#37 - GiveWell picks top charities by estimating the unknowable. James Snowden on how they do it.

Organization Overview and History

• GiveWell is a San Francisco-based organization focused on identifying evidence-backed, cost-effective charities, primarily in international development and global health.
• Founded approximately 10 years ago by Holden Karnofsky and Ellie Hassenfeld, who previously worked in the hedge fund industry.
• Currently directs approximately $100 million annually to nine recommended "Top Charities" and operates an incubation grants program.
• The organization's research team consists of slightly more than 20 employees, split roughly evenly between research and operations.
• GiveWell recently added five new top charities to its recommendations: Malaria Consortium (Seasonal Malaria Chemo Prevention), Sightsavers, End Fund, No Lean Season, and Helen Keller International (Vitamin A program).

Research Methodology and Cost-Effectiveness Analysis

• GiveWell applies three core criteria for recommendations: cost-effectiveness, evidence of effectiveness, and transparency/commitment to monitoring and evaluation.
• The organization prioritizes direct health interventions (e.g., bed nets, deworming) where the evidence base is robust and homogeneous.
• To estimate moral weights and empirical parameters, staff utilize a "wisdom of the crowds" approach, taking the median of evidence-based guesses from the research team.
• A significant portion of GiveWell's cost-effectiveness analysis involves "internal validity" and "external validity" adjustments to headline study results to account for replication risks and geographic applicability.
• The organization performs sensitivity analyses to identify which parameters (e.g., discount rates, malaria income effects, deworming replicability) drive the most uncertainty in their models.
• GiveWell explicitly excludes "disability-adjusted life years" (DALYs) as a baseline for moral weights, allowing staff to prioritize preventing adult deaths over child deaths if their individual ethical frameworks support it.
• Recent work includes a blog post detailing the uncertainty in cost-effectiveness figures, with a public spreadsheet showing 40+ parameters estimated by 10 staff members.

Key Controversies and Ethical Decisions

• GiveWell made a $1.3 million grant to the Center for Pesticide Suicide Prevention to reduce suicide deaths in India and Nepal via means restriction.
• Internal debates exist regarding the moral weight of preventing suicide versus preventing homicide or death from disease, with some staff assigning zero weight to suicide prevention.
• The organization has shifted from ignoring leverage and funding effects to explicitly modeling them, which improved the cost-effectiveness ranking of deworming charities due to high matching funding.
• GiveWell estimates that marginal government spending on social programs is approximately 75% as cost-effective as GiveDirectly's cash transfers, a figure subject to high uncertainty.
• The organization distinguishes between "leverage" (charity funds causing others to spend more) and "funging" (charity funds displacing others), applying these concepts to policy interventions like tobacco and lead regulation.

Research Priorities and Future Directions

• GiveWell is expanding its research scope to include policy-based interventions (e.g., tobacco control, lead paint regulation) that operate outside traditional direct service models.
• The organization is exploring "evidence-based policy" to improve how governments in developing countries utilize research for decision-making.
• Researchers are investigating psychotherapy (e.g., via Strong Minds in Uganda) and antidepressant drugs as potential interventions for depression, though cost-effectiveness remains a barrier compared to current top charities.
• GiveWell faces a strategic choice between optimizing current top charities (filling funding gaps) and undertaking speculative research to find interventions significantly more cost-effective than current recommendations.
• The organization does not engage in global cause prioritization (e.g., comparing poverty to animal welfare), deferring that work to Open Philanthropy.

Work Culture and Hiring

• GiveWell hires for roles including research consultants, conversation note writers (part-time), and research analysts, often seeking candidates with quantitative backgrounds in economics or statistics.
• The internal culture is described as serious and focused, characterized by an "epistemic culture" where staff prioritize finding the correct answer over persuading others.
• Remote work from London is possible, though the San Francisco office is noted for being quiet and intense, with heavy reliance on written communication (Slack) rather than formal meetings.
• Employees report high satisfaction derived from making grants that would not otherwise exist and working with intellectually curious, collaborative colleagues.
• The organization encourages applicants to apply directly via its website, noting that part-time conversation note writing can serve as an entry point to full-time research roles.