How fast should foundations pull out of old areas?
One of the most important principles for grantmakers is the ability to update on new evidence and factors of the world. However, this pragmatic consequentialist updating can lead to tricky social dynamics with other actors in the space. A specific dilemma that I feel connects to this broader principle is the concept of exit grants and pulling out support for an area.
I have seen foundations we have worked with err on both sides of this spectrum, most rarely changing, for reasons more connected to existing relationships and staff than their current assessment of the impact of the area. On the other hand, foundations, particularly more consequentialist foundations, can often change quickly, changing the ecosystem as a whole in dramatic ways. I surveyed a handful of foundations and charities in the effective charity space to try to derive a decent rule of thumb or heuristics for how quickly a foundation should stop funding an area.
Principles that came up
In general, the larger the foundation was, the slower it made sense to pull out. This was both due to logistical concerns but also due to the expected effect on the ecosystem as a whole. Larger foundations are the biggest agenda setters and can have an even bigger impact than their raw budget would suggest.
In general, the less diversified an ecosystem, the more important it is for actors to enter or exit slowly, as a marginal funder has a stronger effect if they are a larger percentage of the pie. There was a bit of debate over whether this percentage amount should be estimated using the entire cause area (e.g., mental health) or a sub-component of it with similar values/epistemology (e.g., cost-effective layworker therapy in LMICs).
Staff are harder to ramp up or down compared to other services or goods. For example, a mass media charity that can double or cut half its budget with minimal staff change can typically handle a faster change than an organization that has a higher percentage of its budget spent on staff (a hard resource to quickly move up and down).
Possible systems
When surveying charities and foundations about the ideal theoretical rate, responses showed two different ways that people considered it:
One was a percentage-based decrease per year, e.g., cutting funding by X% each year from the amount given last year, e.g., a 50% cut per year would look like (e.g., 100%, 50%, 25%, 12%, 6%, etc.).
The other method was a flat decrease based on the largest amount given, e.g., a 30% flat yearly cut would look like (e.g., 100%, 70%, 50%, 20%, 0% etc.).
In terms of timelines, people were pretty consistently in the 3-10 year range for key funders in an area, depending on the importance of the specific funder to the area. There was also a high value placed on clarity and predictability, e.g., a clear plan for a 4-year cut might be better than an unclear 8-year one.
Typically, charities and other actors preferred slower moves over quicker ones, even when it was not clear if they were losing or gaining a funder in the space. Consistently losing a dollar from the space was harder for them to deal with than gaining a marginal dollar. Most funders thought that actions in unrelated cause areas had effects on people's willingness to work with and the general reputation of the foundation in other cause areas, whereas charity founders did not have a strong view on this.
Conclusion
Overall, people were more risk-averse and pro-slower-moving foundations than I would have expected, even when a foundation would move into their space. It updated me toward recommending foundations pull out of spaces slower than I would have previously (1-2 years → 3-5 years) for medium-sized foundations in smaller ecosystems (e.g., ~$10s of millions per year donated and making up between 1%-10% of the total ecosystem).