Click here to jump to Kerrie Markou's case study of the Work Programme and its payment-by-outcome system

Paying public service providers for the social outcomes they deliver, not the effort they exert, is an attractive idea. It is fair. Nobody makes money for delivering services that do not make a difference to people's lives. It reconciles the desire to give professionals freedom to do their jobs effectively with the need to increase accountability. At a time of austerity, it can be combined with upfront investment in services by private providers, so that government does not have to spend any money until savings are made. It seems like the silver bullet.

Except it is difficult to do. Many conversations about payment-by-outcome are so predicable they seem rehearsed. "It's a great idea in theory, but… How do you know what impact providers have made and what would have happened anyway? How do you avoid creating perverse incentives? What if the outcome is impossible to measure?" Indeed, some critics claim it is impossible. In 1996, Henry Mintzberg, Professor of Management Studies at McGill University in Canada, wondered, "How many times do we have to come back to this one until we give up?"

An attractive prospect

Yet governments around the world remain tantalised by the possibility that payment on delivery can ensure important policy objectives are achieved. In employment services, payment-by-outcome has been developing for 30 years, since the Job Training Partnership Act was passed in the US in 1982. In the 1990s, the Australians took the lead in developing the approach. Now, building on the foundation laid in the last decade and with the enthusiasm of a new government, the UK is unquestionably the market leader. Clearly, there are many who believe the benefits of payment-by-outcome justify the search for ways to overcome its challenges.

A range of tools

A new report, Payment by Outcome: A Commissioner's Toolkit, published by the 2020 Public Services Trust, extracts the lessons learned in employment services, chronic disease management and offender management to assemble a set of tools commissioners can use to design effective systems. The study revealed that successful payment-by-outcome relies on a much wider range of tools than just financial incentives.

Reputational incentives, professional norms and regulation are important drivers of behaviour that commissioners can deliberately reinforce, and which may help counteract perverse financial incentives. Australia's 'star rating' system gives welfare to work providers a score based on the numbers of employment outcomes they deliver, weighted towards outcomes for the very disadvantaged. High-performing providers, as indicated by these ratings, may have their contracts renewed without a competition. While the payment system may not be quite right in terms of incentivising providers to work with clients who are not job-ready, this reputational incentive helps to curb cream-skimming and parking. In the UK, providers and the Department for Work and Pensions (DWP) have worked together to establish mechanisms to govern the market, and the DWP Supplier Charter, the Code of Conduct and the Merlin Standard have become important guides for providers in terms of setting the standard for acceptable conduct.

Transferring risk

Moreover, financial incentives can be deployed at varying intensities. It is often assumed that the more risk government transfers to providers, the better. In the current fiscal environment, the temptation to demand that providers carry all of the risk is even greater. However, lower-intensity incentives may still motivate providers to deliver outcomes and may stimulate more innovation, as providers can afford to try radically different ways of working. Indeed, researchers have found evidence that high-stakes performance incentives may cause 'choking under pressure', particularly where tasks require creativity or problem-solving. Commissioners need to understand the reason for using payment-by-outcome and design systems accordingly.

Above all, commissioners must remember that the system is populated by human beings who have their own agendas. It is unlikely that commissioners will design a perfect system the first time, since it is impossible to anticipate how people will respond. The evolution of performance measures under federally funded job training programmes in the US indicates the importance of continuous evaluation and commissioners' willingness to make changes. As a result of commissioners observing unintended behaviour on the part of providers, a measure of providers' cost-effectiveness was phased out. The employment rate measure was also changed so that rather than being taken on the date participants completed the programme, it was measured six months later. What will ultimately decide the success or failure of the experiment is commissioners' ability to learn from experience and adapt to changing circumstances, adjusting systems over time so they deliver better outcomes.