Almost a decade ago I wrote in Treasury guidance that “Fixed percentages without any evidence base do not provide a sound basis for the calculation and award of relevant overhead costs” and that “Under-funding of overhead costs can be detrimental to third sector organisations as it may reduce their ability to maintain their infrastructure and retain quality in management and service delivery, compromise their duty as responsible employers, and stifle innovation”.
So it saddens me to have to point out that Caroline’s recent work - which aims to show that high-performing charities actually spend more on administration costs than weaker ones - in fact falls short of making this case convincingly.
Let’s look at the methodology - the crucial point here is how 'high-performing' is measured. GiveWell – where Caroline's data comes from - use various indicators or criteria to score high performance:
1. Strong evidence of positive impact on people's lives - crucially, this is about whether a charity gathers evidence and demonstrates impact, not whether it delivers it (of course the latter is impossible to understand without the former). GiveWell award more points for those who undertake Randomised Control Trials, for example. They also point out that demonstrating impact can be extremely expensive.
2. Highly cost-effective activities - GiveWell score charities highly for high life impact (measured in things like DALYs) and for low direct costs associated with interventions.
3. Room for more funding - GiveWell look at what would the charity would do if it received more funding. While this can certainly be useful for prospective donors, it isn't remotely a measure of current effectiveness.
4. Transparency and accountability to donors - GiveWell reward charities willing to share enough in-depth information about their work.
So in essence, It looks like Caroline’s work is in fact concluding that (ignoring the irrelevant third criterion) those charities which...
- put central resources into proving that what they do works;
- have low direct costs; and
- put central resources into sharing information about their work...
...have higher indirect costs!
That sounds pretty robust and logical to me. But what it isn't, is proof that high-performing charities spend more on administration costs. This is a real shame.
If I've got this right (and would really welcome any feedback) this means I have busted the myth of the link between administration costs and effectiveness. Sadly, not in the way I would wish.