Written by Erin Crossett and Keir Bradwell

Water is a relatively new area of grantmaking for GiveWell, but we’re excited about its potential. Two billion people around the world lack access to clean drinking water, and unclean water is a major cause of illness and death, primarily through waterborne diseases such as diarrhea and cholera.

Within the water portfolio, we think about which specific programs in which specific places are likely to address these health burdens most cost-effectively, and what additional evidence we need to gather in order to make that determination.

In this blog post, we detail our current approach to our water portfolio, explore the areas we’re excited to investigate next, and share the work we’re doing this year to deepen our understanding of the sector. Through this work, we aim to make more highly cost-effective grants that bring clean water to many more people around the world.

Where are we now?

So far, our grants have focused on improving water quality, rather than access. This is because we think there is a strong link between water quality and health outcomes, and that other donors in the sector are generally more focused on access. Because water quality is relatively neglected, we think there is plenty of room for us to make an impact.

Based on our review of a recent meta-analysis, we estimate that chlorination, a common approach to water treatment, reduces all-cause mortality in children under five by 12%. (The meta-analysis itself reports a much larger effect; our reasoning for using a smaller effect size in our grantmaking can be found here.) This estimate means that in certain locations, the water quality programs we currently support look highly cost-effective, even relative to other opportunities in other sectors that GiveWell could choose to fund.

Since we began our water portfolio, our support has focused on two main interventions: chlorine dispensers and in-line chlorination. We funded Evidence Action to install chlorine dispensers in Kenya, Malawi, and Uganda and in-line chlorinators in Malawi, and to work with state governments to provide in-line chlorination in India. We have also funded exploratory work on providing vouchers for a free supply of chlorine, which can be redeemed in shops and health clinics. When we made these grants, we estimated that these programs looked anywhere between 4 and 22 times as cost-effective as unconditional cash transfers, our benchmark for comparing the cost-effectiveness of different funding opportunities. (When we recommended funding for chlorine dispensers, our cost-effectiveness threshold was lower than it is today.)

For more on the history of our work on water, see this blog post.

What’s next?

We think that water quality (rather than access) is likely to remain the most promising area in this sector for GiveWell. Expanding access to water is undeniably important, but such infrastructure projects tend to have very high up-front costs. In addition, they can be difficult to start and sustain in areas with the highest mortality burden, which tend to be rural communities in which people have little ability or willingness to pay for water. By contrast, devices like chlorine dispensers are relatively inexpensive, can be quickly set up in remote areas, and directly target the health outcomes that GiveWell focuses on, like mortality.

Our current understanding is that GiveWell is one of only a few funders of last-mile chlorination programs like these. This suggests that the grants we are considering are unlikely to receive funding from other sources, and therefore that the funding we contribute is making a real difference. However, there may be factors or alternate funding sources we are unaware of, and we plan to spend time over the rest of the year improving our understanding of this funding landscape.

There is still plenty of room for our grantmaking to evolve as we deepen our subject area expertise. Specifically, this year we hope to:

• broaden the range and location of organizations we work with in order to increase the amount of cost-effective grantmaking opportunities available to us
• reduce our uncertainty around the effect size of chlorination on mortality
• understand the relative promisingness of the chlorination programs we have focused on to date (i.e., chlorine dispensers, in-line chlorination, and chlorine vouchers)
• explore other kinds of water treatment, such as filtration or desalination

Expanding the number of chlorination implementers

So far, the programs we’ve funded at scale have all been implemented by Evidence Action. While we think highly of Evidence Action as an implementer, we also think there may be cost-effective grantmaking opportunities in areas where Evidence Action doesn’t currently deliver these programs (such as in Francophone Africa), and that increasing the number of organizations working on last-mile chlorination is a worthwhile goal in itself. In the places we think look most promising for expanding our grantmaking, we are not aware of organizations that currently deliver programs like chlorine dispensers at scale, and we’re keen to explore whether our funding could change that. If so, this could unlock lots of additional highly cost-effective grantmaking.

