SoGive rates Open-Phil-funded charity NTI “too rich”
View this post on the EA Forum for improved formatting and images.
Exec summary
Under SoGive’s methodology, charities holding more than 1.5 years’ expenditure are typically rated “too rich”, in the absence of a strong reason to judge otherwise. (more)
Our level of confidence in the appropriateness of this policy depends on fundamental ethical considerations, and could be “clearly (c.95%) very well justified” or “c.50% to c.90% confident in this policy, depending on the charity” (more)
We understand that the Nuclear Threat Initiative (NTI) holds > 4 years of spend (c$85m), as at the most recently published Form 990, well in excess of our warning threshold. (more)
We are now around 90% confident that NTI’s reserves are well in excess of our warning threshold, indeed >3x annual spend, although there are some caveats. (more)
Our conversation with NTI about this provides little reason to believe that we should deviate from our default rating of “too rich”. (more)
It is possible that NTI could show us forecasts of their future income and spend that might make us less likely to be concerned about the value of donations to NTI, although this seems unlikely since they have already indicated that they do not wish to share this. (more)
We do not typically recommend that donors donate to NTI. However we do think it’s valuable for donors to communicate that they are interested in supporting their work, but are avoiding donating to NTI because of their high reserves. (more)
Although this post is primarily to help donors decide whether to donate to NTI, readers may find it interesting for understanding SoGive's approach to charities which are too rich, and how this interacts with different ethical systems.
We thank NTI for agreeing to discuss this with us knowing that there was a good chance that we might publish something on the back of the discussion. We showed them a draft of this post before publishing; they indicated that they disagree with the premise of the piece, but declined to indicate what specifically they disagreed with.
0. Intent of this post
Although this post highlights the fact that NTI has received funding from Open Philanthropy (Open Phil), the aim is not to put Open Philanthropy on the spot or demand any response from them.
Rather, we have argued that it is often a good idea for donors to “coattail” (i.e. copy) donations made by Open Phil. For donors doing this, including donors supported by SoGive, we think it’s useful to know which Open Phil grantees we might give lower or higher priority to.
1. Background on SoGive’s methodology for assessing reserves
The SoGive ratings scale has a category called “too rich”. It is used for charities which we deem to have a large enough amount of money that it no longer makes sense for donors to provide them with funds. We set this threshold at 18 months of spend (i.e. if the amount of unrestricted reserves is one and a half times as big as its annual spend then we typically deem the charity “too rich”). To be clear, this allows the charity carte blanche to hold as much money as it likes as long as it indicates that it has a non-binding plan for that money.
So, having generously ignored the designated reserves, we then notionally apply the (normally severe) stress of all the income disappearing overnight. Our threshold considers the scenario where the charity has so much reserves that it could go for one and a half years without even having to take management actions such as downsizing its activities.
In this scenario, we think it is likely better for donors to send their donations elsewhere, and allow the charity to use up its reserves.
Originally we considered a different, possibly more lenient policy. We considered that charities should be considered too rich if they had reserves in excess of the upper limit set by their own reserves policy. However, despite the fact that setting such an upper limit is good practice, it is rare for charities to have such a limit, or report it publicly.
We have conducted shallow investigations of hundreds, likely thousands of charities over the years, and in our experience it's common for charities to have reserves policies of 3 months or 6 months. A policy of 1 year is not an extreme outlier, but anything above that is.
Our methodology errs towards a number which is above the normal levels (i.e. 18 months); it’s possible for there to be unusual or idiosyncratic situations where a charity should opt for an unusually high level of reserves, even if they are employing good risk management thinking.
An important perspective is that it appears that many charities set their reserves policies without employing good risk management thinking. Instead, the target reserves amount is often anchored based on the amount of reserves they have at the time they set the policy. This means that having high reserves is correlated with (but not logically equivalent to) two things:
Popular with funders: the fact that they have tended to have high reserves is correlated with lots of funders wanting to provide them with funds, which suggests that they don’t need funds from you.
Unrestricted (and maybe restricted) donations are less likely to have impact, because the fact of having high reserves increases the probability of certain types of governance weakness. These issues have differing impacts depending on the type of donor/donation. This is expanded upon in Appendix 3.
We are saying that having higher reserves increases the probability of the two things listed above, not that those two things are definitely true of charities with high reserves.
