Can rate cards improve results-based financing?

Last month, world leaders and global development experts gathered in New York City for the United Nations General Assembly and its side events, exploring some of the most persistent challenges our world faces today.

Funding — to end poverty, to create gender equality, and to provide equal access to healthcare — was an underlying theme throughout the week. There has been increasing interest in using results-based financing mechanisms to both increase social impact and draw in more resources towards social causes. This piece will explore the advantages and disadvantages of one tool, known as rate cards, that can be used by outcome payers and donors to potentially streamline measurement efforts and save costs.

Rate cards are meant to tie funding to impact. They also have the potential to create a more efficient market for impact. But many rate cards don’t live up to that promise. Instead, they hew close to traditional funding by tying payments to outputs and activities, rather than true impact. What if there were a way to make them truly transformational by only paying for the results we really care about, like lives saved? In this post, we discuss the possibilities — and challenges — of using rate cards in this way.

Benefits of rate cards

Rate cards provide a “menu” of outcomes, accompanied by “rates” that represent the maximum amount an outcome payer (such as a government or foundation) is willing to pay for each outcome. The outcome payer then contracts multiple service providers who agree to be paid the listed rates (or lower) if outcomes are achieved. A well-known rate card example is from the UK’s Department for Works and Pensions, focused on youth employment. The card specifies what the government will pay per person who achieves outcomes such as improved school behaviour, National Vocational Qualification level (awarded by completing training and assessments), and entry into employment.

Rate cards can help keep costs consistent across service providers, cut down on transaction costs, and make it easier to scale up results-based financing mechanisms as compared to the usual process with impact bonds. Because the work to define outcomes and their associated payments — for example, change in a child’s test scores or entry into employment — is done upfront, this cuts downtime and resources spent on this component of the evaluation.

In addition, the outcome payer can publish the rate card and contract multiple service providers through one request for proposals. Outcome funders already contract multiple service providers at once. However, in their current form, outcome funders do not define the specific outcome metrics and payment rates prior to the contracting process. Instead, they typically do so in conjunction with specific service providers and project managers. This can require additional cost.

Addressing potential pitfalls

Rate cards are currently under development for multiple sectors in the US and for young adult needs such as debt and housing in the Netherlands. A rate card approach has also been considered for the India Education Outcomes Fund. Each of these instances provides valuable examples that illustrate both the benefits and potential pitfalls of using rate cards.

In the UK Department of Works and Pensions example, the rate card only specifies payments on a per participant level. For every person who gets his/her first job, the service provider is paid £3,500. These outcomes are determined by a pre-post evaluation with administrative data. Service providers do not have to demonstrate a causal link between their program and the outcomes of their participants. This has been cited as both a potential benefit and limitation of rate cards — they may cut down on evaluation costs, but do so at the cost of paying for ineffective programs.

However, with a few design changes, one could combine the rate card with rigorous impact evaluation. For example, rather than paying £3,500 for each participant’s first entry into employment, one could pay per increase in job entry relative to a control group not in the program. For example, £3,500 for every additional person entering their first job per 100 people. The price would be very dependent on baseline job entry in the target demographic (it’s probably easier to raise job entry from 4 to 5 per cent, than to raise it from 74 to 75 per cent), but this issue is also present with a per participant rate card. Cream-skimming is an issue for many results-based financing (RBF) instruments — this is when implementing organizations select program participants who are highly likely to achieve the outcome. That is not the focus of this post and there are many other articles that get into the weeds about how to prevent this.

Another potential pitfall lies in the counterpoint of rate cards: if they were to gain popularity and become widely used, would this divert funding away from higher cost, but still high impact, programs? There are many programs that won’t pass the cost-effectiveness benchmark because of lower impact or higher operational costs, and generally that seems like a good thing. But if donors start putting out global rate cards, it could funnel money away from places where costs are high — like sub-Saharan Africa — and toward places where costs are lower, for example like South Asia and Southeast Asia. Different countries could have different rate cards, but even regionally, the same challenges could arise — rural areas where implementation costs are high might wind up under-funded. This could potentially undermine rate cards’ potential if donors don’t take these considerations into account.

The price is right

In order for rate cards to create an efficient market for impact, outcome payers need to select outcomes that capture improvements in people’s lives and get the prices right. While this isn’t straightforward, there are two types of outcomes for which rate cards can be more easily applied.

Prevention

Many results-based financing instruments focus on outcomes that save the government, or taxpayers, money: recidivism, homelessness, unemployment, chronic disease, etc. The government knows exactly how much it costs to keep someone in prison. Therefore, it seems straightforward to develop a rate card for prison years averted.

Existing cost-effectiveness evidence

There is growing literature on the cost-effectiveness of different interventions, which can provide benchmarks for rate cards. This focuses on how much a program costs relative to its effect size (or vice versa), and not solely on public savings. For example, a cost-effectiveness analysis conducted by J-PAL of multiple education programs in Kenya found a cost-effectiveness range of $2.87-$72.26 per additional test score standard deviation. One could potentially use the median cost-effectiveness of these programs as a benchmark, paying $28.06 per additional standard deviation achieved per student. A potential issue with this approach is that cost-effectiveness studies generally focus on marginal costs. For instance, if an NGO already rents office space to run other programs, then the cost of rent is generally excluded from cost-effectiveness estimates. However, rate cards may need to cover some of these costs to be attractive to service providers. And going back to the issue of cream-skimming, it may cost less to improve outcomes for populations that are easier to work with (e.g. working with high-income students versus low-income students). When reviewing cost-effectiveness estimates from the literature, it’s important to compare how contexts may differ from those in which the rate card will apply.

A thought experiment

Looking at the UK Department of Works and Pensions rate card, there’s a laundry list of outcomes, some of which map better onto impact than others. One argument for including intermediate outcomes, and even inputs, on a rate card, is that these will subsidize costs for the longer-term, harder-to-achieve outcomes. If it takes years to get program participants to sustain employment, service providers can at least get paid for improved school behaviour in the meantime.

However, the ultimate goal should be to improve/save lives. What would happen if someone developed a rate card with just one rate: a price per life saved (say $3,000 per life saved)? This could potentially improve service delivery for interventions we already know are life-saving, like vaccines and bed nets. Some of the original promises of RBF were that it would provide implementer flexibility, spur innovation, and crowd-in money for risky interventions. If implementers can focus on outcomes, will that give them more flexibility to experiment with approaches — for example coming up with novel interventions for saving lives, like lithium in drinking water. Or would implementers be less willing to take on risks because they’re responsible for the losses? Would a rate card approach only attract service providers who work on well-established drivers of mortality such as acute disease, and not on slower, more systematic approaches to saving lives? Regardless, this scenario providers for an interesting thought exercise and potential considerations for RBF in the future.