By Matthew Eldridge, Sarah Guminski
This is the second of a two-part blog series considering criticisms of pay for success (PFS). This series builds on an opinion piece written by our Initiative’s co-director, Justin Milner, which addressed four of the most common criticisms of PFS. The first post considered three additional criticisms and this blog offers some concluding thoughts on the risks and challenges of PFS.
Pay for success (PFS) is attracting interest because it offers the opportunity to overcome longstanding obstacles to more effective social service delivery. As this blog post discussed, it has received scrutiny from critics who have raised concerns over the rationale of the model, the challenges of designing and implementing PFS projects, and potential negative unintended consequences. In many cases, these risks of PFS are actually the risks of doing PFS poorly.
Many of these critiques can be satisfactorily addressed by a thorough and transparent planning process—two of the good governance qualities that the PFS process itself hopes to promote. Some criticisms object to the concept of leveraging private money to achieve public outcomes. On these more ideological questions, PFS is merely one more element of an already animated discussion on the role of public-private partnerships, and addressing this sufficiently lies beyond the scope of this Initiative.
All this notwithstanding, there are valid practical concerns that merit a close look by those planning a PFS project or considering the field as a whole. One notable point of caution is that while PFS is an appropriate financing mechanism for many types of programs, it is not a long term solution to the problems that these programs address. A PFS-funded intervention that supports formerly incarcerated people, for example, might help the program participants find jobs and avoid re-offending, but it will not address the many systemic issues with the criminal justice system. Private investment also will not, and should not, replace taxpayer-funded government services. However, the exercise of contracting on outcomes will hopefully shift government focus to how well people are served (outcomes) instead of how many (outputs).
Furthermore, not every jurisdiction, problem, service provider, or intervention is a good fit for PFS. These projects can carry significant transaction costs, require stakeholders to have frank discussions about strategic objectives and challenges, take years to deliver final results, and pose a number of other challenges. For some governments, there may be other, more cost-effective ways to fund the same intervention.
The largest risk of PFS is that at a larger scale, ultimately it changes nothing: that it doesn’t promote broader evidence-based decision making, or a culture of meaningful outcome setting and measurement. Anecdotal evidence thus far provides reason to believe this is false, and that PFS really can yield broader systems changes, or at least provide a model from which governments can learn. Many stakeholders believe that the future of PFS lies in this commitment to aligning performance evaluation with long-term goals, and that traditional methods of financing are not flexible enough to accommodate the complexities of society’s most pressing challenges.
Given the newness of PFS, we are only just beginning to accumulate sufficient evidence to assess whether the PFS model is an effective way to scale evidence-based solutions. That being said, we know that in many cases the current system isn’t working. If governments continue to fund programs with no evidence as to their effectiveness and avoid investing in innovation due to the financial risks of failure, there remains a strong rationale for exploring PFS’ potential to address unresolved social challenges.
Have a Pay for Success question? Ask our experts at PFSsupport@urban.org!
As an organization, the Urban Institute does not take positions on issues. Scholars are independent and empowered to share their evidence-based views and recommendations shaped by research. Photo via Shutterstock.