These points are in addition to the definite fact that you would be giving money to a charity which already has lots of money, and arguably doesn’t need more money.
Note that a donor decision to not fund a “too rich” charity becomes high impact because it can unlock funds which otherwise would be idle. As a donor, you are less effective at unlocking those funds if you don’t communicate your thinking to NTI. Therefore, if you are interested in funding NTI, but decide not to because of the “too rich” considerations, then we encourage you to communicate this to NTI, indicating the amount that you would have donated, and that the “too rich” concern is the reason for not doing so. If you are not already in contact with NTI, you can communicate this information to us at SoGive by emailing sanjay@sogive.org, and we can pass the information on to NTI.
1.1 How confident are we in this policy?
Our confidence in the appropriateness of this policy depends on some fundamental ethical considerations. We have very high confidence (c.95%) that this policy makes sense assuming that (a) we have discussed the topic with the charity to make sure we have got our facts straight (as is the case with NTI), and (b) we adopt a deontological ethical framework (more on what this means in Appendix 4). Under a consequentialist framework (which we normally lean towards), our confidence in this framework is lower, but still quite high (>50%) if the donor is donating significant amounts and communicates their rationale to the charity, or alternatively if they are donating small amounts, but the charity is aware that lots of donors are being influenced by SoGive’s “too rich” rating. Because of these uncertainties, our framework allows for analyst discretion to adjust the rating away from the “too rich” assessment.
The above paragraph glossed over several nuanced topics; we’ve expanded on these in Appendix 4.
2. Background on Nuclear Threat Initiative and their reserves
The Nuclear Threat Initiative (NTI) is a moderately large NGO based in the US which has received funding from Open Phil for its work on biological risks. We understand that Open Phil may issue another grant to NTI.
According to NTI’s most recent form 990 (31/Dec/2021), they hold around $85 million of net assets. They spend around $20 million a year, which suggests that their reserves are in excess of four times as big as their annual spend.
We find that the financial reporting in Form 990s in the US is less clear/transparent than charity accounting in the UK. As a result, we were keen to check with NTI to ensure that our understanding was accurate. As can be seen in the next section, we had a discussion on this topic, and they did not disagree with this interpretation of the accounting, although they also didn’t appear to demonstrate much understanding of their own financial position.
They indicated to us as well that the policy is to hold between three and five times as much reserves as their annual spend, and that, for the purposes of this policy, reserves refers solely to backup funds for the case where their fundraising fails, and not to funds held with a specific spending purpose in mind. This is further evidence that our understanding is at least roughly accurate.
At first glance, this seems excessive when a factor of merely 1.5 times would already be considered high.
2.1 How confident were we that NTI’s reserves were high?
Prior to the call with NTI, we had some doubts, because of the weaknesses in US charity reporting. However, given the indications that NTI had already provided us in writing, it seemed very likely (c.75% chance) that their reserves were in the range 3x-5x annual spend. After the call, this confidence increased to c.90%.
3. What we learned in our call with NTI
On 23 May 2023, we held a call with NTI to discuss their reserves.
The first thing they said was that the only NTI people on the call were development (i.e. fundraising) staff, and they were not able to guarantee that more senior staff would agree with what they said.
SoGive comment: SoGive does not typically allow for generous allowances to be made for situations like this. I.e. we believe it’s naive to say: “let’s assume that if someone more senior were present, then the answer would be a good answer”. Doing so can, in general, provide incentives for organisations to dodge questions by not having the relevant people present.
It is also typical for organisations to include senior staff in such calls where the potential donation size is sufficiently material – as the grant size available was potentially in six figures, this conversation was likely sufficiently material. Note that there are reasons why NTI might have chosen to deprioritise working with SoGive. This is expanded upon in Appendix 2.
The first thing they said when we reiterated the question about reserves was:
Some years ago they did some work on the LEU bank in Kazakhstan
They budgeted conservatively, and had money left over
Because Warren Buffett (a material source of the funds) thinks highly of NTI, he was willing for the leftover funds to stay with NTI
SoGive comment: we found it surprising that this was the first thing mentioned. This answer seems to suggest that reserves are calibrated based on the question “how much do we have” and not on good risk management, nor an awareness of the opportunity cost of the money sitting in reserve.
