Social Enterprise (VCSE) delivery provider staff stress. It also risks disconnection from service users and the SIB’s business case logic.
In Reconnections, the investor response was more thoughtful, combining proactive shared learning with patience for investment returns. New capital was injected before outcomes were achieved and paid for, which sustained delivery partners while changes were implemented. SF’s new director worked with the AGE UK project manager to appoint specialist case workers to take on more intensive cases and supervise the less complex ones triaged-on to volunteers. The recruitment approach and volunteer retention support were also overhauled to ensure better connection with the project aims.
This made for far more effectively configured connections between SIB theory and local practice. Despite this, three VCSE providers were decommissioned, and the outcome tariff was ultimately changed. The re-think process established an acceptable higher-cost case, which is cause for some encouragement. The fate of the three providers, however, could have been different if the business case had been better designed, and more time was allowed for capacity to understand the data and make changes.
The key lesson, set out in the CBO evaluation report, is that in-principle modelling can fail to account for real life complexity, especially when there is no counterfactual to compare the project to. Future SIB funders should consider whether SIB stakeholders have capacity to produce, curate and use accurate, actionable data. This might help achieve that theoretical “win-win-win” ideal and fully meet the needs of the people supported by each project. It may be the reconnection that’s needed in the system, so a SIB can be “the catalyst or trigger that can kick a system [of actors] into doing something differently”.