To that end, this year we’re seeking to find ways of broadening the pool of organizations that deliver chlorination programs in regions with a high mortality burden and low water treatment rates. This could involve:

• encouraging existing organizations in the sector to tailor their programs in ways that we think could increase their cost-effectiveness (e.g., by operating in higher-mortality areas, or targeting those most likely to benefit from chlorine, such as pregnant women),
• funding organizations from outside the sector to expand the range of their work
• providing funding to start a new organization

The fruits of this exploratory work may take a long time to materialize, but we hope to get some signal in the next year about whether additional organizations could implement cost-effective chlorine programs at scale.

Chlorination’s effectiveness

In addition to supporting chlorination delivery, we’re currently funding the Development Innovation Lab (DIL) at the University of Chicago to further study the effect of chlorination on mortality. To date, we are not aware of individual randomized trials of chlorination that have been statistically powered to detect a mortality effect; the meta-analysis we rely on combines multiple smaller trials for its estimate. Though we are confident that chlorination averts deaths, we are unsure of the exact magnitude of the effect. This mortality reduction is a key factor in our analysis of chlorination programs’ cost-effectiveness, which is the primary factor in our grantmaking decisions. If DIL finds a smaller or larger effect than we currently estimate, that could alter our future grantmaking plans significantly.

This year, we expect to make a funding decision on the next phase of DIL’s proposed study, which would fully fund an RCT of chlorine vouchers and in-line chlorination that is powered to detect all-cause mortality in children under two. Should we decide not to fund this phase, we would explore alternate options for improving our estimate of chlorine’s effectiveness, including other trial designs, evaluators, or research methods. We will also learn about the relative promisingness of chlorination in other ways, such as through ongoing monitoring and evaluation data from our existing grants, and through work we’re doing with Evidence Action to scope new chlorination opportunities, including piloting chlorine dispensers and chlorine vouchers in new countries.

The relative promisingness of chlorine dispensers, in-line chlorination, and chlorine vouchers

Important questions we’re considering are which approach to chlorination would be best under which circumstances, and whether one approach looks particularly promising overall.

Our initial hypothesis is that, all else equal, in-line chlorination may be the most promising way to get people to drink chlorinated water consistently. This is because in-line devices dose chlorine automatically, without requiring consumers to change their behavior: when end users turn on the tap as they usually would, the water they receive is already chlorinated. By contrast, chlorine dispensers require people to add chlorine into their water manually, while chlorine vouchers must be redeemed at a shop or health clinic. (For our full report into water quality interventions, see here and here.)

We plan to learn about the relative merits and drawbacks of each approach through the routine monitoring and evaluation data we receive from our grantees, through Evidence Action’s scoping of new opportunities, and through ongoing discussions with water experts. We also plan to conduct detailed assessments of the performance of our existing grants. These lookbacks will be based on the extensive program data we’ve received from Evidence Action, alongside updates to several parameters of our cost-effectiveness model, such as costs, leverage and funging, take-up of chlorination, and the underlying health burdens in program areas.

Alternative water treatment methods

At present, we think that chlorination is the most promising approach to water treatment from GiveWell’s perspective: it averts mortality, and it is cheap, widely available, and comparatively straightforward to implement. However, we have not explored alternative treatment methods in depth, and we know that communities in some parts of the world face water quality challenges that chlorine cannot address, such as rising salinity levels or chlorine-resistant parasites.

We are open to the idea that other forms of water treatment, like desalination or filtration, might look just as cost-effective as chlorination in some contexts. We hope to investigate grants to support these approaches as our capacity allows. These grants would be valuable both for their direct impact and for what we’re able to learn about the intervention by tracking the grants over time. We’re also speaking to experts to learn more about the most promising opportunities beyond our current portfolio.

***

Our team’s agenda for this year will help to create future cost-effective funding opportunities, increase our confidence in the mortality effect of chlorination, better understand the trade-offs between different chlorination interventions, and begin to explore alternative treatment methods in depth. We think this agenda is well-suited to our goal of making the most cost-effective grants we can in this sector. Ultimately, these grants will provide safe drinking water to some of the poorest communities in the world, and we’re looking forward to learning more about the best ways to scale these programs to many more people.

If you work on improving water quality in low-income countries, particularly as an implementer of chlorination programs, we’d love to hear from you. Email us at info@givewell.org.