In addition, they made the following points
The risks from nuclear weapons are important
Those risks are particularly elevated at the moment
The amount of funding available to tackle nuclear risks has declined significantly in recent years.
NTI has a good track record.
SoGive comment: we believe those points have merit, and constitute generic arguments in favour of providing funding to NTI. These also argue against the “popular with funders” concern set out above. Indeed, in SoGive’s view, it would be unsurprising if some charities tackling nuclear weapons risks faced severe financial hardship in the near term (but not NTI). On their own, these points are not enough to explain why the reserves appear to be so high. Indeed those arguments seem to suggest that if the organisation had $85m of net assets, they would use them to tackle the pressing concerns which lay ahead of them.
3.1 NTI may have plans to spend down their reserves
When we asked them whether they were willing to spend down their reserves, they said:
They acknowledged that it was possible to spend down their reserves.
They mentioned no plans to do so.
They pointed out that the amount they raised last year was less than the amount they spent – i.e. they ate into their reserves last year.
SoGive comment: At first glance, it seems that this may be sufficient to explain their high reserves – if NTI ate into their reserves last year, and expect to do so in future, then the reserves are not simply sitting idle. However, we find this confusing. Firstly, several months prior to arranging this call, NTI had indicated to us that their reserves target was to hold 3 to 5 years of reserves. Under our understanding of the term “reserves”, it’s a contingency fund as a backup in case fundraising is not sufficient. It’s not funds which are set aside where there is a plan for the funds to be used. In those previous written communications, NTI confirmed that this was their interpretation of the words as well.
When we asked whether they could share with us any forecasts indicating how they will spend down their reserves in future years, they did the following:
They did not agree to share any such evidence
They questioned whether other charities ever provided such forecasts (Our reply was that although we have reviewed a large number of charities, we rarely encounter charities with such a large amount of reserves, so we have never requested a multi-year forecast before. For brevity, we did not mention that charities frequently provide us with a budget for the coming year, and that this is an entirely normal level of transparency)
3.2 How confident are we that NTI’s reserves are high after our call with them?
The call increased our confidence that NTI holds high reserves. This is because:
NTI were told in advance that the call was about reserves, but did not choose to prepare for the call by having the facts available about their reserves. This is a logical strategy if they know that their reserves are high.
NTI hinted that their rationale for holding high reserves was to do with past fundraising success, and not grounded in good risk management.
We actively sought ways that their reserves might not be as high as they look – i.e. the possibility that they might forecast a spend-down of their reserves. When we encouraged them to demonstrate that this is the case, they declined to do so.
We believe that it is still logically possible that NTI is not as rich as it appears, however our experience speaking with them made us less likely to believe that this is the case. It moved our assessment from c.75% to c.90% chance that they are too rich.
4. Conclusions: We rate NTI as “too rich”
For charities which hold reserves in excess of 1.5x annual spend, our methodologies indicate that we should default to a rating of “too rich”. In NTI’s case, we have explored with them whether we have reason to deviate from this default assessment, and as things stand, we do not.
It is possible that if they chose to share with us more information about how they determined their reserves policy, including forecasts of future income and expenditure, we might revise our rating.
As outlined in the section on SoGive’s methodology above, where donors are choosing not to donate to NTI because of “too rich” concerns, we encourage donors to communicate this to NTI, either directly or through SoGive (by emailing sanjay@sogive.org). Note further that in SoGive’s methodology we condense a number of nuanced considerations into a simple rating to make things cleaner for users of our work, however under certain circumstances (the donation is large, restricted, and the donor doesn’t have certain deontological preferences) we may judge that it’s suitable to overturn the “too rich” rating.
Appendix 1: Risk management considerations for setting a reserves policy
When setting a reserves policy, you’re balancing two competing desires:
To hold as much money as possible to avoid the risk that the charity goes bankrupt
To hold as little money as possible because every penny has an opportunity cost and the charity’s work is important.
To balance these competing desires, you should consider relevant factors to help you decide what your target should be. Examples include:
Donor concentration/diversification: if you are heavily reliant on a small number of donors, you’re more exposed to risk if one of them decides to stop funding you.
Donor attitudes to your organisation: if donors have indicated that they have very high trust in your organisation, this reduces the probability that your organisation will be short of funds.
Certain types of financial risk. Examples:
FX risk: if you receive funds denominated in a currency other than the currency you spend, you may be exposed to the risk of currency fluctuations.
