To what extent do solutions to global problems differ in terms of effectiveness?

27th June 2024, Reading time: 5 mins.

As far back as 2013, Toby Ord highlighted the major differences in global health in his article “The Moral Imperative towards Cost-Effectiveness in Global Health” . Ord examined the cost-effectiveness of measures drawn from the World Bank’s publicly available “Disease Control Priorities in Developing Countries (second edition – DCP2)” dataset. DCP2 compared 108 interventions in low-income countries, ranging from surgical interventions to vaccines and mosquito nets, as well as public health programs such as free condom distribution for AIDS prevention. Ord examined how many disability-adjusted life-years (DALYs) were gained for every thousand dollars invested.

His analyses show that the most effective intervention was around 15,000 (!) times more effective than the least effective. The 2.5% most effective measures were around 50 times more effective than the median, and 23 times more effective than the average. For every thousand dollars invested in the most effective intervention, they could contribute 50 times more quality-adjusted life-years (SALY) than the median (or 23 times more than the average). (The differences are so enormous that it’s worth reading the above paragraph twice to realize them).

Cost-effectiveness of health interventions according to the Disease Control and Priorities Program 2 (DCP2). 

The most effective interventions therefore generated a disproportionate share of the total benefit. If all DCP2 interventions were funded equally, 80% of the benefits would be generated by the 20% most effective interventions. If we choose an average DCP2 measure rather than one of the most effective, we lose over 90% of the potential benefits that the most effective measures could have generated with the same resources. 

Recently, Benjamin Todd of the 80,000 hours organization investigated whether this effect was found in other datasets. To do so, he examined a wide range of measures and thematic areas, including the recent DCP3 study, WHO-CHOICE data, health measures in the UK and social policy in the USA, but also a renowned study on the cost-effectiveness of climate protection measures and education data. Again, the same huge differences: in all datasets, the 2.5% most cost-effective measures were around 20 to 200 times more effective than the median, and around 8 to 20 times more effective than the average.

Taking a conservative approach, Todd concludes that the differences may in fact be smaller than his results suggest. For example, it could be that the best interventions are already no longer in need of funding and that, as a result, the best available interventions are slightly less effective than those studied. In addition, the data are generally retrospective and perhaps over-optimistic for the future. Finally, all analyses are based on assumptions which, by necessity, cannot perfectly reflect reality. As a result, good results often seem even better than they are in reality, and bad results even worse – the margin between highly effective and less effective measures may therefore seem greater than it actually is.

But there are also reasons to believer that these studies underestimate even the difference between very good and average interventions. For example, the data sets mainly include easily measurable metrics, whereas the most effective interventions (e.g. in the field of advocacy), are not only more difficult to measure, but also very rarely recorded. What’s more, studies generally focus solely on the direct effects of measures. Indirect effects, also known as co-benefits, are not usually taken into account. An association that promotes its impact and participates in numerous scientific studies helps to ensure that humanitarian aid and development cooperation as a whole focuses more closely on the issue of effectiveness. Yet most cost-effectiveness analyses fail to take this effect into account.

In the case of impact grants, we assume that the best of all measures within a thematic field are around 10 times more effective than the average, and even up to 100 times in some cases.

What role does the choice of cause area have on the impact of an intervention?

The differences mentioned above only concern interventions within a cause area, for example interventions in the field of global health. However, the choice of cause area itself makes an equally important – or even more important – difference in the impact achieved.

The cause areas that are the most impactful, in terms of the amount of good they achieve, share a number of common characteristics. They are :

• Large-scale – involving a particularly large number of people or other living beings

• Neglected – they receive comparatively little attention or resources

• Have high potential for improvement – there is evidence that additional resources, including donations, can improve the situation.

Two simple examples illustrate the differences. In the EU, around 2.7 million people are diagnosed with cancer every year, and 1.3 million die from it. The cost of cancer treatment in Europe is around 103 billion euros a year. By comparison, malaria still kills around 600,000 people a year worldwide, but global expenditure is only around $4.3 billion a year. In relation to the number of deaths each year, the fight against malaria therefore receives far fewer resources than cancer. As a result, additional donations to malaria charities are likely to have a far greater positive impact. 

What’s more, malaria is most prevalent in the poorest countries in the world where money can buy much more than in France. So with the same amount of money, you can achieve a lot more by donating money to poorer countries. 

Gross domestic product per capita (in purchasing power parity) – 2022

Source: The World Bank

As a result, the choice of cause area multiplies impact by 10 or more. And the choice of intervention within a cause area also multiples the impact. Therefore a donation to a highly effective charity in the most effective cause area easily achieves 100 times more good than an average donation – great news for donors who want to do as much good as possible with their money.