Pension scheme risk: if you have provided a defined benefit pension scheme to staff (this is less common in charities set up recently) you may be exposed to the risk that assets or liabilities within the scheme fluctuate, potentially leading to a deficit which the sponsor would be obliged to cover.
Contingent liabilities: e.g. if an orderly wind-up of the organisation required the organisation to pay redundancy payments, it seems reasonable to incorporate this need in the target reserves amount.
Beneficiary needs: if the charity’s work involves working directly with beneficiaries, and if the charity feels that by starting a lengthy programme of work with a beneficiary, they are obliged to finish what they started, they may wish to hold enough reserves to ensure they are not forced to stop a programme of work midway through. This would likely only apply with certain types of work, e.g. if stopping partway through was likely worse than not working with the beneficiaries at all.
A charity which considered itself in a favourable position on the above may choose a lower reserves target (e.g. 3 months) and charity which considered itself to be in an unfavourable position may choose a higher target (e.g. 12 months).
Appendix 2: Why NTI was less willing to work with SoGive
NTI articulated two issues, both of which are potentially reasons why they might be less willing to invest time in conversations with SoGive. This appendix sets out what those issues are, and why they don’t cause us to deviate from our rule of taking what NTI has to say at face value.
We wrote a piece about NTI in 2020 which portrayed them positively. NTI said that they felt “betrayed” because we had not done enough to flag in advance that we were going to write this piece.
It’s hard to comment on this when it’s flagged 3 years after the fact, however we normally find that charities we work with value the opportunity to have another channel of communication with the public domain, and so we’re not always very careful about warning about this. Therefore it is entirely possible that I[1] neglected to mention this. So I apologised straight away for this. (Note that we did give NTI a chance to review our draft before publishing, which charities typically value greatly, and is a standard part of our process.)
NTI became less positive about EA after the FTX collapse, and are therefore less keen to invest in talking to EA groups.
We believe that NTI are entirely entitled to prioritise whichever donor groups they prefer, and they have every right to expend less senior time and effort speaking to us. However, we don’t think this is a strong enough reason to deviate from our rule from taking what they say at face value (as opposed to assuming a more charitable intervention than what NTI said).
Appendix 3: Why being too rich might mean donations are less likely to lead to more work
As outlined above, unrestricted (and maybe restricted) donations are less likely to have impact, because the fact of having high reserves increases the probability of certain types of governance weakness, which in turn means that the donation is more likely to simply increase the reserves rather than lead to more work being done.
Part of an effective governance structure is that the target amount of reserves should be based on risk management considerations. Examples of relevant considerations are outlined in Appendix 1.
Another element of strong governance is having mechanisms which help the charity to notice if they reach levels where they consider themselves to be holding too much money, and take actions to address this (e.g. spending more, regranting, or relaxing fundraising efforts).
Although it’s in principle possible to be strong in one area and not another, there tends to be a correlation between the two.
If a charity is not managing the risk that they might hold too much money, then it’s more likely that an unrestricted donation will simply increase reserves without ultimately leading to the charity doing something they otherwise wouldn’t have been able to do.
This may also apply to restricted donations in some cases, because of fungibility, aka illusory restrictedness.
This means that the charity being too rich has differing implications for different donors. For a donor providing a substantial restricted donation, the effect of the donation is likely to be more activity, i.e. the charity being too rich doesn’t materially reduce the impact of your donation, although it does mean that you’ve missed an opportunity to encourage the charity to produce a stronger justification for why they hold so much money – money which could be used to do (hopefully high-impact) work. For someone donating smaller amounts (or anyone making an unrestricted donation), it’s much more likely that the donation will lead to an increase in reserves without ultimately leading to more activity happening.
Appendix 4: Our confidence in our policy
Earlier in this document, we said that our confidence in the appropriateness of our approach to rating charities as too rich depends on some fundamental ethical considerations.
Appendix 4.1 Deontological ethics
Under a deontological framework, we have much higher confidence that donors should avoid donating to charities which are too rich. The deontological framework we have in mind is the Kantian categorical imperative, which says
"Act only according to that maxim whereby you can at the same time will that it should become a universal law."
(Note that there are other ways of thinking about deontological ethics, which we have ignored for brevity)
If we apply the categorical imperative to this context, we are saying that if everyone chose not to fund charities which are too rich, then good outcomes would happen (e.g. either the charity uses up its excess reserves, or they come up with more spending plans).
Conditional on it definitely being true that the charity is too rich, it’s entirely clear that choosing not to donate to them is a good choice for donors following the categorical imperative (assuming some other standard assumptions). The reason for our confidence in this being below 100% is that we might not be totally sure that the charity really is too rich, e.g. because it’s possible that the charity is making accounting errors or otherwise has poor governance.
Appendix 4.2 Consequentialist ethics
The consequentialist perspective is indifferent to the question: “what if everyone acted the way I’m acting”. This is because it’s simply not true that everyone is acting the way I’m acting.
Consequentialist donors are still likely to converge on the conclusion that they should avoid funding charities which are too rich, however with less confidence and with more caveats.
The consequentialist concern with SoGive’s recommendation is that the charity’s decision to hold high reserves de facto “locks away” that money, and as far as the donor’s decisions are concerned, they can’t influence that money. This would mean that their donation may be enabling the charity to do high impact work.
We think this is typically an overly simplistic way of looking at things.
Firstly, as outlined above, our experience reviewing a large number of charities suggests that charities which are too rich are normally popular with funders.
Secondly, also as outlined above, “too rich” charities tend to have the governance concerns outlined in Appendix 3. This governance consideration centres around the question: at the margin, how likely is another donation to lead to impact, as opposed to propping up already healthy reserves?
Thirdly, as a donor, you can have influence over the organisations you fund. By pressing them for more details on their reserves, you can encourage them to either improve their governance and clarify what plans they have (if any) for their money, or encourage them to unlock reserves and use them if that makes sense.
Appendix 4.2.1 Applying consequentialist thinking to NTI
Does this apply to NTI?Is this a pro or a con for funding NTI?Brief comments (more comment below)(A) Popular with fundersProbably notProWe don’t see NTI as crowded with funders.(B) Governance issues (marginal $ → impact or reserves?)ProbablyConIt’s hard for us to comment on this, but on balance we are more inclined to believe that this is an issue.(C) Will respond to donor pressure?Unclear, probably not?Unclear, maybe a pro?Multiple factors are at play (see comment below)
Applying these considerations to NTI, it seems that:
NTI is not unusually popular with funders (c.90% confidence – we have seen several charities which are popular with funders, and they are typically in cause areas which are much more popular than nuclear war)
It is hard for us to comment on whether NTI has the governance concerns outlined in Appendix 3, because of the lack of transparency on NTI’s part. As outlined earlier, we tend to err in the direction of not assuming the most generous assumption when a charity chooses not to be transparent, so we think there’s a reasonably high (c.80%) chance that NTI has these governance concerns, although note that the probability that this is problematic reduces for donations which are large and restricted.
We think there’s a fairly reasonable chance that NTI won’t listen to our concerns if they come simply from SoGive alone – we were never likely to send them much more than $100k in a year. Note that if we were joined by several other donors, this might be more likely to effect change. Furthermore, the fact that NTI views SoGive as being linked to the effective altruism movement makes them less likely to heed our perspective, as they may be less prone to prioritise seeking funding from the EA community now. On the other hand, it is very plain that not wanting to fund someone who is too rich is not specific to the EA community (indeed not obviously a trait of the EA community). As outlined earlier in this appendix, it fits better with a deontological viewpoint, which is not typically associated with effective altruism. This again reduces the chance that our recommendation to avoid donating has an impact.
It is more common for the above three items to consistently point towards it being suitable for consequentialists to avoid supporting the charity, and in NTI’s case items A and C are not so favourable to our recommendation to avoid donating.
We reiterate that a consequentialism based decision to donate to a charity which is too rich could reasonably be summarised as “I believe the charity’s weak governance is a good reason for me to donate to them”, which highlights the risks with this strategy.
Nonetheless, we think it’s around 65% likely that our recommendation to donate is appropriate as viewed through a consequentialist lens. Note, under consideration (B), this probability decreases if you are a large donor making a restricted donation. Note that under consideration (C) this probability would increase if there were a large number of donors (who might be donating small amounts) who followed our advice, or a small number of major donors.
Switching from first person plural (“we”) to first person singular (“I”) to avoid the sense of hiding behind organisation-wide responsibility when I’ve made a mistake.